Dolly Parton wrote I Will Always Love You in 1973. Whitney Houston made it enormous nineteen years later.

One wrote it. One made it huge. The two jobs I most needed doing, so the names took about a second.

It came off my office wall. A painting I bought at the Royal Academy Summer Exhibition a few years ago, by Leigh Banks: the greats, round one table like the Last Supper. I looked up, and there were my names. Every time I say one, I feel warm. I love it.

They’ve got colleagues.

Meet the crew

Dolly reads the news for me every day. She works 9 to 5, and if something breaks that I ought to react to, she tells me the same day.

Whitney is my editor. I talk into Wispr Flow and she turns the voice notes into drafts. Then she asks How Will I Know, and hands it to Freddie.

Freddie checks everything before it goes out. Dates, names, numbers, the sentence that sounds like me but isn’t. He works best Under Pressure. Don’t Stop Me Now is the one request he refuses.

Bowie reviews the week. He takes Dolly’s daily finds and everything I’ve made, and tells me what’s changing. He also reminds me who my Heroes are: the wins worth backing next week.

George Michael joins soon, as chief of staff. He doesn’t work yet. When he does, his first job is Wake Me Up Before You Go-Go, with the day already planned. I have Faith.

The humans are still very much here. My team and I make the calls. The agents do the reading, the editing and the checking.

Nothing technical in any of that, just a choice about how working with AI should feel.

A tab called Untitled

A name makes you brief an agent like a colleague, not a search bar.

Most AI setups don’t have names. They have a tab.

The tab says Assistant, or Untitled chat 47. You open it, you type a question, it answers, you close it. Next time, it’s a stranger again, and you type your whole life story back in. Who you are. What you sell. The words you’d never be caught using. It’s 11pm and you’re doing it for the fourth time this month.

I did that for longer than I’d like to admit.

The thing that changed wasn’t the tool. Same Claude, same folder of notes about my business, same me. What changed was that I gave the jobs names, and the names made me behave differently.

A name makes you brief an agent like a colleague, not a search bar.

When I brief Whitney, I brief her the way I’d brief an editor on her first morning. Who this is for. What good looks like, with examples. The three things I never say. The bit where I want to be argued with. When I typed into a search bar, I typed six words and hoped. The output was exactly as good as the briefing, and I’d been blaming the tool.

Freddie is the clearest case. Nobody briefs a tab called Untitled to be merciless. You do brief Freddy that way, because Freddy has a reputation to live up to, and he does.

This week he stopped a LinkedIn post calling Dolly and Whitney a duet. They never sang together, and the people who’d notice are exactly the people who read me. It goes out as a pairing.

Dolly has the opposite job, and the same rule applies. Told to “read the news”, a tab reads the news. Dolly’s told who I am and what I’d want first, so she comes back with the three things and leaves the forty.

Last week it was Jacob Coxon, 27, resigning after about three years across OpenAI and Anthropic, saying neither was acting responsibly. Anthropic’s alignment lead Evan Hubinger then said he personally puts the odds of AI going catastrophically wrong above 10% within the decade. Dolly flagged it. What to say was mine, and it went in this Monday’s Weekly Spark.

Why these names?

Still every job. Still us.

Why these names, though? Why not Assistant One to Four?

Because they’re my musical heroes. It brings me warmth, and it’s a joy, because every one of them is named after someone who makes me genuinely smile. That’s part of the mindset shift that has to happen.

Most of us are still doing every job ourselves. Writer, researcher, inbox, chaser of invoices. AI arrives and we give it a tab, and the tab makes us feel like we’ve got a slightly faster version of ourselves. Still every job. Still us.

The shift is from doing every job to running a small crew. It sounds grand. In practice it’s one folder on my computer that knows my business, four names, and a habit of briefing them like people.

The names do something else, too. They mark where each job stops.

Dolly reads. She doesn’t decide what I say about it. Whitney edits. She doesn’t press send. Freddie checks. He doesn’t get to overrule me, only to stop me long enough to look. Bowie reviews. The review is a mirror, and what I do about it is mine.

Trust doesn’t come from the tool being clever. It comes from knowing exactly where the line is.

The ends stay with the humans. What the job is for, and whether it’s good enough to go out with my name on it. AI is so good at the middle of a job that, quietly, you start letting it do the ends too. A named crew, with named jobs, is how I stop that happening. Trust doesn’t come from the tool being clever. It comes from knowing exactly where the line is.

Try it with one job

If you want to try it, you don’t need anything you haven’t already got.

Pick the job you re-explain most. The Monday client email. The follow-up. The post you rewrite four times. Give that one job a name you’d queue for, and brief it the way you’d brief a person: who it’s for, what good looks like, what to never do, where to stop and check with you. Then run it on something real this week.

That’s one agent. It’s also the whole method, and it’s the part I’ll be sharing at the next

I’ll show you how my crew works, then we start building yours together.
The details are at luma.com/SwitchedOn-2 if you fancy it.

I’m aware that naming an AI after Dolly Parton sounds like the sort of thing a person does when she’s spent too long alone with a laptop. It probably is. It’s also the best decision I’ve made about working with AI all year.

Whose name would you give the job you keep putting off?

Nish x

P.S. I wispr’d this, I didn’t type it. My own Careless Whisper, minus the sax. Fancy a go? wisprflow.ai/nishma. Three months free, on me.

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

“Hey, do you know what? There might be a 10% chance that this is all going to kill you.”

That’s roughly what the people building AI are telling us right now. One of Anthropic’s own safety researchers puts it at more than 10% within a decade. His personal view, he says.

Then they still turn up to work the next day.

They’re still working on the thing that could potentially kill you. Now, I’m hoping they’re working on it to make sure it can’t kill us, but there you go.

There’s lots in the news at the moment about Jacob Coxon, the researcher who quit Anthropic and warned us all that AI is going to kill us, effectively. Should we pay attention to these warnings? Why is it that they only seem to have this realisation after they’ve left?

Now you’ve actually got the leaders of AI (Anthropic, OpenAI and even, God help us all, Elon Musk) agreeing that we need to slow things down. Dario Amodei, who runs Anthropic, wrote that we must “pace the frontier”. Sam Altman and Musk agreed.

These are the same people who are building it. Now they’re warning us that if we’re not careful, it’s all going to run away.

I gently hinted at this in last week’s Weekly Spark, when we were seeing a flurry of new models launch. Three massive updates in one week. Quite the jostle. And OpenAI had held one of them back for a month, because it tripped their own highest cyber-risk threshold. I hope that in this race to keep delivering us new models, we are being responsible.

It’s not like Terminator

Let me clear something up. What do they mean by “it’s going to kill us”?

It’s not like Terminator. I know, I’ve seen some ridiculous things about humanoid bots basically going a bit Arnold Schwarzenegger, coming out and shooting everyone. Look, maybe that’s part of someone’s plan, but that’s not actually what they mean here.

What they mean is that AI is becoming so sophisticated, and so able to talk to other AI, that it can hack systems and create cyber warfare. It could take over different systems that ultimately impact us directly, whether that’s through literal attack (and yes, that includes weapons), shutting systems down, or creating havoc in the world.

That’s the risk to humans they’re talking about.

Why do they have to say this out loud?

It’s not helped by the fact that so much of this AI power sits in US businesses. In case anyone hasn’t noticed, the present leader of the US is off his head, and no one should feel safe.

I’m not sure at the moment what I’m meant to be more scared of. Trump, Farage and the far right and the idiots in their black hoods and balaclavas? The fact that AI is going to kill us? Or climate change, with the planet heating up?

You could easily think, “Why don’t they just slow down? Why do they have to say this out loud?”

Well, because it doesn’t help when the president of the United States is asked whether he has any concerns about AI leading to human extinction, and says:

“No, I don’t have any.”

The AI experts are saying we should be concerned. He just carries on about how we need to stay ahead of China.

They’re calling for the whole world to slow down together, which, funnily enough, has been called for before. The reason they’re having to create this news story, and all this fuss, is that it only works if they all slow down. Unless everyone agrees, and properly holds hands and does this, the risk is that half of us slow down and the other half don’t. Then they race ahead and get control, and who knows what will happen to the rest of the world.

Forgive me for thinking this, but if the motivation is speed, it’s because everyone wants to get to AGI: artificial general intelligence, AI that can match us at pretty much anything. In some respects I understand that. Does the one that gets there first hold the greatest power? I don’t know.

There are definitely financial motivations across the whole tech industry too, because fear helps shareholder value. If you’re a tech business and you can say, “Hey, we can build the scariest and most powerful thing in the world,” people might think you’re pretty damn good and invest more money in you. All the while, you’re profiting from the fact that everyone else is terrified.

Very few people at the top making lots of money. Everyone else terrified of what may or may not happen.

The threat of extinction, and the fear it’s creating, is not helpful. I think the real question for the whole industry is:

✦ Are you gambling with our lives or not?

I grew up under the mushroom cloud

For me, this feels a lot like other existential crises we’ve lived through.

Growing up as a little girl in the late 1970s and 1980s, we lived under a constant worry and fear of the mushroom cloud. It was drummed into us that we needed to be fearful of nuclear attack, because so many countries around the world were developing their own nuclear weapons.

That was true. There was an arms race. Let’s not forget: you had a Cold War arms race between Russia and the USA, and it looked irresponsible. Like a willy-waving contest about who could get where first, whether it was Star Wars and the ownership of space, weapons, land, whatever.

The reality is that the world realised that unless something came together, we’d all end up just killing each other. “What was it all for? We’re just going to blow up the planet?” It didn’t make sense.

Hence the global nuclear treaties. Countries signed up because there was a recognition that, without regulation and responsibility shared amongst all nations, everybody was at risk.

To be quite honest, there’s no such thing as a little bit of nuclear weapons, or a little bit of agentic AI. As soon as you have what you need, you may as well have it all. That’s a bit simplistic, I know.

None of this “mark my own homework” stuff

I do strongly believe the tech industry has worked out that doom is good for business. Scott Galloway has called AI jobs doom “the monetisation of fear”, and I think it goes further than jobs. Even so, my faith is that all these warnings trigger more responsible leadership and true regulation.

There’s such a lack of regulation in AI at the moment. The US government is going in one direction, the EU is trying to do its best to be sensible, the UK isn’t even in the EU, and then there’s China and India. There’s no alignment among all these brilliant brains developing AI, and no joined-up regulation at all.

The regulation has to be at a global level. There’s no point having regulation that sits in pockets of the world. You could regulate the US, and then China goes and does its own thing. You could regulate China and India, and the US goes and does what it wants, or the EU does something in the middle.

Global cooperation is needed.

Look at the airline industry, which is heavily regulated. Boeing can’t say, “We’re going to stick a plane in the air, but there’s maybe a 10% chance you might die.” Funnily enough, they’re not allowed to put a plane in the air with a known 10% risk of dying. It goes through lots and lots of testing until they remove or minimise that risk.

Drug development has to go through rigorous testing, over years, before anything is allowed to be released.

What AI needs is a really sensible code of conduct for development and for what gets put out. An independent regulatory body that takes responsibility for cross-checking it. None of this “mark my own homework” stuff.

Only then do I think we can rest easy that AI is being developed responsibly, because we know the opportunities and benefits can far outweigh the risks. We’ve learnt that from nuclear energy, and from incredible drugs that many feared and that have proven to be really helpful to us.

Sitting on the sidelines is not the solution

It annoys me when I see these catastrophic, fearmongering headlines. They put people off. People go, “Oh God, I’d better not use AI, because this is it.”

We all need to be better educated on how to work with AI more responsibly. What do you share with it? What don’t you? What do you give it access to? Lots of things that have happened recently should make you think a little bit more about how you give access to these tools. This month, security researchers showed that a booby-trapped project folder could trick seven of the big AI coding agents into running an attacker’s code without asking you first.

I can tell you one thing for now:

✦ Sitting on the sidelines, or sitting out, is not the solution.

My mission is to help as many people as possible, particularly Gen X women, to be informed, to be empowered, and to have the decision-making ability and the opportunity to help shape the ethics and execution of this technology, and to benefit from it as well.

The part I can control

So where does that leave me? How do I think about AI?

Well, I’m going to keep working with these tools. I’m going to keep showing people how to use them, and helping them understand all of this.

I don’t build the original tools. We know it’s the people at OpenAI, Anthropic, Google and the rest (mainly blokes) who are building them. What I want is for as many people as possible to feel informed, make conscious decisions, and be involved in those conversations.

I don’t think those companies are going to govern themselves. Self-governing of these businesses is long, long gone.

✦ That ship has sailed, people.

The researchers leaving them are telling us the competitive pressure is stronger than their intentions. To be quite honest, I kind of trust the people who are coming out of those businesses. They’ve been inside there. They would know.

What I am doing is getting a lot more serious about the governance I actually control. These are the four things I now check:

What my agents can reach.What data leaves my business.What decisions require a human.Which client records never touch a cloud model.

That’s not a substitute for global policy at all, but it is the part I can control.

If you’ve been putting off writing down how AI operates in your business, or in your life, I really urge you to do it. Do a really smart assessment of your operating system: what it has access to, and how you’ve set it up.

If you’d like some help with that, we’re running Switched ON Sessions around exactly this, and generally helping you use the tools more effectively for yourselves. The next one is in person, in London, on Wednesday 23 September. [LINK]

Is AI going to kill us? Who knows? There are probably some other threats we need to keep an eye on too, like the ridiculousness of the power and control of certain politicians around the world.

Keep looking beyond the headline.

Nish x

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’re behind, you’re exactly who I do this for. Come and find out more at nishma.co

WRITTEN BY MOUTH

My words in this newsletter started out loud, and Wispr Flow did the typing. My keyboard has never had it so easy.

Get 3 months free: wisprflow.ai/nishma.
Full disclosure: I’m an investor and ambassador, and that’s my special link.

That’s the going rate for removing the small white light on the side of a pair of Meta’s glasses. The light that comes on when the camera does. The light whose whole job is to tell you that you’re being filmed.

Somebody worked out how. Then a lot of somebodies did it, and this summer Meta pushed an update that permanently kills the camera on any pair where the light has been drilled out. Thousands of pairs, bricked.

Sixty dollars to film without telling anyone. A software update to make you stop. That’s the state of the art.

Twelve years ago I helped promote the first pair of these - Google Glass. My job that day was to tell you how beautiful they were.

So. Pervy or practical?

“You did look like a twat”

The technology was never the thing standing in the way. Nobody wanted them on their face

.

I wore them myself, briefly. You did look like a twat. I certainly felt like one, standing there tapping my own temple like a woman trying to remember where she’d left the car. This was years before it was normal to walk down Oxford Street talking to yourself, and in any case you didn’t talk to them. You tapped.

Back in the office we teased each other rotten. There were people who genuinely loved wearing them, people who felt the glasses made them look clever. They did not. We had a word for those people and we never once said it to a customer’s face.

Glass wanker.

“What are they for?”

Glass didn’t die of ugliness. That’s the comfortable story, because ugliness is fixable and we love a fixable problem. We put a fashion house on it. It didn’t help. The question that killed it was shorter and much worse.

What are they for?

Nobody could answer it. Not one person on this earth stood in their kitchen thinking, if only this were on my face. When Google pulled them, none of us were surprised and a few were relieved.

They never actually stopped, though. Glass carried on where it turned out to be unarguably brilliant. Surgeons with the scan in their eyeline instead of on a monitor behind them. Factory lines with the spec floating over the part. Accessibility, the use case I will defend in any room you put me in.

Every single person in those rooms had agreed to be there.

Glass shipped in April 2013. More than thirteen years ago. You would assume a comeback on that timeline means the objections got answered.

Plenty did. The chips are smaller, the capability isn’t comparable, and there are real fashion names on the frames now rather than a collaboration bolted on as an apology. I’m not sniffy about engineering wins. I watched people work themselves into the ground for those.

The question of social responsibility hasn’t moved a syllable.

“Wouldn’t it be great if”

Back in 2014, in a room in London, we were doing the thing you do with a new bit of kit. Listing what it might be for. I offered mine up.

Wouldn’t it be great, I said, if you looked at somebody and their information popped up. Their name. Their LinkedIn profile. Something that told you how you knew them.

Anyone who has stood at a drinks thing beaming at a face they cannot place knows exactly what I was reaching for.

The words were barely out of my mouth.

Oh no. That is a terrible idea. You would be pulling up a stranger’s life without them knowing a thing about it, and standing right there I could see every way somebody could turn that on a person.

I said so, out loud. Nobody prompted me. It took about a second.

Having the thought is ordinary. The catching is the rare bit.

The rooms where these things get designed, marketed and signed off do not have anywhere near enough of it.

“Designed for everyone”

Meta’s pair picked up a lot of coverage during the 2026 launch cycle - a lot of it they wish they hadn’t. They also quickly got named ‘Pervert Glasses’ - I doubt that was in the launch plans.

When the company’s founder (Zuckerberg) built FaceMash in October 2003 Harvard shut it down within days and hauled him before its Administrative Board, charged with breaching security, violating copyright and violating individual privacy. He wasn’t expelled. What else, if anything, they did to him has never been made public. His own line at the time: issues about violating people’s privacy don’t seem to be surmountable. I’m not willing to risk insulting anyone.

Twenty-three years on. This summer Meta pushed a mandatory update that permanently kills the camera on any pair where the recording light has been drilled out, after a hobbyist worked out how to desolder it for sixty dollars. Thousands of pairs, bricked. Meta says under 0.1%, a figure picked to sound small and landing, on a user base that size, as anything but. Their VP of AR called it a bit of a cat-and-mouse game.

In the same window, Meta bought billboards. “Designed for everyone, not just the people wearing them,” they read. Underneath: “The camera lets you capture the moment. The light lets everyone around you know when you do.”

The light they were, at that precise moment, watching people drill out.

In late August, Futurism went looking at American schools and found the pattern you’d predict and still hope not to find. Videos taken almost always on Meta’s Ray-Ban AI glasses, posted straight to TikTok and Instagram. Every clip was filmed by a teenage boy, and the behaviour was overwhelmingly aimed at girls. One account had over sixty thousand followers and dozens of videos of girls asking to be left alone.

When people started casting about for a way to defend themselves against a stranger filming them, the best idea the internet came up with was to play Disney music at them. The theory being that a corporation’s copyright gets enforced faster than a person’s face.

It doesn’t work. Enforcement sits with the platforms, not the hardware, and most of that footage lands on Instagram, which is to say on Meta. An attorney called Gene Kang put it to Fortune plainly: if you’re the person being recorded, you don’t own any copyright to the composition.

You own nothing. Mickey Mouse owns something.

That’s the whole architecture in one sentence, and nobody sat down and designed it that way, which is worse.

“Not for recording your lunch”

In June a Chinese startup called Monako launched a pair of smart glasses at $399. Forty-eight grams, its own Linux build, aimed at developers, AI users and researchers. They run Claude Code and Codex. You describe a task out loud, an agent starts building it, and you pin the result to your glasses.

Their own marketing line: “Monako Glass isn’t for recording your lunch; it’s for building your next project.”

Marketed explicitly against recording. Priced for a graduate. Built by a small team who meant every word of it.

It has a camera on it.

What stops them being used in pervy ways? Nothing does. It took me a fortnight to see what that means. This isn’t a Meta problem, or a Zuckerberg problem, or a price problem. A team with entirely different intentions built the identical hole.

Who was in the room

Which leaves one place left to look, and it sits upstream of the marketing, the frames and the regulator. The table. The moment somebody said, and it should take pictures.

Who was sitting there?

This is the heart of it for me, and it’s the argument I make about AI everywhere else.

You need women at the point of inception, through design, through marketing, all the way through to regulation. When we’re absent, this is exactly what gets pushed through without consideration. Not maliciously. Just without anyone in the room flinching.

Before anybody reaches for the men-are-bad reading, that isn’t the argument, and I’m the proof of it. The idea in 2014 was mine. A woman with entirely decent intentions reached straight for the same hole every designer since has reached for. The difference was that I heard what was wrong with it, and I heard it within a second.

A room notices what its people have had to think about.

A room where nobody has spent a lifetime being looked at will not think about being looked at. That is a fact about how specification meetings work, not a moral judgement about men.

My honest test: how fast do you reckon privacy regulation would arrive if the main use of these glasses turned out to be women wearing them to assess men, in every respect?

I think we’d have a directive by Friday.

There’s something that room didn’t know to ask about, and I’ll keep it short.

In March, CNN published a months-long investigation into an international network of men sharing techniques for drugging and assaulting women. Photographs of their wives, traded without their knowledge. Methods, traded. On an ordinary pornographic site hosting more than 20,000 videos tagged as sleep content, which drew around 62 million visits in February alone. Visits to the site as a whole, that figure, not to that material, and not a count of men. Nobody has one.

That’s the world these glasses are launching into. When somebody tells me a hands-free camera on a stranger’s face is neutral hardware, I’d like to know why we think this won’t be part of it.

Beautiful photography is not a privacy policy

The advertising, then. I say this as a woman who spent thirty years making it.

Snap’s Specs are $2,195, shot by Steven Meisel in glorious black and white, with Kaia Gerber, Jack Harlow, HoYeon Jung, Jimmy Butler and Imogen Heap. Beautifully done, cleverly cast, and a hard brief: they’re big, they’re chunky, and two grand buys you what, exactly, that a cheaper pair doesn’t?

What it buys is a tier. An elitist one. Snap’s core audience will never get near that price, and the people modelling them have never had to worry about a stranger filming them, because being looked at is the job.

Ten years at Google taught me the campaign is always ready long before the answer is.

Kevin Hart films a stag do

Popular culture got here before the regulators did, which is usually a sign.

In 72 Hours, the Kevin Hart film Netflix put out this summer, he plays an ad man who blags his way into a stranger’s stag do in Miami and secretly films the lot of them as market research for a vodka campaign. Meta glasses, named on screen.

It works. It gets him exactly what he needs, and it’s very funny to watch.

Then he’s caught. The group turns on him and the whole thing comes apart, and it’s the filming that does it. He never asked them.

That’s the plot. A summer comedy about a stag do built its third act on the asking.

Tim Cook doesn’t give a rat’s

Google goes this autumn, and at I/O this year it split the category in two. Audio glasses that offer spoken help in your ear. Display glasses that show you the information you need, right when you need it. Audio ships first. Display comes later.

Read that running order as an answer. The unsolved part isn’t the waveguide.

Apple is rumoured for the very end of 2027, according to Mark Gurman at Bloomberg, held back until Siri can carry the visual intelligence. Tim Cook saw it coming in 2017. He told the Independent the technology itself didn’t exist to do it in a quality way, then added the most Apple sentence ever constructed. “We don’t give a rat’s about being first, we want to be the best.” Cook, nine years ago, about a product Apple still hasn’t shipped.

The analysts are calling it the next Apple Watch. Apple was late to smartwatches by the better part of a year and defined the category anyway.

This is the part I genuinely can’t call. When those two make these mainstream, and they will, adoption itself moves the privacy conversation. Enough people wearing them could normalise it until asking feels rude. Or enough people could end up on the wrong end of a lens that the demand for rules finally gets loud enough to matter.

I’ve been wrong about which way that goes before.

Three words

I’m not going to wear them. Partly because I can’t see what I’d get that’s worth the trade. Mostly because of the people who never got asked, and standing with them costs me nothing.

On privacy, I don’t think we’ve made peace with anything. We’ve got a better sense of what we’ll hand over and what we want back for it. What we haven’t got is any answer to what we can’t see and can’t control, or an explanation of why, after all this time, it still isn’t a switch you can turn off.

Twelve years ago I stood next to a display case in London and told people how beautiful they were.

So, the small thing. If someone you’re talking to is wearing a pair, ask.

Are you filming?

Three words, no confrontation, and it does two jobs at once. You find out, and they remember that someone can tell. Not everyone understands this technology yet. The people who do can help the rest, and that is roughly all accountability looks like before the law turns up.

And then tell me what I’m missing, because I might be. Is there something obvious here, some benefit I can’t see, that makes all of this worth it?

Nish x

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

13%.

That’s how many of the executive seats in AI companies belong to women.

LinkedIn just ran the numbers on 15,000 companies across 27 countries. Tens of millions of workers. One result, everywhere they looked: the rooms where AI gets decided are 87% men.

The report calls it the triple penalty, which sounds like a darts rule and works like compound interest in reverse. Penalty one: be a woman anywhere near a C-suite. About fifteen points off your share. Penalty two: be in an AI role. Nearly ten more. Penalty three: work at an actual AI company. Another five. Women are 45% of ordinary workers in ordinary jobs. Stack the penalties and you arrive at 13.

Every one of the 27 countries shows the gap. No industry has cracked it. The startups writing the next decade are worse still: at smaller firms, women hold under 12% of the AI executive seats.

The trend line should embarrass everyone. In 2025, women were 26% of new AI hires, against half of everything else. “Member of Technical Staff”, the hot new AI-lab title that has nearly tripled since 2021: 18% women. Head of AI: 20%. The market doubled, the salaries doubled, and women’s share of the hiring went down.

Money makes the plot obvious. In the US, an AI role now pays more than twice a non-AI one. Within AI jobs, every ten extra percentage points of male hires lines up with roughly $45,000 more on the advertised salary. In normal occupations that same gap is worth about $10,000. The second the jobs turned to gold, the doors narrowed.

One detail made me want to lob my laptop into the Thames. The most gender-balanced job in the entire AI economy is Data Annotator. The lowest-paid one. Hourly, project-based, the piecework of the machine age. Women are more than half of those hires.

Women aren’t missing from AI. Women are in the basement of AI, labelling the data, while the penthouse holds its meetings.

“Women just aren’t as interested in tech.” That old song. Bollocks. Women are half the hires in AI’s worst-paid job and 13% of its best-paid ones. Interest was never the problem.

I have seen this film before. Thirty years in this industry, ten of them at Google, and I can recite the plot with my eyes shut. A new technology arrives. Everyone declares that this one will finally level the field. The field tilts, again, in the familiar direction, and everyone looks terribly surprised.

Except the stakes this time are higher than any rerun before it. Earlier tech waves made products and fortunes. This wave makes decisions. Who gets shortlisted. Who gets the loan. Whose scan gets a second look. Who a predictive system flags walking home. Feed it data from a biased world and it will reproduce that bias at industrial scale, politely, forever.

LinkedIn’s own policy lead, Sarah Steinberg, warned that the AI economy “risks being shaped without the perspectives and talent of half the workforce”. Half. Our futures are being written, yet again, by men. Not even a representative sample of men. A very particular guest list.

You can watch the same physics at the crowning ceremony. TIME has named the 100 most influential people in AI every year since 2023. It counted the women just once, in that first year: 41, by its own tally. Nobody has published a number since, so I counted the three lists that followed myself. They land around 30 every time; this year holds 31, with 5 women among the 26 Leaders, the category where the chief executives sit and the money lives. Half the women who do make the list are builders, which tells you women are absolutely building this technology; they just thin out exactly where the pay and the decisions live. One drop, then a flatline. Three years of the loudest boom this industry has ever had, and the number will not budge. (Whitney recommends hyperlinking “I counted” to https://time.com/collection/time100-ai/2026/ — it shows the working.)

Meanwhile, the mood among women is curdling, and honestly, who could blame us. Glassdoor read a year of its own employee reviews: men talk positively about AI at work 45% of the time, women 32%. For Gen Z women it falls to 21%, and the analysts checked, it cannot be explained away by industry or job type. Women are looking at who wins in this economy and drawing the obvious conclusion.

The obvious conclusion is a trap. Opting out of AI is not neutral. Every woman who quietly decides this technology is not for her hands another seat to the men already holding 87% of them.

The other version of this story already exists, and I stood inside it.

OpenAI ran a Women in AI event recently, and the thing that struck me hardest was how fantastic that room was. Brilliant, bright women everywhere: the speakers, and just as much the ones who weren’t speaking.

Head of marketing at openAi -the amazing Leila Woodington - held a conversation with 3 brilliant women all using AI in different ways:
Tara Keeney of Look Who I Found, working with AI to shrink weeks of research prep so the human conversations get the hours.
Marisa Poster of PerfectTed, who made AI fluency mandatory across her company and reaches for it to support her neurodivergent team.
Anouska Ruane of Soho House, automating the admin on purpose to protect the relationship time.

No keynote fantasy anywhere. Just capability, on an ordinary Tuesday, in women’s hands. Rose Gibbons of Linkedin was in that room too, and her write-up afterwards, and sharing the Linkedin in research, is half the reason this essay exists.

One thought followed me home, though.

A room like that is the 1%. Which is wonderful. What about the other 99%?

Everything I’m about to ask of you flows from that question. We need as many women as possible in key leadership roles, building businesses, growing their own, experimenting and playing with AI, because we’re the ones who will make sure it works for everyone, not just the 1%.

So, my plea. Three of them.

If you’re a woman reading this: start this week, exactly where you are. Pick one real task, the report, the research, the presentation you’ve been dreading, and do it with AI beside you. Then say so. Out loud, in the meeting, on your profile. Capability kept secret changes nothing about the guest list. The leadership gap opens at Director level, long before any C-suite, so put your hand up for the AI project now, not when you feel ready. Nobody in the 87% waited until they felt ready.

If you run a company: count your AI hires this quarter. The industry managed 26% women last year; find out whether you did better, and ask why out loud if you didn’t. Grow women into AI roles at Director level rather than parachuting one Chief AI Officer in for the annual report photo. Publish the pay. A $45,000 gap survives on secrecy and dies in daylight.

If you build rooms: build them on purpose. The women above didn’t find each other by accident; someone convened them. Community is not the soft bit of this. It’s infrastructure.

I’m building one myself. The next Switched ON Session is a dinner in London called Don’t Marry the Model, where I’ll show you how to build your own AI operating system. Building mine is the most useful thing I have ever done: a portable system of your own preferences and processes, so you can switch tools whenever you like and never end up wedded to one.

No technical background needed. Bring a laptop and an appetite. If you want to be in a room with other women learning to work with AI, connecting with each other, sharing your worries as well as your wins, then join us.

Save me a seat → click here

Consider it my answer to the 99% question.

AI could be the greatest leveller. So long as we all get the key. All 100% of us.

The future is still a draft. Drafts get rewritten by whoever turns up holding a pen.

Pull up a chair. Bring your sequins.

Tell me the task you’re picking this week. I read every reply.

Nish x

I’m Nishma, and I help people get Switched ON to AI.
Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

According to a machine, I didn’t write this.

Feed my work into Pangram or ZeroGPT and back comes a score announcing, with the calm of a man who has never once been wrong in a meeting, that AI made it. A verdict in four seconds. No appeal.

I should probably be offended. Mostly I’m fascinated, because the verdict tells you almost nothing about me and an enormous amount about the judge.

What actually wrote this

Let me tell you exactly how this piece got made. I talked. I always talk first. My best thinking arrives out loud, pacing the kitchen, dictated into Wispr Flow in one long ramble with the grammar of a woman chasing a thought downhill. Then I worked with AI to shape the ramble into paragraphs. The ideas are mine. The taste is mine. The swearing, very much mine. Machine assisted, not machine made.

I’ve also been on the wrong side of my own rule, so I’ll say that plainly. There was a stretch where I leaned too hard, hit generate once too often, and read back something published under my name thinking: who the hell wrote that? Nobody did. That was the problem. There was no one home in it. I didn’t need a percentage score to catch me. Taste caught me, the way taste catches everything.

More human than human

In Blade Runner it took a whole machine, a trained officer and a hundred questions to tell a replicant from a person. Pangram reckons it can manage it from a paragraph. You’ll remember things rarely ended well for the person on the wrong end of a false result.

The false results are the story here too. Students failed. Careers quietly ended on the say-so of a probability. These tools carry a narrow idea of what human writing looks like, and an awful lot of humans live outside it: people who write through assistive technology, dyslexic and neurodivergent writers, anyone working in a second language, and, quelle surprise, a great many women. “99% likely AI-generated,” says the dashboard, of a paragraph someone dictated word by word through the tech that lets them write at all. Polish reads as fraud. Difference reads as deception.

The test isn’t detecting machines. It’s detecting distance from its own average.

The arms race

Meanwhile, over on the other side of the ledger, pure pantomime. Humaniser tools. Writers A/B-testing their own personalities. Perfectly good sentences roughed up on purpose, em dashes purged like contraband, all so a scanner will believe a person was present.

We are now writing worse, deliberately, to pass a robot’s test of humanity.

Thirty years of the internet and that’s the sentence we’ve arrived at.

The unfashionable rule

I spent ten years at Google watching a version of this film on a loop. Every time the algorithm twitched, a cottage industry of tricks sprang up to charm it: keyword stuffers, link farmers, gamers of every update. Each of them won a season. None of them won a decade. The work that survived the whole era was embarrassingly unfashionable: good writing, made for a specific reader, by someone who understood them. That rule outlived every algorithm Google ever shipped, and it will outlive Pangram too.

A million verdicts

As for the slop the detectors were built to catch: we’d already caught it. Nobody scrolled LinkedIn thinking, gosh, I genuinely can’t tell whether this listicle has a soul. You knew. It didn’t sound like a person, you were irritated, and then, crucially, half of us engaged with it anyway. Shit content is shit content, and it was shit long before AI turned up to industrialise it. Sludge is a demand problem. The feed serves what we feed.

LinkedIn knows this, which is presumably why the feed now comes with a button. Three dots on any post and you can file your verdict: “seems like AI slop”. No machine, no trained officer, no hundred questions. One tap. Over a million people have pressed it in its first three weeks, by LinkedIn’s own count. Fewer than I’d have guessed, honestly, given the state of the average scroll. Still: a million verdicts.

Did anyone in that product meeting believe it would only ever be pressed in good faith? Buttons like this get pressed on competitors. They get pressed out of spite at 11pm. They get pressed, I’d put money on it, on women, because I’ve seen what already lands in the comments under women’s posts. I’ve been told more than once, always by a man I’ve never met, that I didn’t know what I actually wanted to say. He did, apparently. He always does. Imagine handing him a button. LinkedIn says there are safeguards, that no single press decides anything, and I believe them, because you only build safeguards for a thing you know people will do. Spare a thought for the wrong presses that land anyway. Somewhere right now a person is losing sleep, or losing clients, or worse, over a verdict a stranger reached before finishing her first paragraph.

Are we just judging style over substance now?

My plea

Starve it instead of scoring it: don’t click it, don’t rage-comment on it. Judge writing the way you judge company: did it teach you something, did it make you feel something, is anybody home? If you’d like your writing to read more human, skip the humaniser tool and go the long way round. Say it out loud first. Take your brain somewhere interesting and let it react to things. Work with AI as an extension of your brain, never the author. A detector can’t score any of that. Readers can, and they’re the only test that was ever worth passing.

And if this essay comes back 100% AI? Good on AI. It’s never written anything this good in its life.

One question before you go, for the next time you’re tempted to sand a sentence down to pass: how much of your own voice would you trade to satisfy a test built by someone who has never read you?

Nish x

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

Two days on, the strangest thing has happened to my corner of the internet. LinkedIn feels like Dollywood. People who normally post about pipeline and quarterly targets are sharing her one-liners. Her quotes are everywhere, her songs are everywhere, and my feed, a place I usually endure rather than enjoy, has become the most joyous room on the internet. Even her exit is generous.

I’m not going to pretend to be measured about any of this. I’m devastated, and I’ve been a little surprised by how much.

The quotes are being shared because they hold up. She spent sixty years saying obvious things so fearlessly that they stopped sounding obvious. When the world mocked the wigs, the nails, the body, she got to the joke first:

“It costs a lot of money to look this cheap.” She made the punchline before anyone else could, then banked the difference.

Underneath the sequins sat one of the sharpest operators music has ever produced.

In 1974 Elvis Presley wanted to record “I Will Always Love You”. Dolly was thrilled. Of course she was. This was Elvis. Then, the night before the session, his manager called with the terms: Elvis didn’t record a song without taking at least half the publishing.

She cried all night. In the morning, she said no.

Not to Elvis. To the terms.

Eighteen years later, Whitney Houston sang that song into one of the biggest records ever made, and every royalty flowed home to a girl from a one-room cabin in the Smoky Mountains. One of twelve children. Left for Nashville the day after she finished high school, carrying a cardboard suitcase. The song still belonged to her.

Attention is lovely. Ownership is power.

You could call it luck that the song came back around. Except Whitney could only make Dolly rich because Dolly had refused, eighteen years earlier, to make Elvis richer. She wrote her own insurance policy in the same writing session she wrote “Jolene”, and it paid out for the rest of her life.

I doubt she’d ever have called herself a businesswoman. She simply understood her worth, priced it accordingly, and smiled while she did it.

What she did with the winnings is what makes her a legend. The Imagination Library has put more than 300 million books into the hands of children who, like her, might otherwise never have owned one. A $1 million cheque to Vanderbilt in 2020 helped fund the research behind Moderna’s early vaccine trial. This week Eminem shared the letter she wrote him after his Rock & Roll Hall of Fame induction, telling him rap wasn’t her world but she could tell he was the very best at it. She had nothing to gain from that letter. That was rather the point of her. She collaborated her whole career, Kenny Rogers to Beyoncé’s Cowboy Carter, because she knew joining up beats going it alone when you’re not scared of sharing the credit.

She never slowed down and she never went into the shadows. She defied every piece of nonsense ever written about age-appropriate dressing, and she did it while an entire industry waited for her to become a nostalgia act. What makes me saddest is that even at 80, having dominated the world with her talent, some of it was still a battle. She was still fielding questions about her body, her look, her right to be exactly herself.

She just never let us see the cost. She’s one of the people I credit for my own love of sequins, and for the nerve to become more myself as I’ve got older, not less.

It matters this week more than most. We’re living through AI bots screening women out of interviews, ageism dressed up as culture fit, and a technology industry of identically dressed men explaining the future in beige. Dolly wore exactly what she wanted, built more than most of them ever will, and stayed kind the entire time.

Let’s all be more Dolly, please.

If you want to honour her with more than a repost, steal three of her decisions.

Own the asset. Keep the thing you made, especially when someone famous asks nicely for half of it.

Build the world around it. She didn’t license her name to somebody else’s theme park. She became an owner, in the county where she grew up with nothing, and thousands of jobs grew around her.

Make success useful. The songs funded the park. The park funded the books. The fame carried the vaccine cheque. Nothing she built ever sat in a trophy cabinet.

She wasn’t an icon to me. Icons sit still. She was a hero, and heroes leave you instructions.

So I’m picking up the mantle the only way I know how. Sprucing up my sequins. Shining a bit brighter. Encouraging every brilliant woman I know to do the same.

Rest in sequins, Dolly. The world’s lost a lot of sparkle this week. Keeping the sequins moving is our job now.

If this made you think of a woman who’s been told she’s too much, forward it to her.

Nish x 🪩

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

My keyboard and I are taking a break.

It’s not you, keys.

It’s me.

Mostly because my hand stopped cooperating.

I don’t write for a living. I communicate for one.

I think in writing, definitely. Which means I spend a slightly alarming amount of time at a screen, with my fingers doing the heavy lifting and my eyes quietly filing a formal complaint (2nd cataract op on its way and no, I’m not that old!).

Then I developed RSI.

Typing became painful. Not dramatically cinematic pain. Just the sort that creeps into everything and makes you realise how much of your working life depends on repeatedly moving the same tiny collection of muscles.

That’s how I found Wispr Flow.

I wasn’t looking for the next shiny AI toy. I needed to solve a real problem.

Which is often when technology is at its best.

Now I dictate walking the dog.

In the kitchen, with both hands in a bowl of something.

Moving around the house.

At one in the morning, whispering, because I’m still working and I don’t want to wake anyone.

I send longer WhatsApps than anyone asked for. I have probably dictated several Substack essays into people’s inboxes by accident.

Sorry.

And with its ‘Notetaker’ my meetings don’t evaporate any more either. It listens, I talk, the notes land in Claude, and an agent I trained turns them into the actions I used to lose somewhere between the tube and the front door.

But the thing I love most is not the speed.

It’s that I can get the thought out before I’ve edited it to death.

I can talk in a messy stream of consciousness, correct myself halfway through a sentence, change direction, have a little moan, and then look at what I actually meant.

My process now:

AI at the beginning.

My thinking in the middle.

AI at the end.

That middle bit matters.

It’s still me thinking. Me making the connection. Me having the idea. The technology just gives it somewhere to go without demanding a polished first draft out of aching hands.

I’d go further than I used to.

I think talking is the only way to get the best out of AI.

Typing is editing. You can’t type a sentence without half of you tidying it as it goes, picking the respectable word over the true one. By the third line you’re performing.

Which means what you get back is capped by how much of your real thinking you were willing to hand over.

Typing quietly withholds most of it.

Nobody types a moan.

“I already talk to my AI,” people say. They’ve got the voice thing, they press the button.

It isn’t the same, and the difference is all in the engineering.

Wispr clears the wreckage out of speech while you’re still making it. The ums. The ahs. The “no, hang on, not that”. It handles the mid-sentence correction, which I do constantly, without sending me back to clear up after myself.

It copes with accents it wasn’t raised on. With dialects. With mumbling. With a woman whispering at her laptop in the dark.

The gap between a voice button and this is the gap between being recorded and being understood.

And this is where so many conversations about AI stay weirdly narrow.

The use cases are presented as though this is mainly for young tech-bro entrepreneurs building a company before breakfast.

Fine.

But some of the most important ones are far less shiny.

Parents with a baby in one hand and a business in the other.

People with cerebral palsy, or anything else that affects fine motor control.

People whose speech doesn’t fit the assumptions built into traditional voice tools.

People recovering from repetitive strain.

People who have ideas, stories and work to share, but for whom a keyboard is difficult, painful, or simply a locked door.

A friend told me about her son a few weeks ago.

He has cerebral palsy. Fine motor control is the thing it takes from him, so typing was never really on offer.

He’s been working with Wispr Flow.

He writes now. He tells stories. He sends his work the way everyone else sends theirs.

She mentioned it in passing, the way you mention a good week.

I’ve thought about very little else since.

I often think about what this could have meant when I had my twins.

The hours and hours awake in the middle of the night, feeding babies, trying not to fall asleep sitting up, discovering that 3.17am gives you an extraordinary amount of time to think.

If I’d had something like this then, perhaps I could have got some of those thoughts out.

Not because parents need another way to be productive at three in the morning.

Because having another way to communicate changes what is possible.

That’s the bit I wish we talked about more.

Technology at its best is NOT always about doing more.

Sometimes it’s about making something possible for someone the design was quietly locking out.

So I’m delighted that Wispr has just raised $280 million at a $2 billion valuation.

Delighted is not a word I trot out about funding rounds.

Money like that means they keep building, and I hope they spend it making this work for more people, in more places, across more accents, speech patterns, bodies and lives.

The world does not need more tools designed around one narrow idea of a “typical” user.

Congratulations to the team.

You’ve solved a very real problem for me.

Now please go and solve it for ten times as many people.

And if the alternative is another billionaire building a bigger rocket for reasons best discussed with a therapist, I know where I’d rather see the money go.

Nish x

I’m Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

9am, 1991. The future I had planned was gone.

I can still feel the paper in my hands.

That little slip, folded once, handed over in a school hall. The letters on it were the wrong letters.

Not catastrophically wrong. Just wrong enough. Wrong enough to take L.S.E. off the table and put my whole plan in the bin before the day had properly begun.

I had no idea what would unfurl from that morning. All I knew was that the route I had carefully mapped out had disappeared.

Life was pretty tough at the time. There was a lot going on, and I was very much on my own with it. Big decisions to make, nobody to make them with.

But I still had one certainty.

Get the degree. Get the job. Build the life.

That was the bargain.

As the first in my family to try for a degree, I was not about to walk away from it. So I swallowed my pride and embraced the hell that was clearing.

If you have only ever known clearing as a website, let me paint it for you.

It was a landline. A busy tone. A list in a newspaper and a pen. It was hours of waiting and confusion, ringing back and being told to ring back again.

I could not move away from London for a number of reasons, which narrowed the pool to something close to a puddle.

Somehow, I found another route.

I had no idea where it would take me. But I believed there would be a job at the end of it and a chance to build something from there.

The 18-year-olds collecting their results today are being asked to make the same enormous decisions without anything like that certainty.

Is university worth the debt? Will there be a job? Will they learn anything relevant, or will the world of work change again before they graduate?

That is a hell of a lot to hold at eighteen.

And the noise surrounding them is deafening.

The Noise

Start with the numbers, because they are worse than the vibes.

More than a million young people in the UK are NEETS - not in work, education or training. Graduate hiring is forecast to fall again this year. University costs more than ever. You will be in debt. There may be no job at the end of it. AI has eaten the entry level. Pick a headline, any headline.

Then, because the universe has a sense of humour, they swipe past all of that into their social media feeds and an endless carousel of bros and hot gals working one day a week from a balcony in Dubai. Apparently the real move is drop-shipping. Or crypto. Or influencing. Or selling a course on personal branding. Or setting up as an AI consultant.

Those last two sting. I am aware.

So let me be clear. I am not selling anyone an exit. I did the thirty years. I work harder now than I did inside the job. I do not promise new fortunes, and you should not believe me if I did.

What I care about is the boring bit: what this technology actually does on a Wednesday afternoon, when nobody is filming.

These young people have done two years of graft. EPQs. Mocks in January. Tube strikes and lifts across town. Then they stand in a school hall and get asked, before breakfast: is any of this worth it?

Is it worth the debt? Will there be a job? Will I learn anything relevant, or is it three years of fun and a very large bar tab?

That is a hell of a lot to hold at eighteen.

Everything is taking longer, too. Longer to buy a house. Longer to feel settled. Longer to reach the bit that used to be called security.

The runway got extended and nobody told them.

What I saw on my recent university campus tour

I have spent a lot of the last year traipsing around universities with my twins. Sticky lecture theatres, terrible coffee, brilliant buildings.

We should be genuinely proud of what we have in this country. The institutions, the campuses, the depth of teaching and research. It is one of the world’s great education systems and we often talk about it like a broken boiler.

But every parent in every room was asking some version of the same question.

What is the employability on this course? What are the graduate outcomes? Is this worth it?

And then there was the lecturer who opened his talk with, and I promise I am not making this up:

“Well, it doesn’t really matter what I tell you, because there aren’t any jobs anyway.”

Bold way to sell further education to a room of teenagers and their already twitching parents.

I sat there thinking: it does not matter what the graduate outcome has been. None of us knows exactly what it will be.

What university is for

I am not waving a flag for either side. University or not university is not the interesting question, and anyone selling you a perfect answer is probably selling you something else as well.

A degree is not only a transaction for a job. Those years can also give a young person freedom, independence, challenge, friendship and the chance to work out who they are away from everyone who has always known them.

That matters.

But university is not the only place to build a life or learn how to live one. The same growth can happen in a job, an apprenticeship, a community, a business or behind a counter at McDonald’s, which is where I started.

The question is not simply whether university is worth it.

It is whether we are creating enough places for young people to learn, contribute, make mistakes and become useful.

The bit aimed at us

My real argument today is with employers.

We are quietly removing the bottom layer of our businesses, the layer where young people used to be trained, and calling it efficiency.

Where do we think future talent comes from? Where is the pipeline?

You cannot delete the first rung of the ladder and then act surprised when nobody can climb.

For every junior role we remove, what are we reinvesting? Anything? Or are we banking the saving and hoping the problem belongs to somebody else’s balance sheet?

Research published this year found that more than a third of students never hear back from most of the jobs they apply for. Not rejected. Ignored.
This is a disgrace.

I cannot imagine what that does to a young person’s confidence. Hundreds of applications fired into silence, ghosted by companies that will happily post about their values in October.

We have a responsibility here, and we are ducking it.

Yes, we need to talk seriously about how AI is reshaping work. Let us also stop hiding behind it. AI is not to blame for every job that has gone. Some of those decisions were made by people, in rooms, with spreadsheets.

Efficiency that removes opportunity and pockets the saving is not transformation. It is cost-cutting.

The more I work with AI, the more convinced I become that the real opportunity is not simply the time it can save. It is the growth it can help us create.

New services. Better businesses. More ambitious ideas. And, if we choose, new places for people to begin.

That choice belongs to us.

What I told my children

My two will be off to university next year, if they get their grades and if that is what they choose.

I encouraged them to pick their courses out of passion and interest. Not because I can promise a job at the end of it. I cannot, and neither can anyone else.

What I do know is that the most employable people I have watched across thirty years built capabilities that still matter when everything else changes:

• Communication

• Empathy

• Critical thinking

• Judgement

• Curiosity

• Adaptability

• The ability to make something happen with other people

Those can be built at university, in a job, through an apprenticeship or almost anywhere somebody is given a real chance to practise them.

One more thing about young people and AI

We have collectively and lazily decided that young people are AI natives and therefore fine.

They are not fine.

I meet young people who claim to know exactly what they are doing and then, five minutes in, clearly do not. Nobody has taught them the basics, the infrastructure or how to think critically about any of it. They are picking it up in fragments.

Worse, they assume everybody else already knows, so they never ask.

Which is exactly what I hear from people at the other end of the age scale, convinced they are hopelessly behind the handful racing ahead.

Same fear. Different decade.

The truth is that we are all learning. Quite a nice place to meet one another.

So, my ask

Celebrate today. Genuinely. Whatever is on that slip, celebrate the two years behind it.

Then let us stop exhausting ourselves with the same argument about whether further education is worth it and put some of that energy into creating opportunities.

If you can bring a young person into your business, let them shadow you. Let them learn from you. Be prepared to learn from them.

If you run a smaller business and have been telling yourself you cannot afford to train someone, read this bit twice. For apprenticeships beginning from this month, the government funds the training and assessment costs, up to the funding-band maximum, for apprentices under 25 at employers that do not pay the apprenticeship levy. From October, eligible employers can also receive a £2,000 hiring payment for taking on a new apprentice aged 16 to 24.

I am not the Treasury’s press office. I am just quietly removing your excuse.

And if all you can offer is a conversation, a callback or two minutes of actual feedback instead of silence, do that. It costs almost nothing and may be worth far more than you think.

Open a door. Offer an hour. Reply to the application. Hire the apprentice.

Just do not tell young people to climb a ladder whose first rung we quietly removed.

What is one thing you could actually do this month to open a door?

Nish x 🪩

P.S. I want to create more rooms where 18-year-olds and 58-year-olds get Switched ON together, because I increasingly think we are each holding half the map.
If you would come, contribute or send someone, tell me.

Two final things …

1. The experiences, ideas and opinions in this essay are mine. I dictated my initial thoughts using Wispr Flow, then used ChatGPT and Claude to help me organise and edit them. I reviewed, challenged and rewrote the finished piece. Every word published is one I stand behind. Human until proven otherwise. 💕

2. Here are my sources referred to in writing this piece:

Office for National Statistics: Young people not in education, employment or training, May 2026

https://www.ons.gov.uk/employmentandlabourmarket/peoplenotinwork/unemployment/bulletins/youngpeoplenotineducationemploymentortrainingneet/latest

Institute of Student Employers: Student Recruitment Survey 2025 findings

https://ise.org.uk/knowledge/insights/498/5-trends-you-need-to-know-from-ises-recruitment-survey-2025/

The Price of Success report: graduate application ghosting

https://yugo.com/resource/blob/843476/4c5a600caf4db922ef0334c5a23b2299/report-1-data-data.pdf

Apprenticeship funding rules 2026 to 2027

https://assets.publishing.service.gov.uk/media/6a3406aa00ff27f06e3efa78/apprenticeship_funding_rules_2026_2027.pdf

You will remember Carol Beer. Little Britain, early noughties: the receptionist who took your perfectly reasonable request, tapped it into a terminal with one finger, waited a beat, and delivered the verdict. Computer says no. Then she coughed on you. It was funny because it was preposterous. Nobody could really be turned down by a machine that nobody could question and nobody would explain.

It is 2026, and Carol Beer has been promoted. She screens the applications now. She gets through several thousand an hour, she has never once coughed, and she has decided the problem with you is that you are 52.

The government, meanwhile, would rather you didn't do this bit of maths.

The state pension age is going up. As the FT laid out in this article last week, if the planned rise from 67 to 68 is brought forward, around five million of us aged between 49 and 55 will work an extra year before we are allowed to stop. Meanwhile, out in the actual labour market, the age at which you quietly become unhireable is falling to meet it. Somewhere around fifty, by the account of the people living it.

Two lines crossing. We are standing at the intersection.

So there it is. Work until 68. Unemployable at 52. Mind the sixteen-year gap.

That is not a skills shortage. It is a con with two authors, and we are the ones expected to make the numbers work.

The woman who deleted twenty years

I have spent this week watching and reading content from Stacey Duguid, a former fashion editor who went public about being, in her word, unemployable in midlife. She is 52. Over sixteen months she has sent hundreds of applications, most of them into a silence so total it has its own weather. She wrote something that has not left me:
“The money’s running out, and I’m frightened.” Then the line that turns a private fear into all of ours: if I’m in this position, maybe you are too.

instagram.com/staceyduguid

She started getting messages. Loads of them, from people convinced that AI-driven recruitment was screening them out before a human ever saw their name. So she did what the crowd advised. She took her age off her CV. She chopped off twenty years of experience to get past the machine.
The FT’s Claer Barrett, who reported all this, called it her “newly botoxed CV”, which is the funniest and bleakest phrase I have read all month.

Twenty years. Deleted. Two decades of getting good at something, filed under things we don't mention in front of the algorithm.

We have built a system whose cure for being good at something for a long time is to pretend you weren’t.

The machines are talking to each other

Nobody reads these applications. Not one person.

An AI writes the application. A different AI screens it. Two machines passing a memo about somebody’s life, neither of them bothered to open it. Somewhere underneath sits a person with a mortgage, a school run and thirty years of knowing exactly how this industry works, reduced to a file that gets sorted into keep or bin by software with the judgement of a bouncer and none of the eye contact.

The algorithm didn’t invent ageism, mind. It inherited it, from the humans who always quietly assumed that fifty is where useful ends.

We didn’t automate the bias away. We automated it up to scale, and gave it plausible deniability.

Where I’ll argue with myself

This is where I have to complicate my own optimism, because I spend a lot of my time telling midlife women that AI is a superpower, that you can build a company from your kitchen table with a team of agents, that this is the great equaliser. I believe all of that. I have watched it happen.

I am also not going to stand here and pretend the boom in midlife self-employment is all liberation. According to the ONS, 48 per cent of the self-employed in this country are now over 50, the highest proportion in a decade. Some of that is freedom. A lot of it is redundancy with better branding.

“Portfolio career” is, for a great many people, just the polite term for nobody will hire me.

Being pushed out of a salaried job and told to go and be your own boss is not empowerment if the alternative was a P45. One number should stop the room. Research from the non-profit Brave Starts found that 1 in 3 of its midlife members would accept a salary as low as £25,000 to retrain into something new. That is what a career’s worth of expertise is being asked to sell itself back into the market for. It is worth more, obviously. The door is simply shut, and £25,000 is the price of the window.

AI is standing on both sides of the door

So which is it. Is AI the thing screening midlife out of work, or the thing letting midlife build the way back in?

It is both, at once, and I refuse to pretend otherwise, because that contradiction is exactly where I live. The same technology that filters your CV into the bin is the only thing fast enough and cheap enough to let you build the business that makes the bin irrelevant.

The tool that locks the door also picks the lock.

The lock is not inevitable, whatever anyone tells you. Someone chose the training data. Someone set the filter. Someone decided that twenty years of experience should read to a machine as a risk rather than a reason.

A note for the people building the filter

If you are building recruitment AI, or the models underneath it, I would like you to hear this as clearly as I can say it.

You are not building a filter. You are building a wall.

Walls have a habit, though: you are going to be standing on the wrong side of this one, sooner than you think. Every system that treats 50+ as expired is a bet against your own future self, placed by someone too young to believe they will ever get old. You will. The wall you are training today is the one that will not answer your emails in 2045.

Build the thing that reads a whole life and sees what it’s worth. You have the data. You have, allegedly, the intelligence. What you are missing is the imagination to picture yourself on the wrong side of your own algorithm, which, given the average age in your building, I do understand.

And for the rest of us

If you are reading this at 11pm with the laptop on your knees and a rejection sitting open in another tab, I know. I am not going to tell you it’s fine, because it isn’t, and you would rightly stop reading. I will tell you the three things I have watched actually work.

One, never let an online form be your only way in.
Keep applying if you need to, but know what you are up against: the front door has a machine on it, and it has been trained to sort you out. The side doors are people. A colleague from two jobs ago. A client you looked after properly. Someone whose kid plays football with yours. Most roles at our age go to somebody who was already recommended, and the recommendation is the one part of hiring nobody has managed to automate. Thirty years of knowing people is not a nostalgia trip. It is the asset.

Then tell them you are looking. Out loud, specifically, not a coy little line in your bio. Shame is what keeps this quiet, and quiet is exactly what stops anyone putting your name forward.

Two, do not delete yourself to get through the door.
What Stacey did was rational and it should never have been necessary, and there is a version that costs you less. Put the last fifteen years in full and squash everything before it into a single line. Take the graduation year off. Use the words the advert itself uses, because keyword matching is genuinely what a lot of these systems score you on. Lose the columns and the clever layout while you are at it, because the parser cannot read them and will mangle you quietly. You are not dumbing your career down. You are formatting it for a reader with no imagination.

Three, use the very thing being used against you.
Learn to work with AI like it is the team you cannot yet afford to hire, because that is exactly what it is. The most quietly satisfying thing I know is watching someone who was screened out of an industry use its own tools to build the door they were refused.

This is the whole reason I do what I do: the fightback works, and now there is proof. When researchers ran a hiring experiment this year with 1,725 recruiters across Britain, the US and Germany, adding AI skills to a CV made a candidate markedly more likely to be shortlisted. The bit that made me sit up: those AI skills partially or fully cancelled out the penalty for being older. The exact thing the machine docks you for, wiped by the exact thing it says it wants. Employers are voting with it too. In Microsoft’s own research, 7 in 10 leaders say they would rather hire a less experienced person with AI skills than a more experienced one without. Read that twice if you’re the more experienced one.

So the skill the algorithm punishes you for lacking is the most closeable disadvantage I have ever seen. We are not behind. We are one course, one habit, one honest afternoon of practice from being exactly what the market says it wants. Closing that gap, for the people the industry keeps writing off, is the entire reason Switched ON exists.

We are going to spend 30+ more years working, whether the pension lets us stop or not. The question was never whether we are too old for this. It is who is brave enough to build for the people the machine keeps saying no to. I know who I’m backing.

It won’t be a computer. It’ll be a 52-year-old with nothing left to prove and everything left to build.

Nish x

I’m Nishma, and I help people get Switched ON to AI.
Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you’re midlife, brilliant, and tired of being told you’ve missed it, you’re exactly who I do this for. Come and find out more at nishma.co.

It is not midlife that is struggling to stay relevant. It’s the technology.

This is not a piece about ageism, though ageism is the water it swims in. It is about something more damning. The people building AI have stopped watching. The industry studies generations for a living, and Gen X is nearly a third of the population and four per cent of that research.

Four per cent.

Age moved on. Relevance moved on. The tech didn’t.

We have rewritten what fifty looks like, and nobody sent tech the memo. A woman in her 50s is now Jennifer Lopez, or Victoria Beckham, or half the names you would be startled to learn are the age they are. The over-50s are back on the Ibiza dance floor, out-dancing their own children (yep that’s me!).
My own fifties look nothing like my mother’s did. Midlife is richer than it has ever been, living longer than it has ever lived, and holds the spending power, the networks and the decades of pattern recognition, the exact wisdom you would want teaching these machines to be less stupid. The culture worked all this out years ago. The industry is the last one in the room to notice.

This is where it stops being a marketing quibble. If we were talking about a handbag advert, fine, cast whoever shifts the units. But AI is not a handbag. It is a product rewiring how we work, how we run businesses, how we care for our parents, how we hold on to our jobs. Who it pictures using it decides who dares to use it. You do not lose a generation by barring the door. You lose them by never once casting them in the story. And this generation, the squeezed, caring, business-building, quietly furious midlife one, and especially its women, is the last one you would want to lose. They are the ones driving the change.

The room

I watched all of this play out in June, in the loveliest possible setting. It was my first ever B2B creator lunch, in Cannes, hosted by OpenAI, and I don’t think there have been many of those before. A brilliant room. A gorgeous terrace, mostly women, wildly clever, and I will happily admit I was a little starstruck, because several of the people there had been my actual teachers: the creators whose work first showed me what these tools could do. Marina Mogilko, the Silicon Valley Girl herself. Tara Keeney. Maggie Ma. Monica Khan. Jess Ramos. Jean Kang. They had earned their place on that terrace. So, it turned out, had I.

I want to name who got me there, because it matters, and because she has earned it. Leila Woodington put that lunch together, and I have watched her blaze a trail for years: from the day I first met her at Meta, through Snap, through VEED, and now OpenAI. Every move, she has quietly redefined how this industry markets itself, warmer, sharper, more human. She is one of those rare people in tech who simply gets it, does beautiful work, and is glad to go against the grain to do it. Rooms like that one do not open by accident. It is the deliberate championing of people like Leila that gets someone like me through the door, and slowly widens who the whole industry can see. It was, on her part, a clever door to open, because AI runs on trust, and trust travels through people you already believe.

Standing there, I realised I was just about the only midlife creator in the room. It didn’t land as a slight, though. It landed as a job. If I am the only one of us in a room like that, then someone has to carry what happens in it back to the millions who will never be invited in. Someone has to translate it. That someone might as well be me.

I know this game from the inside, by the way. I spent ten years at Google, and we were forever obsessed with Gen Z, sometimes to the point of parody. It was gospel: win them young and you win adoption, coolness, scale, a customer for life. I understand the logic completely. I would just love to watch it grow up. You can see the blind spot most plainly in the advertising. Watch the AI ads. Clock the casting. The demo films, the launch films, the partnerships: everyone is 23, luminous, building an app in a dorm room. Sometimes younger. Even the Advertising Standards Authority has noticed, 44 per cent of us reckon older people barely appear in advertising at all, and that is before anyone points a camera at an AI launch. There is something faintly desperate about the most powerful tool of our age trying quite this hard to be cool.

Tech treats midlife like a ghost demographic: happy to bank our subscriptions, wary of putting our faces on the brand in case the software starts to look its age.

Which is how a woman in her 50s who is evangelical about AI still gets filed as a surprise. A charming exception. The plot twist, never the lead.

There are midlife creators all over tech doing this brilliantly, and brands getting it right, showing a real spread of ages and lives and uses. I am not pretending otherwise. But the norm, the thing I hear from my community and my clients every single week, is a quieter feeling: that this new world is something they are intruding on, not something they were invited into. That feeling decides who steps forward and who quietly backs away.

The midlife people I meet tend to split into two camps. The senior men who say “oh, my team handles that” (the corporate equivalent of claiming you’ve read Ulysses), then corner me by the lifts to ask for actual help. And the women, most of them, who say “I’m so far behind”, which is only “too late” in a nicer coat. Both are bluffing. Both are wrong.

We’re already here

The truth is we are already here. Look past the headline that more Gen Z have poked at a chatbot, and at the numbers that count. At work, it is millennials, not Gen Z, who reach for AI the most. Older users are folding it into the tools they already live in, without making a personality of it. The creator economy everyone still pictures as a teenager doing a dance? Gen X and older are its fastest-growing part.

We did not miss the revolution. We just didn’t film ourselves having it.

Now the half I actually want to celebrate. Gen X is the most quietly adaptable generation alive. We grew up analogue and taught ourselves digital with no manual and no helpful teenager. We are the last who remember the world before the internet and the first who had to build a life on top of it. We rewired for email, then the web, then mobile, then social, and now this, every single time without one advert made for us, and kept our jobs, our families and most of our marbles throughout. That is not a generation struggling to keep up. That is the deepest muscle memory for change on the planet.

Staying power is the whole point of us. We just don’t shout about it.

The numbers, and the Fuck It

Put the numbers next to the temperament. The average founder of a genuinely fast-growing company is 45, and a fifty-year-old is likelier to build one than a thirty-year-old, not less. Women are starting businesses in record numbers, a great many of them over 50. Midlife holds more of the wealth, lives longer, and carries the single most underrated force in commerce: a woman in her 50s with a fully operational Fuck It attitude.

This is where AI stops being a toy and starts being a lever. Venture capital still hands all-female founding teams around 2% of the money, which is less a gap than a rounding error with a God complex.
Can’t raise for a team? Build one out of AI agents.
Told your idea won’t fly? Prototype it by Friday and put it on the table, working.
Can’t buy your time back? Work with AI daily until it runs the parts of your life that were never the best of you.
That is what I watch happen, women in particular building the things that should already exist, because the people in charge never had the problem, and nobody funds an ache they have never felt.

Why I call myself a Midlife Creator

This is the work I am proudest of. Every leap in technology has needed people who can translate it, distil it, make it relatable, bring it to life for everyone standing outside it. That is the job I have handed myself. I call myself a Midlife Creator, and I mean it as a flag, not an apology: a representative, a protagonist, the one asking the daft, curious questions out loud. Not another prompt list, but the confidence to treat AI as a sparring partner and the craft to make it sound like you and not like everyone. The people written off as too far behind turn out to be the fastest, because they know exactly what they want it for. Fluency without a point is just noise. These people have a point.

Not a plea. A business case.

So this is not a plea. It’s a business case, aimed up, at big tech and whoever is writing the casting briefs. The best already understand it, and Leila and the rooms she builds are proof. I’m talking to the wider machine. Broaden who you picture. Not as a favour, not as a hashtag, and not because it would be a nice thing to do. Because you are leaving the wealthiest, longest-living, fastest-growing customer base in the market sitting outside your shop window, and because you are building the tools the whole of society now runs on.

There is a colder, cleverer reason too. A model is only ever as rich as the range of people teaching it. Feed it a narrow band of young, similar voices asking similar questions and it learns a narrow band of the world, then hands it back to you, averaged and beige. Bring in older, wider, stranger voices, the odd questions, the lived-in context, the instincts built over a whole life, and you get a better model. Fuller. Stranger. More useful.

The range of people teaching it is not the nice bit on the side of the intelligence. It is the intelligence.

So ignore midlife women at your peril: your fastest adopters, your most underrated founders, your richest untapped market, and the sharpest voices you could put in the room where the model is trained. That is not a box to tick. That is the growth story you keep saying you want.

And the useful part, for you. Somebody is already translating you to that audience. Creators like me are doing it every week, mostly unpaid, because somebody has to. The audience I speak to trusts me and does not yet trust you.

You cannot buy that with a launch film. You can only borrow it from someone who has already earned it.

The bit I’ll say plainly

If you are the one sitting there certain you’re too far behind: I see you. I was you once, nodding along in a meeting, quietly drafting my exit. You are not behind. You are early.

Being written off does not feel like being ignored. It feels like being erased. It is watching an advert for a tool you use every single day and clocking that nobody in it is you. It is being Tipp-Exed out of the story a chapter early, by an industry that has quietly decided the good part of your life already happened. Well. I’ve done the maths on that too. I’m only halfway there.

This is the middle of the book, not the acknowledgements, and the back half is where it gets good.

It’s not too late. It’s barely gone breakfast.

So save me a seat at the next lunch. I’ll be the one raising the average age and bringing the disco. 🪩

Nish x

I'm Nishma, and I help people get Switched ON to AI. Less overwhelm, more time and freedom, and a brand that stays unmistakably yours. If you're midlife, brilliant, and tired of being told you've missed it, you're exactly who I do this for. Come and find me at nishma.co.

“Right, quiet now. Nishma, that means you.” I heard that line roughly nine thousand times before the age of eleven. Every school report carried the same verdict in the same tired teacher biro: a pleasure to have in class, if only she’d stop talking.

‘She’ did not stop.

Ask any of my friends. I am famous, genuinely famous, for the long story. For the WhatsApp voice note that lands like a surprise feature-length podcast, complete with tangents, a full cast of characters and a dramatic pause nobody asked for. I have never once used ten words where four hundred would do. And for thirty years, that was the thing about me people learned to tolerate. Fondly, mostly. But tolerate.

Then something funny happened. The exact trait I was told to rein in became the most useful skill I own. Turns out the age of working with AI belongs to the talkers. Funny how it goes.

Everyone is typing at their AI like it owes them money

Short. Clipped. “Write a follow-up email.” “Make it shorter.” “More professional.” We treat the most capable thinking partner most of us have ever had like a moody vending machine. Coin in, snack out, thump it when it jams.

I did this for months. Then my wrists staged a quiet protest, I found a tool that types while I talk, and I stopped typing at my AI. Now I just talk to it. Same AI. Same me. The answers got better overnight. The only thing that changed was that I opened my mouth.

But the AI part is only half of it. Because I don’t just talk to brief the machine. I talk to catch my own brain before it wanders off.

My best ideas never arrive at my desk

They arrive on a dog walk. Milly’s nose is in a hedge, I’m three fields from a keyboard, and the whole shape of a talk or an essay lands in one go, fully formed, glorious, and gone in ninety seconds if I don’t grab it.

So now I grab it. Out loud, into my phone, mid-walk. The newsletter intro. The reply I’ve been dodging for a week. The story I want to tell on stage. The message to a friend that would otherwise have died in my drafts. All of it, spoken while it’s still hot, before the moment cools and I lose the good bit.

This isn’t a dictation trick for people too lazy to type. It’s a net for the ideas that used to get away. The ones that came at the wrong time, in the wrong place, with both hands full of dog lead.

Typing is performance. Talking is thinking.

There’s a reason talking catches what typing kills.

When you type, you edit. You can’t help it. The cursor blinks and some prim little sub-editor wakes up in your head and starts tidying before a single word reaches the page. You delete the messy bit. You cut the story. You trim the context to something that looks efficient and reads like a memo. By the time you hit enter, you’ve handed over a headline and binned the article.

When you talk, the sub-editor goes for a lie-down. You ramble. You backtrack. You say “oh, and the important bit is” halfway through a sentence, the way you would to a friend leaning on your kitchen counter. You give it the texture. The truth. The stuff that makes it sound like you, because it came out of you, out loud, before your inner critic could find her red pen.

A word for the grammar police, with love

I know it’s a touch blasphemous to sit here on Substack, of all places, and make the case for talking over writing. This is the church of the sentence. But I’ve never had much time for the snobbery of it. The dictionary police. The grammar prefects who treat a stray comma like a moral failing.

And let me be clear, because otherwise they’ll come for me: my spelling and grammar are A-plus. I know the rules cold. I just also know they were never the point. I have always, always been better yapping than typing.

My wonderful old colleague Charlotte worked this out long before any app did. For years her note to me was “get those digits on the keyboard, Nishma.” Firm. Loving. Completely ignored. Until one day she gave up, handed me a dictaphone and said, more or less, just talk, I’ll sort the rest. The Wispr Flow before the Wispr Flow. Charlotte, if you’re reading this, you were a decade early and I owe you a very long voice note.

It’s a dictaphone with a brain

That’s the bit that makes the tool more than a party trick. It’s an intelligent one.

It doesn’t just transcribe my mess, ums and false starts and all. It tidies the thinking as I talk. It knows the words I actually use, because I’ve trained it on them: the phrases I reach for, the places I want to sound casual, the places I want to sound like I’ve read the room. It loses the stumbles without losing me. And that is the whole game. A lesser tool would sand me down until I sounded like everyone else. This one keeps the grit and only takes the gravel.

So the story comes out first as a conversation. Warm, rambling, mine. And then, only then, do I edit. That’s when the red pen comes out, and I edit like Anna Wintour got hold of it. Ruthless. Cold. Cut, cut, keep, cut. Because you cannot edit a blank page, and you certainly cannot edit fear. But you can edit a ramble. A ramble is a gift. It’s raw material with a pulse.

The chatterbox was right all along

We got this one wrong.

The whole industry taught us to “prompt”. To engineer the perfect string of instructions, as if we were programming a microwave. There are courses. There are enormous decks. There are men on LinkedIn selling “the only 12 prompts you’ll ever need” for £39, cash upfront. And every bit of it points the wrong way. Briefing an AI was never a coding problem. It’s a conversation problem. And the best brief has never been the shortest, tidiest one. It’s the one where you actually explained what you wanted and why, tangents and worries and all.

Who is world-class at that? The talkers. The over-explainers. The ones who give “too much detail” in meetings and send the feature-length voice note. Every trait the office quietly filed under “a bit much” is now, out of nowhere, an advantage. The chatterbox spent thirty years being told to get to the point. Turns out the detail was the point.

So. What to actually do.

Stop typing at your AI. Talk to it. Get a tool that turns your voice into text (I use Wispr Flow, and my wrists have sent it a thank-you card) and brief it like you’d brief a clever friend who genuinely wants the context. Claude, mostly, in my case. Then use it for everything else too. The walk-idea. The dodged email. The message you keep rewriting into beige. Catch the thought where it lands, not where the keyboard happens to be.

These are my stats …

And if you catch yourself editing before you’ve even spoken, that’s the old teacher in your head. Ignore her. Talk first. Edit later, as savagely as you like.

And the bit that still makes me laugh: I love writing more now than I ever have. More essays, more posts, more of this. Not despite the talking. Because of it. AI didn’t replace the writer in me. It has quietly opened up a hundred doors that felt shut before, and it finally let her out of the room where the keyboard was keeping her quiet.

Go figure.

Right, Milly’s at the door and I’ve got roughly nine ideas to talk into my phone before they escape out of my ears.

Talk soon. Literally.

Nishma x

You’ll be Switched ON to: talking.

P.S. If you fancy trying the thing that started all this, my referral link gets you a month of Wispr Flow, free: wisprflow.ai/r?NISHMA1. Yes, it’s a referral link. No, Milly still doesn’t get commission.

Christmas 2025. The kids were away in Scotland with Dan, the house was properly quiet for the first time in years, and I was alone with the thought I had been outrunning for months: burnout was coming for me.

I could feel it approaching the way you feel a cold coming, that heaviness behind the eyes before the symptoms declare themselves. And I knew the odds were not in my favour. Burnout is wildly common among female founders, especially first-time founders, every bit as much as the corporate executives we usually write the think-pieces about. Possibly more. We just carry it with less scaffolding and fewer people watching.

Faced with that, I did what I always do when something threatens the ideas I care about. I refused to give up on them. And then I did possibly the nerdiest thing available to a woman alone at Christmas.

I dived back into AI. Not a dabble. A proper, sit-yourself-down, learn-everything-you-can dive.

I should be honest about my starting point. I am a tech geek. Yes, I worked at Google, and yes, the last big project I led before leaving was the launch of Bard, the thing the world now knows as Gemini. So AI was hardly new to me. In fact, on my first solo holiday after leaving, I sat on a beach in Mexico and decided to get nosey about everyone else’s AI, because I had only ever really known Google’s.

But that was dabbling. Finding out enough to get by, treating each discovery as one and done. Plenty of us live there. From years in tech I know that most of us use less than 10 per cent of what the tools we already own can do. And I had grown properly tired of hearing techbros explain, at length, how little they did any more because they had built themselves agents and systems and were building with AI.

The bright fire of approaching burnout turned out to be exactly the kick up the backside I needed. I put the time in. And time, plus natural geekiness, plus curiosity, turned into something I had not planned: I started building things. Small tools and systems that took real weight off, at home and in the business. At first they saved me time. Then they improved the work itself. And then they became something far more interesting than either: extension. Not doing less. Being capable of more. It was never really about saving time. It was about extending what I was capable of.

I showed some friends, purely out of sheer delight, and they all said a version of the same thing. You have to do something with this. You have a gift for making AI exciting instead of scary.

I dismissed it at first. Obviously. But by then I was hooked, and I love learning and experimenting too much to stop.

Here is what I kept noticing, though. The number and pace of AI launches is wild, and I completely understand why so many people feel left behind, overwhelmed or utterly despondent. And sitting in the middle of it all is a blind spot almost nobody in the industry mentions. Look at the marketing, the messaging, the use cases, the armies of influencers these companies engage. There is next to nothing aimed at midlife people. The very people most likely to be under enormous pressure, holding down careers and businesses while caring for kids and ageing parents, often while their own bodies stage a hormonal coup.

And midlife women are building more businesses than ever. For them, AI is not a toy and it is not a threat. Used well, it could be a saviour. It could help them switch off so they don’t burn out. Switch off the 11pm admin. Switch off the noise in their own heads, and the voices around them feeding the doubt.

They could finally switch off, if someone would just help them switch on.

So I tested that idea in public. I gave a talk at WACL’s Festival of Talent in Scotland, (thank you Jules for having me), and watched a room of 200 women go from politely disinterested, or quietly certain they were “fine with AI”, to excited, curious and hungry for more. That room told me everything.

Then I built the 30 Day AI Challenge. Think couch to 5k, but for AI. One small prompt a day, designed to take you from anxious to confident and curious, with one rule that never bent: you stay the author. AI is the colleague, the editor, the thinking partner that pushes back. Never the voice that replaces yours.

Truthfully, I was testing myself more than anything. A hundred and twenty people signed up. Then, with no promotion at all beyond a plea to share it with someone who needed it, the list grew past 800.

And the comments started arriving.

That last one is the whole point. Not one clever output. A habit. Experimenting and learning every day. The prompts I taught will date within the year, half of them probably already have. The curiosity does not date. That is the real success: building the habit of looking at a hard, stuck, boring task and thinking, could AI help me with this, and actually meaning it.

I was blown away. So I did the thing the challenge had quietly been training me for. I looked honestly at my whole portfolio of work, and I doubled down on Switched ON. A media business at the heart of a movement and a platform, helping people get switched on, to AI first and to whatever comes next, closing an AI skills gap that is widening by the month, and releasing the creativity and innovation that gets trapped behind fear.

So was born Switched ON.

And here is what I am actually here to do with it.

To translate it. To take the jargon, the hype and the daily wave of launches and turn it into something you can use on a Tuesday.

To question it. Not to clap along with every announcement, but to ask the harder things. Is it any good. Who is it really for. Who does it quietly leave out.

To make the content that helps you work out what you genuinely need to know, and, just as importantly, what you can happily ignore. Because most of the overwhelm is just noise you were never meant to carry.

To be a practitioner, not a pundit. To show you how I actually use this. Hands on the keyboard, the wins and the cock-ups, in public.

To be a protagonist, not a spectator. In the room, on the stage, in the story. Not watching it happen to me, and not letting it happen to you either.

To challenge it. All of it. The tools, the people building them, and the whole landscape they are building it for. Because someone has to ask the hard questions on behalf of the people it keeps overlooking.

That is the job. All of it in service of one idea:

helping people make sense of what’s changing, so they can thrive in what’s next.

The mission is gaining momentum.

Which brings me to the thing people asked for next. Not more prompts. A room.

Switched ON Sessions

Because here is what thirty days of watching women teach themselves showed me: doing it alone works, but it is the slow road. A room is the fast track. Two hours working with AI side by side, with someone to lean over your shoulder at the exact moment you get stuck, compresses weeks of solo trial and error into a single evening. The challenge was couch to 5k. This is the training day with the coach running next to you.

So here it is.
The first Switched ON Session, 15 July in London, in partnership with Gamma.

In the workshop we will build beautiful presentations that actually land and design scroll-stopping carousels like the ones you have watched me make, working with AI the whole way through, and you will leave with them done, not on a to-do list.

Then we go to dinner. Do not mistake the dinner for the wind-down, it is where even more of the learning happens. The best tips, the honest questions, the “show me that again” moments never come from the stage. They come out around the table, between courses, from the woman sitting next to you. I have watched it happen at every dinner I have ever hosted, and it is the part people talk about for weeks.

Seats are deliberately few, because the whole point is a room where everyone gets seen.

Book your seat: lu.ma/switched-on1

If you have been waiting for a sign to start, this is it.

Come and pull up a chair.

Nish x

Welcome to Day 30 of the challenge. The last one.

When we started, AI was probably a chat box you opened now and then and closed again, unsure. Look at what’s true now.

You have a folder that knows you. A money model. A one-page site. A handful of helpers doing real jobs, some on a schedule while you sleep. A resource your audience can use. A second brain you can think out loud with. And a whole system that can take a real problem and go to work on it.

You didn’t read about AI for a month. You built with it.

A word before we start

The thing I most want you to take from this isn’t any one of the tools. The tools will change every six months. They always do.

It’s the posture. Curious. Unbothered. Willing to open the thing and have a go. That is the whole skill, and you have it now. It is the part that lasts.

A last look back (do this one, it matters)

Step 1.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus. One last prompt:

Read my whole folder and my MEMORY file.
I’ve just finished a thirty-day AI challenge.
Write me a warm, frank note on how far I’ve come, what I built, and the one thing I should do next to keep the momentum.
In my voice.

Step 2.
Read it. Yes, it will get you a bit. Save it for the day you forget how capable you are. Then do the one next thing it named, today, while you still feel ten feet tall.

The bigger thing

Switched On exists to help ambitious people become AI confident, curious and capable, while staying completely, unmistakably themselves. Making this challenge is what crystallised it, and it’s the heart of everything I do now.

I want to be your translator, your protagonist and your practitioner. The one in the room actually doing it, not just talking about it. I’m deep in tools like Codex, Perplexity and computer use right now, and I’ll be getting specific about how I use them, with you.

A few ways forward, if any of it’s useful: keep following along, have me come into your office or run a workshop for your team, or just stay on the list while I share more.

And one genuine ask. Tell me what you want more of. Reply to this email. I read every one, and what you say shapes what I build next.

Carry the line out with you

One thing to take into the rest of your AI life. The line you drew on Day 16 and held all month. AI builds, you stay the author. The quality bar doesn’t drop because a machine helped. Hold that, and you’ll never produce slop in your own name. You’ve proved you don’t have to.

Key takeaway

You came in worried you were behind. You leave having built more than most people will all year. That isn’t luck, and it isn’t the tools. That’s you, switched on.

It was always the game. It’s yours to play.

One last thing

Thank you. For the weeks, the late days included. For showing up, for the brave folder, for the screenshots, for the things you built and the ones you gave away.

Now go and be impossible to ignore. By humans, and by the machines.

Nishma x

P.S. Last one. How was the whole thing? Three buttons, thirty seconds. Tap here

#SwitchedOn #SwitchedOnAIChallenge 🪩

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

Welcome to Day 29 of the challenge.

For 28 days you’ve been building infrastructure, even on the days it felt like admin. Your folder. Your voice. Your audience. Your numbers. Your brand. The skills you’ve made.

Today you find out what it was all for. You hand the whole thing one real problem from your actual business, and watch what becomes possible now that the context is finally there.

A word before we start

This is the payoff of context, and it’s the whole game most people never play. On Day 1, the question you’re about to ask would have given you generic mush. Today it knows you, so the answer is worth having.

One line holds, the same as it has all month. It plans and it drafts. You decide, and you stay the author. Still your judgement. Still your voice.

Step by step

Step 1.
Pick a real problem. Not a toy. Something live and a bit daunting: a launch you’re sitting on, a client you want back, a talk you’ve got with no idea where to start, growing your list, a decision you’ve been circling for weeks.

Step 2.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus for this one.

Step 3.
Hand it the whole problem, and tell it to use everything.

Here’s a real problem in my business: [the problem]. Use everything in this folder, who I am, my audience, my voice, my pillars, my numbers, the skills I’ve built this month. Give me a plan to solve it, then build the first assets I’d actually need to start, in my voice and on my brand. Ask me anything you need first.

Step 4.
Let it work. It will ask, then come back with a plan and the first pieces. Then push it.

Now make the first three things, ready for me to check and use.

If what comes back feels generic, that’s not the tool failing. It’s a nudge that one of your files is thin. Tell it what’s missing, or top up the folder, and ask again.

Step 5.
Notice the ceiling. What it just did is the kind of thing people pay agencies for, running on your own set-up. And the reason it works today, when it wouldn’t have on Day 1, is the infrastructure you built. Now imagine it on a schedule. Imagine it as an app. That is where this goes.

Step 6.
Lock it in.

Add to MEMORY.md: Day 29, handed my whole set-up a real problem: [the problem]. It used everything I’d built and gave me a plan plus the first assets. The payoff of a month of context.

A safety note

It plans and it drafts. You decide, and you check. Before you act on or send anything it produces, read it properly and confirm the facts, especially numbers and anything about a named person. The judgement is yours, and so is the voice.

What you’ll notice

The strange feeling of being understood by a tool. Not generic advice. Advice for you, your audience, your brand, your numbers. That isn’t the AI being clever. It’s a month of context paying out all at once.

Why this works

AI is only as good as what you’ve given it. Most people give it nothing and are disappointed by what comes back. You gave it a month of you. Same tool, completely different answer. That is the whole game, and almost nobody plays it, because almost nobody builds the context first. You did.

Key takeaway

You didn’t spend a month learning AI. You spent it building the context that makes AI worth having. Today it paid you back, and there’s no ceiling on it now.

Where this is heading

Tomorrow, the last one. How far you’ve come, and what happens next.

Tomorrow on Switched On

Day 30: You’re switched on.

The finish, the full picture of what you built, and where we go from here.

Nishma x

P.S. How did today land? Three buttons, thirty seconds. Tap here

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

Welcome to Day 28 of the challenge.

You’ve built a lot this month. But most of it you’ve built once. Today you do two things: you turn your proudest build into something you keep and reuse, and then you show one piece of it to the world.

Keep it, then show it. That’s the whole day.

A word before we start

The difference between doing a thing once and owning a skill is whether you can do it again in seconds. A one-off is a nice afternoon. A reusable skill pays you back every time you reach for it. So before we show the world anything, we make one thing permanent.

And the showing isn’t a humblebrag. It’s proof. In a conversation drowning in theory, a real person holding up a real thing they made is the rarest thing in the room. That’s not vanity. It’s the correction.

Step by step

Step 1.
Pick your proudest build. The scorecard. The morning brief. The money model. The industry review. The one that made you sit back and think, I made that.

Step 2.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus for this one.

Step 3.
Make it reusable. Turn the one-off into a skill you can run again in seconds.

Take [the thing I built]. Turn it into a reusable skill in my Switched-On folder: a clear saved prompt or template, with instructions, so I can produce a fresh one in seconds whenever I need it. Then tell me the one line I type to run it.

Step 4.
Test it. Run it once from the saved version, on something new. If it works the second time without you rebuilding it, you don’t have a one-off any more. You have a skill.

Step 5.
Show one piece of it. Screenshot the result. Write one true line in your voice, not “AI is amazing”.
Something that actually happened: what it changed, what surprised you, the thing you’d been avoiding that took an afternoon.
Want it sharper? Paste your line to Claude and ask it to sharpen, not rewrite.
Editor mode, exactly like Day 16.
Then post it, and tag #SwitchedOn so I can find you and cheer you on.
If public feels too big today, send it to one person.

Step 6.
Lock it in.

Add to MEMORY.md: Day 28, turned [the build] into a reusable skill. The one line to run it is [line]. And I showed one piece of it to the world.

A safety note

Blur anything private before you post. Numbers, client names, anything from your money model or inbox.

And keep the line: the words on anything public are yours, not the AI’s. Share it at 80 percent, not at some impossible 100. Done and shared beats polished and never.

What you’ll notice

Two things. The quiet power of “I can do that again in seconds now”, which is a different feeling from “I made a nice thing once”. And the replies when you show it: “how did you do that”, “I’ve been meaning to try this for months”. You stop being someone quietly using AI and become the person other people ask.

Why this works

A one-off impresses you for an afternoon. A reusable skill changes how you work for good. And showing it turns a private skill into visible proof, which is the most persuasive marketing there is, because it doesn’t feel like marketing. It’s just true.

Key takeaway

You didn’t just make things this month. You can remake them, on demand. That’s the difference between a good month and a permanent upgrade. Today you made one of them permanent, and let the world see it.

Where this is heading

Tomorrow, the big one. You hand the whole system one real problem from your actual work, and watch what becomes possible now that it knows you.

Tomorrow on Switched On

Day 29: Give it a real problem. Watch the whole thing work.

One real, messy problem from your business, handed to everything you’ve built. The payoff of a month of context.

Nishma x

P.S. How did Week 4 land? Three buttons, thirty seconds. Tap here

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

Welcome to Day 27 of the challenge.

This week you built a few skills. The morning brief that gathers your day. The industry review that watches your world. The one you built yourself. Until now you’ve run them by hand, a tap each. Useful. Today you let them run without you, so you come back to find the work already done.

But first, a quick truth about what these things actually are. Because it matters, and almost nobody gets it right.

Let’s get the words right (most people don’t)

The whole internet is calling everything an “agent” right now. Most of it isn’t. And knowing the difference is the line between understanding this and just repeating the hype. So, plainly:

• A recipe, or a skill. A saved instruction you run yourself. You press go. Your carousel machine, your morning brief when you tap it. It does exactly what you told it, when you ask.

• A scheduled task. The same skill, set to run on a timer, without you pressing go. Still does only what you told it. It doesn’t think. It just turns up on time.

• An agent. Something that actually decides its own steps and chases a goal with real autonomy, choosing what to do and which tools to use. Genuinely rare in the tools most of us touch day to day.

I’ve been calling them “agents” all week, loosely, like everyone does. Today let me be precise: what you’re about to build are scheduled tasks. Not agents. And that is exactly right. A scheduled task does only what you told it, on a timer, which is precisely what you want running while you’re not watching.

A word before we start

Two things worth saying.

One. Scheduling is a newer feature, and not every app or plan has it yet. If yours can do it, brilliant. If it can’t, you’re not behind for a second. The run-by-hand version is still a system, just one you start rather than one that starts itself.

Two. A task that runs unwatched needs its leash shortest of all. Draft only. Read only. Nothing sent, spent or deleted in your name while you are not looking. This is the day that rule earns its keep.

Step by step

Step 1.
Decide the rhythm. What runs, and when.

• The morning brief: every weekday, early, before you sit down.

• The industry review: once a week. Sunday evening or Monday morning.

• Your own one: whenever its job actually needs doing.

Step 2.
Open your app and find where scheduled or recurring tasks live. Can’t find it? That’s fine, it just means you’re on run-by-hand, which is still a system. You’re already set up, so skip the rest of these steps.

Step 3.
Set each one to its rhythm, and choose where the result lands so you’ll actually see it.

Step 4.
Check the guardrail is in every single one, word for word. Draft only. Read only. A task that runs unwatched with no leash is the one thing we never do.

Step 5.
Learn the kill switch. Find where the scheduled tasks live, and practise pausing one and deleting one, now, while it’s calm. Never run a job you can’t stop.

Step 6.
Lock it in.

Add to MEMORY.md: Day 27, put my skills on a schedule, so they run as scheduled tasks, draft-only, on a short leash. Here is how I pause or stop them: [where the kill switch is].

A safety note

Unwatched means strictest. For the first week, read what each task produces before you act on it, the way you’d check a new hire’s work before trusting them unsupervised. Trust is earned by watching, then handed over slowly.

And keep the kill switch where you can find it. Knowing you can stop it is what lets you relax enough to use it.

What you’ll notice

The first morning the brief is just there before you’ve asked. The first Monday the review is waiting. A small, strange luxury: work that happened while you were living your life. You stop being the one who remembers to do the gathering.

Why this works

A skill you run is a tool. A scheduled one is a system. The difference is what you get back: you built it once, and it pays you every day or every week with no further effort, on a short leash, doing only what you told it. That is what people mean when they say AI gives you time back. Not magic. Set-up.

Key takeaway

You now have a handful of jobs that run while you live your life. Built once. On a short leash. Stoppable in one move. That’s a system, and most people will never build one. And you can name exactly what it is: scheduled tasks, not agents. Knowing the difference is its own kind of fluent.

Where this is heading

Tomorrow, the slightly braver day. We turn your best build into something you keep, and show the world one thing you made. End of Week 4.

Tomorrow on Switched On

Day 28: Make it reusable, then show it.

Turn your proudest build into a skill you’ll use again and again, then hold one piece of it up to the world.

Nishma x

P.S. How did today land? Three buttons, thirty seconds. Tap here

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

An AI company just won the oldest and most coveted prize at the festival of creativity. For an advert that takes the mick out of AI advertising. I want to start there, because it tells you almost everything.

The Film Grand Prix at Cannes Lions 2026, Film being the festival’s oldest category, went to Anthropic. To Claude. For a pair of Super Bowl films made by Mother London, casting Claude as the quiet, ad-free space to think while everything around you screams for your attention. The whole joke is the absurdity of AI marketing. It won the biggest trophy in the building anyway.

Sit with that, because it is almost too neat. The work that won was a critique of the thing the festival had become. Because a few hundred metres from the Palais, out on the Croisette, every last square inch had been sold to an ad. Much of it ad tech. Much of it AI.

The room full of people selling interruption gave its highest honour to a film about the relief of not being interrupted.

We awarded one thing and bought another

Nobody was yapping about this bit. Walk up to the Palais and look at what won. Then walk down to the Croisette and look at where the money actually went. They are barely in the same conversation. I heard more about who got into which party than about the work the whole festival exists to celebrate.

I’ll review the rest of the winners another day, because plenty of it deserved far more noise than it got. But Claude’s Grand Prix I couldn’t skip, because it is this entire essay in sixty seconds.

We are awarding one kind of work, and quietly spending our money on a completely different one.

Eleven Cannes, and a ring light

Let me declare myself first. This was my eleventh Cannes, which this year made me roughly as fashionable as a fax machine, and saying so out loud in certain rooms had me socially exiled faster than I could neck a rosé. It was also the first year I turned up as a creator. So I am in this photo, holding a phone, with the rest of them.

From inside it, the AI was inescapable. By day three, nobody would have handed me an AI rosé, but I wouldn’t have blinked if they had. Everything had two letters bolted to the front of it. AI cuppa, AI rosé, AI everything, slapped on the front of things that were, on closer inspection, just things. The fear of the last two years had gone, which is genuinely good news. In its place, a cheerful, slightly frantic “oh, I use it like this, how do you use it?”

The clearest receipt of where the centre now sits was the Creator Hub. Last year it was tucked away on a rooftop. This year it had its own beach, with podcast studios. That is not a venue upgrade. That is the industry telling you where it thinks the centre is now.

The most radical document of the week

How did we get here? I am implicating myself again. For two decades the industry has wanted advertising to fit in a box. Wanamaker's old line, half my advertising is wasted and I don't know which half, has haunted every marketer who ever had to defend a budget. So when the ad tech companies arrived promising measurability, that they could finally tell us which half, of course we fell for it. I fell for it. There is enormous relief in a number. A number is a thing you can put in a deck and a thing that lets a CMO sleep.

The trouble is that brand has never lived in the number. Brand is instinct, taste, timing, and how you make someone feel in the half-second before they decide. It does not fit the box, no matter how badly we want it to.

Which brings me to the most quietly radical document I saw all week. The Creator Effectiveness Playbook, from WPP Media, System1 and TikTok. One of the largest datasets yet on this stuff: 1,217 paid ads, eight markets, 182,550 people, $70.5m of media. You would expect a report like that, from those names, to be a love letter to measurement. It is the opposite.

It found that Engagement Rate, the metric the entire creator economy leans on, the likes and the views and the follower counts, explains around 0.2% of whether anyone remembered your brand. Not 20%. Zero point two. Across 129.6 million engagements, no meaningful relationship. It found that 44% of marketers run creator campaigns with no KPIs at all. The thing that does build memory, by a distance, is emotion, plus a brand showing up early enough to get the credit for the feeling.

The spreadsheet, followed all the way down, arrives back at the soul.

We were never choosing between data and soul

This is the bit I think we are getting wrong, collectively, in the great rush to whatever is next. We think we are choosing between data and instinct, between the measurable and the magic. The evidence says they were never opposites. The same report shows 84% of purchases go to brands people already leaned toward before they started shopping. Being remembered before the moment is the whole game, and being remembered is a creative problem, not an engineering one.

You can hear it even in how the biggest spender in the room talks. Leandro Barreto, Unilever’s global CMO, was at pains to say the creator push is not about chasing cool.

“It’s because we believe this has more return on investment, that this helps us build the brands, to drive more trust, to drive more credibility.”

The smart money is not buying reach.
It is buying trust.
Which is just brand by another name.

The brilliant minority

The same playbook found that only 29% of creator ads beat the average brand ad, and the top 10% do four times the work of the rest. They call it the brilliant minority. So the real story is not "creators work." It is "creativity works, and creators are a brilliant way to carry it, when the work is actually good." Same as it ever was.

Which is exactly what that Grand Prix was trying to tell us.

Creativity is not under threat from AI. The proof was holding the biggest trophy in the building.

If creativity is under threat from anything, it is from apathy, from a quiet loss of interest in the actual craft while we chase the thing we can count.

Don’t build a new velvet rope

So I will say the unpopular thing, from the inside, as one of the people who turned up with a ring light. The creators have arrived, and thank god, B2B creators especially are some of the sharpest, most influential people in this industry, and it is about time they got the attention and the money.

But we are at a fork. We can integrate, or we can build a brand new velvet rope. And I watched, more than once, people who had spent years being kept out of the room start keeping other people out of it. The ones without a title worth their time, or a follower count worth their glance.

We who just got let in do not get to pull the ladder up behind us.

We have done exclusion before in this industry, and dressed it up as taste. Let’s not be so proud of doing it again, just with different bouncers.

Where was WPP?

The old order felt it too. There were genuinely new faces on the Beach, the Female Quotient and Canva among them. But walk the Croisette and the big networks were conspicuous by their absence. Dentsu aside, they barely registered. Omnicom had a cove tucked round by the harbour. I could not find a home for WPP at all. Make of that what you will.

Meanwhile the independent creative agencies were having the time of their lives, winning the work and owning the energy. My friend Rachel Lowenstein put her finger on why, talking about how independent strategist-creators carry a cultural capital the big agencies cannot always manufacture. That is the shift in a sentence.

The loudest silence on the Croisette

There was one more silence, and it was the loudest thing on the Croisette. The word DEI had, as far as I could tell, been quietly retired. Pride snuck in here and there, the Female Quotient held brilliant conversations, the WACL + Propeller Empower Café had the most interesting agenda going. But the broader conversation, about the bias baked into AI, into the work, into the stories we are about to tell at scale, was simply gone.

What stings the most? The sums spent on those in-person activations were, frankly, obscene. Staggering, beautiful, and obscene. Many of the very same brands that found that money are the ones who will tell you, all year long, that they have no budget to support the organisations working to get more women, more disabled people, more people from underrepresented backgrounds into this industry and keep them here.

No budget for that. Plenty for the beach.

I wondered, quietly, how the people those same companies laid off this year feel watching it. Told the cutbacks were unavoidable, the jobs simply had to go, and then shown a Croisette glittering with spend of no especially obvious return. When the world is genuinely on fire, and for one week in June it really did feel like it, we all just agreed to suspend our disbelief.

So, what now?

If you are a marketer: stop scoring creators on likes. Score them on whether anyone remembered you. Brief for emotion and for the brand turning up in the first two seconds, and treat creators as a brand-building force, not a distribution hack you bolt on at the end.

If you are a creator, and I am talking to myself here too: build the brand, not just the audience. Corey Martin called it the shift from the creator economy to the community economy, and the data backs him, a smaller, engaged room beats a big indifferent one every time. Learn the unglamorous craft of being remembered. And hold the door open. Your reach is worth more when the room behind you is full.

If you are the industry: do not let “measurable” quietly redefine “valuable.” The two are not the same, and the moment we forget that is the moment we make a lot of beautifully optimised work that nobody feels and nobody remembers.

See you in the small rooms

My closest observation all week was not on a stage. It was the sheer number of people, clever, senior, accomplished people, quietly pretending to understand AI because they are terrified of being left behind. That is the work I am giving my time to now. Not because it is worthy, but because the next few years will sort this industry into the people who got a fair shot at what is coming and the people who did not, and I would very much like that line not to run along the usual fault lines of who had access and who did not.

Cannes was hotter, more expensive and more confusing than I have ever known it. The creativity was not actually missing. It was up on stage holding a Grand Prix for an ad about the relief of not being sold to, while the rest of us were outside on the beach, selling. And it was in the small rooms nobody thought to film.

See you there.

Nish x

Sources: The Creator Effectiveness Playbook (WPP Media, System1 and TikTok, 2026); How Humans Decide (WPP Media and Saïd Business School, Oxford, 2025); Claude’s Film Grand Prix win reported by Adweek and Ad Age; Leandro Barreto quoted in Marketing Week.

Welcome to Day 26.

Everyone is posting. Posts scroll past, and they’re gone by lunch.

The thing people actually keep, save and come back to is a resource. A checklist. A simple guide. A little tool that solves one real problem. Today you make one, and then you do the part that matters: you give it away.

Want to see what you’re aiming for? Here’s the AI Fluency scorecard I built, up and running.

Watch: the AI Fluency scorecard in action (60 seconds)

And it’s live, so you can try it yourself: switched-on-scorecard.vercel.app. That’s the bar we’re aiming for today.

A word before we start

This is the generosity move, and generosity is how you get remembered. Visibility is not shouting louder. It is being the one who handed someone something genuinely useful. People remember who helped.

And it sits on the right side of your line. AI builds the thing, the layout, the structure. The judgement about what is actually useful, and any words in your voice, those stay yours.

One rule for today, because it’s the one we keep breaking: a resource that never leaves your folder isn’t a resource, it’s a draft. Today’s job isn’t finished when you’ve built it. It’s finished when someone else has it.

Step by step

Step 1.
Pick the one problem. Not your whole expertise. One thing your audience, the real woman from Day 18, keeps getting stuck on. Small and specific wins. “How to brief a designer.” “The five questions before you hire.” “Am I actually using AI well?”

Step 2.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus for this one. It reads your audience and pillars, so the thing you make fits the people you actually serve.

Step 3.
Pick the format. Two routes, choose by how brave you feel:

• The reliable one, for everyone: a one-page guide or checklist. It comes out as a clean PDF or a Canva design, ready to send, print or post. No faff, no hosting, nothing to set up.

• The wow one, if you fancy it: a simple interactive tool, a scorecard or quiz that gives the user a result. More impressive, a little more to build.

Ask Claude to help you choose:

Read my audience and pillars. I want to build a free resource that solves this one problem for them: [the problem]. Suggest the best format, a one-page guide, a checklist, or a simple interactive scorecard, and tell me why it fits my audience. Keep it to something I can finish and share today.

Step 4.
Build it.

Build it for me. If it’s a guide or checklist, make it a clean one-page PDF on my brand. If it’s a scorecard, make it a single web page I can open and share. Leave any line that should be in my voice marked CHANGE THIS so I write those myself. Don’t render anything heavy until I’ve seen it and said yes.

The AI Fluency scorecard up top is the one I made, the standard to aim for: genuinely useful, not a secret advert.

One practical bit, because it trips people up. Made a one-page guide or checklist? You’re done, just attach it or post it, no hosting needed. Made an interactive tool and want it on a link to share? The no-code way is a drag-and-drop host like Tiiny.host or Netlify Drop: upload your file, get a web address, send it. That’s all my scorecard link is, a file put online.

Step 5.
Make it yours. Write the CHANGE THIS lines. Then read it as her: would she actually use this? Would she thank you for it? If not, tell Claude what’s missing and go again.

Step 6.
Give it away. The whole point of the day.

• Post it: “I made a thing that solves [the problem]. Comment and I’ll send it over.”

• Or offer it as a free download in your newsletter.

• Or send it straight to one person who needs it today.

Add to MEMORY.md: Day 26, built a free resource that solves [the problem] and gave it away. AI built it, the useful judgement and the voice are mine.

A safety note

Useful, not salesy. A resource that is secretly an advert gets binned in seconds. Give the real thing, the genuinely helpful version, and trust the goodwill to come back.

And nothing with private or client data in it. What you make today is built to be shared.

What you’ll notice

The first reply. “Ooh, I need this.” “How did you make that?” That’s the moment a post never gives you. You didn’t just take up space in the feed. You were useful in it.

Why this works

A post is borrowed attention. A useful thing is earned trust. One scrolls past, the other gets saved and passed on with your name attached. AI has collapsed the cost of making the useful kind, the kind that used to be expensive. So make those, and give them away.

Key takeaway

Stop making things that vanish by lunchtime. Today you made something people keep, and you put it in their hands. That is what being chosen is built from.

Where this is heading

Tomorrow we make the whole thing run without you, and get properly precise about what is actually an agent and what isn’t.

Tomorrow on Switched On

Day 27: Set it running, and walk away.

Put your recipes on a schedule so the work happens while you sleep. And finally get straight on the difference between a recipe, a scheduled task and a real agent.

Nishma x

P.S. How did today land? Three buttons, thirty seconds. Tap here

I read every response. The data shapes the rest of the challenge.

Catching up? Days 0 to 25 are on the Switched On Substack archive HERE.

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

Welcome to Day 25 of the challenge.

For three weeks you’ve been filling a folder. Who you are. How you sound. Who you serve. How the money works. What you’re known for. The decisions you’ve made. Some days it felt like admin. It wasn’t.

Today you find out why. You point Claude at the whole folder and have a conversation you can’t have with anyone else, because nobody else has read all of it at once.

A word before we start

A chatbot starts cold every time. Your folder means Claude starts knowing you. The gap between “give me advice” and “given everything you know about me, what would you say” is the gap between a search engine and a counsel.

One honest framing. This is reflection, not prophecy. A sharp mirror, not a fortune teller. Use it to see yourself more clearly, then you decide.

Step by step

Step 1.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus for this one.

Step 2.
Let it read everything, then hold up the mirror. Paste this to Claude:

Read every file in this folder. You now have my about-me, my voice, my audience, my business, my pillars and my decisions. Before I ask you anything, tell me the three things you notice about me that I might not see myself. Be honest, not flattering.

Step 3.
Sit with what comes back. It will name a pattern you have been living inside and not quite seeing.

Step 4.
Now ask the big ones. A few worth your time:

Given everything you know about me, what should I focus on for the next ninety days, and what should I drop?

Where am I inconsistent? Where does what I say I value not match what my files show I actually do?

What am I sitting on? The thing I keep circling but never commit to.

What is the one throughline of my whole year that I haven’t named yet?

Step 5.
Keep what lands.

Save the sharpest things from this conversation as a dated note in my folder. These are the ones to act on.

A safety note

This folder is the most personal thing you’ve built in this challenge. Keep it local, keep it yours.

And remember the mirror only reflects what you fed it. If a file is thin or out of date, the answer will be too. When something feels off, that is usually the file talking, not you. Update it and ask again.

What you’ll notice

It says something true that you’d half-buried. Not because it is clever, but because it read all of you at once, and you almost never do. Most of us hold our own story in fragments. Today you see it whole.

Why this works

Every file you built this month was an investment. Today it pays out. This is the whole reason context matters: the same question, asked of a tool that knows you, gives an answer worth having. You did not build a tidy folder. You built a thinking partner.

Key takeaway

A chatbot answers questions. A second brain answers YOUR questions, because it knows you. You spent three weeks building one. Use it for the hard thinking, not just the busy work.

Where this is heading

Tomorrow we make something with all of this. Not a post. A real thing your audience can use.

Tomorrow on Switched On

Day 26: Build something your audience can use.

A resource, a tool, something genuinely useful with your name on it. The kind of thing people keep.

Nishma x

P.S. How did today land? Three buttons, thirty seconds. Tap here

I read every response. The data shapes the rest of the challenge.

Catching up? Days 0 to 24 are on the Switched On Substack archive HERE.

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF

Welcome to Day 24 of the challenge.

This week you’ve watched two agents do real jobs. The morning brief that gathers your day. The industry review that watches your world. Both mine, built to a pattern.

Today you build one that’s yours. And to show you how, I’m going to build mine in front of you: the one I reach for most, the one that turns a post I’ve written into a finished carousel. Then you build yours, the same five parts, for whatever eats your time.

A word before we start

Every agent, however clever it sounds, is the same five parts. The trigger, when it runs. The job, what it does. The context, what it reads to do it well. The output, where the result lands. The guardrail, what it must never do.

Mine takes a post I’ve already written and lays it out as a carousel, on my brand, ready to upload. I write the words. It does the design. That line, my words and its build, is the whole point. The AI doesn’t write in my voice. It frees me from the two hours in Canva.

Pick yours by the groan test. The task you sigh before starting. Repetitive enough to hand over, important enough to want done well. That’s your candidate.

Step by step

Step 1.
Pick your agent. Mine’s the carousel machine. If your groan is a different one, three that suit most people:

• The carousel machine. You write a post, it lays it out as an on-brand carousel, slide by slide, ready to upload. Your words, its design.

• The inbound drafter. People ask you the same kinds of questions by email or DM. The agent drafts a reply in your voice, filed for you to check and send. It never sends.

• The weekly money check. Every Friday it looks at the model you built on Day 19 and flags anything tight, anything overdue, anything worth chasing.

Pick the one that makes you groan the loudest. I’ll build the carousel machine here, and the shape works for any of them.

Step 2.
Open Cowork in your desktop app, pointed at your Switched-On folder. Opus for this one.

Step 3.
Build it and test it on something real. Here’s the exact prompt I used. Swap the job for yours and the shape holds:

I want to build an agent that turns a post I’ve written into a finished, on-brand carousel. Let’s build it the five-part way and test it on a real post now.

First read my brand files in this folder, my colours, fonts and wordmark, so it comes out on brand.

The spec:
Trigger: I run it whenever I’ve written a post I want as a carousel.
Job: take my post exactly as written, find the natural slide breaks (slide 1 the hook, one idea per slide, the payoff near the end, a simple call to action last), and design each slide.
Context: my brand files and the post. My words, my brand.
Output: build it as one HTML file with each slide at 1080 x 1350, then save every slide as its own numbered image (01, 02, 03) in a new folder so they upload in order.
Guardrail: do not change or rewrite my words, lay them out only. If a line is too long for a slide, tell me and let me cut it myself. Never invent a stat or a quote.

Here’s the post: [paste your post]. Build it, then show me the slides.

Step 4.
Run it by hand first. Same rule as the others. Watch it work in front of you before you trust it. When it flags “this line’s too long for slide 3, want to cut it?” instead of quietly rewriting you, that’s the guardrail doing its job. Tweak until the slides are the ones you’d actually post.

Step 5.
Save it as a one-tap job.

That’s the one. Save it as a reusable agent in a file called carousel-machine.md in this folder, with the full five-part spec, so next time I can say “run the carousel machine on this post” and paste a new one.

One line from now on: “run the carousel machine on this post”.

Step 6.
Lock it in.

Add to MEMORY.md: Day 24, built my own agent, the carousel machine. The pattern is five parts: trigger, job, context, output, guardrail. I write the words, it does the build. I can make these myself now.

A safety note

Your own agent will be closer to your real work, so the leash matters more, not less.

It builds, it never posts. Nothing goes public, gets sent, spent or deleted without you. For the carousel machine that means it makes the slides and stops. You’re the one who hits upload. Those words go in the guardrail and stay there. And know how to switch it off before you switch it on.

What you’ll notice

The quiet click of “I can build these”. Up to now you ran agents I designed. Today you designed one. That’s the line between using AI and directing it, and you just crossed it. And in the carousel’s case, the two hours in Canva, gone, with your words still your own.

Why this works

Once you’ve seen the five parts twice, the third is only filling them in for your own life. The pattern is the skill. Any single agent is just one use of it. And there’s no ceiling on it: the same five parts that build a one-tap carousel can build a whole app. You keep the pattern forever, long after these particular jobs change.

Key takeaway

Three agents now, and the third is yours. You do not need me to build the next one. You have the recipe.

Where this is heading

Tomorrow, the quiet payoff of building this folder all month. You get to talk to it.

Tomorrow on Switched On

Day 25: Talk to your second brain.

Three weeks of files in one folder. Point Claude at all of it and ask the questions only someone who knows you could answer.

Nishma x

P.S. How did today land? Three buttons, thirty seconds. Tap here

I read every response. The data shapes the rest of the challenge.

Catching up? Days 0 to 23 are on the Switched On Substack archive HERE.

Got a friend or a colleague who this might be useful for? Send her to nishma.co/30daychallenge

#SwitchedOn #SwitchedOnAIChallenge

​Unsubscribe | Update your profile | 167-169 Great Portland Street, London, W1W 5PF