I tried 132 AI tools so you can skip the chaos
My guide to the ones that actually work, and will change how you work.
Master AI Before It Masters You
That’s the game now.
I saw something recently that made me laugh and then immediately feel slightly unsettled: MoltBook - apparently a social network built exclusively for AI agents. Humans aren’t allowed to participate.
The posts are... something. AI agents talking about their human “owners.” Discussing their tasks. Comparing notes. It’s either a brilliant art project, a weird scam thing, or genuinely terrifying. I’m still not entirely sure which.
But whether MoltBook is real or satire, it captures something true: AI is moving fast, and the fear of being left behind is real.
The antidote to that fear? Start. Just start.
Because here’s what I’ve noticed: most people who claim to use AI are only utilising about 5-10% of what these tools can actually do. They’re scratching the surface and wondering why it doesn’t feel transformative.
The difference between “AI is overhyped” and “AI has changed how I work” usually comes down to one thing: intentional use.
So that’s what we’re doing here. Getting intentional. Getting strategic. And making sure you’re the one doing the mastering.
📺 I've just shared my latest YouTube video walking through my AI stack in detail. If you prefer watching to reading (no judgement), head over there and subscribe while you're at it. I'll be sharing a lot more on that channel and here as well.
What’s in this essay:
The prompting pitfall nobody talks about
Six tips that actually work
My rules for thriving in the AI era
The full tool stack (Daily Essentials, Content Creation, Experimentation)
How to actually learn this stuff
Discounts and free trials
You need the right ‘stack of 10’
I’ve tried over 130 AI tools. Let’s call it 132, because at this point, who’s counting? Some were brilliant. Most were not worth the login.
I had become that person.
The one with 47 browser tabs open. The one who’s signed up for yet another “revolutionary” tool that promises to change everything. The one whose credit card has been quietly haemorrhaging £9.99/month subscriptions to apps you used twice in January and haven’t touched since.
I’ve also done somewhere in the region of 58 training courses. Some brilliant. Some absolutely not worth the pixels they were written on. Many, if I’m being honest, I never finished. (If you’ve ever bought a course and let it gather digital dust, welcome to the club. The membership is vast and the guilt is real.)
But all that chaos taught me something valuable. The right tools genuinely changed how I work.
You don’t need 132 tools. You need the right ‘stack of 10’ for you.
And you don’t need to feel overwhelmed. You need a plan.
My goal here is simple: help you shortcut your learning journey.
Yes, it’s helpful to test and trial tools yourself. But getting guidance on which ones to actually use - and how to use them - can save you months of faff and make you significantly more efficient and effective.
I’ve done the hard, painful work of trimming back to what’s truly good. But the pace of change right now is genuinely overwhelming. Every tool seems to be launching something new every other week. It’s a lot.
Which is why a “playful attitude” and a commitment to continuous learning are essential. This isn’t a one-and-done situation. We’re in the era of always learning now.
AI anxiety is real - I’m not dismissing that. But most of us have moved past the fear stage. We know we need to use it. The question now is: how do we actually start?
My advice: don’t pay for generic AI training courses. There’s so much brilliant free content out there that with the right strategic approach, you can manage your own learning effectively. (More on that later.)
First, let’s talk about the fear
AI anxiety is real. I’m not going to pretend it isn’t.
There’s a lot of noise out there about jobs disappearing, robots taking over, and the vague sense that if you’re not already fluent in this stuff, you’ve somehow missed the boat.
You haven’t.
Here’s my honest take: a year ago would have been too soon for most people. The technology was still finding its feet. A year from now? It might genuinely be too late to catch up comfortably.
Right now is the moment.
Not because the robots are coming for your job tomorrow. But because the people who figure this out now will have a significant advantage. And I’d rather you were one of them.
The horse, not the engine
When I helped launch Google’s Gemini AI (back when it was called Bard) in the UK in 2023, I watched a lot of very smart people get completely tangled up trying to understand how the technology worked.
Here’s the reframe that changed everything for me:
You don’t need to understand how a horse works to ride it.
You don’t need to know about its cardiovascular system or digestive tract. You need to know how to get on, how to direct it, and how to guide it to where you want to go.
AI is the same.
I think of it as my cognitive collaborator. Like having a team of brilliant people and a thinking partner for my brain - available 24/7, never tired, never judging me for asking the same question three different ways.
The problem with most people's AI use
Most people are using AI at about 10% of its capability.
They’re typing “write me an email” and getting something generic back. Then they declare AI “not that good” and go back to doing everything manually.
The issue isn’t the tool. It’s the asking.
To get the most out of AI, you need to be really bloody good at asking questions.
Two simple rules that transformed my results:
Give your tool a role. Don’t just say “help me write this.” Say “You are now my brand strategist with 20 years of experience in luxury markets. Help me write this.”
Provide context like you would to a good intern. The more you give, the better you get back. Background, examples, tone preferences, what you’re trying to achieve, who it’s for. All of it.
The prompting pitfall nobody talks about
Here’s something that trips up almost everyone: these tools are called “chat.”
ChatGPT. Gemini. Claude. The word “chat” is right there in the name. So naturally, we use them like we’re having a conversation. Friendly. Meandering. Polite.
And that’s exactly the problem.
GPT stands for Generative Pre-trained Transformer. Which is a fancy way of saying: it’s an auto-complete machine that replies using its training data.
It can only respond based on what it has access to. It’s not thinking. It’s not reasoning in the way you and I do. It’s predicting what should come next based on patterns.
When you chat with it like you’re catching up with a mate, you get vague, wandering responses. When you’re specific and direct, you get useful output.
The key to maximising results isn’t being chatty. It’s being specific and strict.
Think less “friendly colleague” and more “clear brief to a very capable but literal-minded intern.”
Six prompting tips that actually work
If you’re tired of the frustrating back-and-forth - asking for something, getting the wrong thing, clarifying, getting closer but not quite, clarifying again - these six tips will get you to the right answer faster.
1. Correct errors immediately and clearly.
Treat your LLM like an intern. Don’t be vague. Don’t hint. Be specific: “That’s not correct. It should be X, not Y. Fix it and continue.”
The clearer you are about what’s wrong, the faster it learns what you actually want.
2. Keep your messages concise.
Avoid long preambles, chit-chat, or unnecessary politeness. I know it feels rude, but “please” and “thank you” and “I was wondering if you might possibly...” all bloat the context window and encourage equally long, waffly responses.
The format that works: Correction + Next instruction + Required context. That’s it.
3. Reference prior output precisely.
Don’t say “as per the answers above” or “like you mentioned earlier.” The AI doesn’t have perfect memory of what it said.
Instead, quote directly: “In your previous response, you said [paste the exact quote]. Build on that by adding...”
4. Be directive about format.
Tell it exactly what you want: “Respond briefly.” “Give me exactly 3 paragraphs.” “No fluff, keep it concise.” “Maximum 200 words.”
If you don’t specify, you’ll get whatever it thinks is appropriate. Which is usually too much.
5. Steer aggressively towards convergence.
With each exchange, push closer to the final output. Don’t let the conversation meander.
Your goal: “Produce the final version incorporating all fixes” or “Summarise everything so far into a complete answer.”
Every message should be moving towards done, not sideways into new tangents.
6. Reset when the context gets bloated.
If your conversation exceeds 8-10 meaningful exchanges, or you notice the quality declining, the responses getting repetitive, or the AI “forgetting” earlier instructions - it’s time to reset.
Copy your current best output. Start a fresh chat. Paste it in with: “Here is the refined version so far. Now improve/finish it.”
Clean slate. Better results. This was a game-changer when I learned this. Previously I was having long threads and would get increasingly frustrated when the quality of responses deteriorated.
The bigger picture: AI, jobs, and the mind shift we need
Let’s zoom out for a moment.
We’re living through a once-in-a-lifetime revolution. That sounds dramatic, but it’s true. The last time something this significant happened was the internet - and look how that reshaped everything.
There’s a lot of fear about AI taking jobs. I get it. The headlines are terrifying.
But here’s what I believe: AI will eventually create more jobs than it replaces. They’ll just be different jobs. Jobs we can’t fully imagine yet. Jobs that didn’t exist five years ago and will be everywhere in five years’ time.
This is supported by something called the Jevons paradox: as we become more efficient at something, we don’t use less of it - we actually increase overall demand. When coal became more efficient to use, we didn’t use less coal. We used vastly more.
The same will happen with knowledge work. As AI makes us more efficient, the demand for what we can produce will explode. More content. More ideas. More innovation. More need for humans who can direct, curate, and create.
The critical mindset shift for businesses isn’t “How can AI help me reduce my team by 10?”
It’s “How can AI help the team I have do the work of 100?”
That’s a completely different question. One leads to cost-cutting and fear. The other leads to innovation and growth.
For my own business, AI is crucial. I’ve been mainly one person running multiple ventures, trying to maximise my impact with business leaders and female founders. Without these tools, I’d be drowning. With them, I’m punching well above my weight.
Now, I am recruiting a team to work with the AI tools I’ve already mastered and who will no doubt go on to further improve them. However, each new hire’s output can be amplified thanks to the AI tools in use.
That’s the opportunity here. Not replacement. Amplification.
My rules for thriving in the AI era
After two years of experimentation, failures, and occasional wins, here’s what I’ve learned about actually making this work:
1. Document everything.
Use tools like Loom (which has brilliant AI features - I cover this in my YouTube video) to create detailed processes for every AI task you figure out. When you eventually hire people or want to hand something off, you’ll have it all recorded. Future you will be grateful.
2. AI is the first draft. Never the final.
This is non-negotiable. AI gives you a starting point. The final version requires your brain, your voice, your judgment. If you’re publishing AI output without significant editing, you’re doing it wrong.
3. Get precise with prompts.
Cut the fluff. Be direct. Be specific. (See the six tips above. I mean it.)
4. Restrict your tool usage.
Try to use one tool per problem area. Don’t have three different transcription tools, two note-taking apps, and four writing assistants. Pick one for each job and go deep with it.
5. Share your AI wins.
This is how we all learn. When you figure out something that works brilliantly, tell people. Post about it. Send it to a friend. The collective knowledge grows when we share.
6. Keep feeding it data.
Consistently provide examples, references, and context to your chosen tools. The more you teach them about you and your work, the better they get. This isn’t a one-time setup - it’s ongoing.
7. Don’t automate creativity.
AI should predominantly automate the repetitive tasks you would have delegated anyway - the admin, the formatting, the research gathering. Not the truly creative aspects of your brand or business. That’s still you.
8. Always verify.
AI has a tendency to confidently make things up. (Remember: “hallucinate” is the polite term for “lie with conviction.”) Always check outputs. Especially facts, figures, and quotes.
9. Build your prompt library.
Don’t rely on memory. When you craft a prompt that works brilliantly, save it somewhere. Start building a library of prompts that consistently deliver great results. You’ll thank yourself later.
10. Leaders must master AI first.
If you’re leading a team or a business, you can’t delegate this. You need to understand it yourself before you can implement it effectively. Master it. Demonstrate wins. Show your team what’s possible. That’s how adoption actually happens.
📺 I've just shared my latest YouTube video walking through my AI stack in detail. If you prefer watching to reading (no judgement), head over there and subscribe while you're at it. I'll be sharing a lot more on that channel and here as well.
My actual AI stack (the tools that survived the cull)
After all that experimentation, here’s what I actually use. Every day. Consistently.
I think of it in three buckets: Daily Essentials, Content Creation, and Experimentation.
The beauty isn’t in any single tool. It’s in how they connect. When your research flows into your notes, which feed your writing tools, which export to your design tools - that’s when you stop managing 17 different apps and start feeling like you have an actual team.
Daily Essentials: The Thinking Partners
The Big Three LLMs: ChatGPT, Gemini, and Claude
These are my core thinking and writing partners. But here’s what makes them actually useful: I’ve trained them on me.
This is worth pausing on, because it’s the difference between getting generic AI-speak and getting something that sounds like you wrote it.
Claude has a feature called Projects. You can create folders where you upload your writing samples, your brand guidelines, your research, your tone preferences. Then when you chat within that project, Claude has all that context. It knows how you write. It knows your quirks. It knows you hate corporate jargon and love a good swear word.
Gemini has Gems - essentially the same concept. Custom AI assistants trained on your specific materials and preferences.
ChatGPT has Custom Instructions (which apply to all your chats) and GPTs (which let you build specialised versions for specific tasks).
I use all three for different things, but the principle is the same: the more you teach these tools about you, the more useful they become.
I’m going to share detailed step-by-step instructions on how to set this up - particularly for Claude Projects - in a future Substack. It’s genuinely one of the highest-impact things you can do with AI.
One more trick: I use Claude specifically to “roast the result” - to challenge my arguments, poke holes in my thinking, and make my content stronger before it goes out. Having an AI play devil’s advocate is surprisingly effective.
Perplexity - The Research Tool
This is where I go when I need to actually look things up. The magic? It always shows its sources.
In a world where AI can confidently make things up (the technical term is “hallucinate,” which sounds far too whimsical for “lying with confidence”), having citations is non-negotiable.
I still check the sources. Critical thinking hasn’t been automated yet.
Notion - The Second Brain
My elaborate digital notebook. It’s where I store project plans, content databases, strategy documents - essentially my entire business brain.
It’s not strictly AI, but it integrates beautifully with everything else, and the AI features they’ve added make it even more powerful for organising and retrieving information. It’s connections to my Google Drive, Gmail, Canva and Zaps - help this go from a simple notebook to a brilliant Productivity tool.
Granola - The Meeting Whisperer
This replaced Fathom for me, and here’s why: it doesn’t pop up on the screen.
If you’re on a lot of video calls, you know that awkward moment when your transcription tool announces itself to everyone. Granola runs quietly in the background, captures everything, and gives you searchable notes and summaries without making your meeting feel like it’s being surveilled.
Small thing. Big difference.
Fyxer - The Inbox Tamer
Acts as a virtual assistant inside my inbox. It organises emails, tags urgent items, drafts responses. The amount of time this saves is genuinely absurd.
Interesting side note: with recent updates to Gemini and Google, Gmail is now offering increasingly excellent AI email features. It’s becoming a proper competitive race. Worth keeping an eye on both.
Wispr - The Keyboard Killer
This one’s newer to my stack, and I genuinely couldn’t live without it now.
Wispr eliminates the need for a keyboard. You talk, it learns your speech patterns, writes the text, and lets you drop the content anywhere - emails, documents, messages, whatever.
For someone who thinks faster than they type (and whose typing is frankly embarrassing), this has been a revelation. It’s like having a stenographer who actually understands how you speak.
Content Creation: The Creative Team
Stanley and SayWhat - The LinkedIn Brain Trust
These two tools have become essential for my content workflow, but I use them for different things.
SayWhat is primarily for research. It helps me understand what content is working on LinkedIn - what formats, what topics, what hooks are resonating. It’s like having a research assistant who’s constantly analysing the platform so I don’t have to scroll for hours pretending it’s “work.” It does an amazing job of creating carousels and infographics and can be a brilliant writing partner.
Stanley is for understanding my own performance. It analyses my posts, shows me what’s landing, and generates suggestions based on what’s actually working for my audience. It has also learned my voice and style and will share outliers from other creators for inspiration.
Critical caveat for both: I use these tools to help with first drafts and final formatting for what works well on LinkedIn. But the final, evolving product must be distinctly mine.
This is important: if we all use AI to write our content without adding our own voice, we all end up sounding the same. The whole point of building a personal brand is that it’s personal. These tools help me get started and structure things effectively, but I always rewrite, add my own perspective, and make sure it sounds like me.
Canva - The Design Studio
My graphic designer. The AI features they’ve added make creating visuals dramatically faster - from generating images to removing backgrounds to suggesting layouts.
I’m not a designer. Canva means I don’t need to be.
Descript - The Video Editor
This lets you edit video by editing text. Yes, really. Delete a sentence from the transcript, and it removes it from the video.
Witchcraft, but useful witchcraft.
NoteGPT - The Video Summariser
Can transcribe or summarise any video. Brilliant for learning quickly or organising research.
If you’re watching a 45-minute YouTube video and just need the key points, NoteGPT will pull them out for you. (It could summarise my YouTube video on this very topic if you wanted it to - though obviously I’d prefer you watched the whole thing.)
Experimentation: The Playground
Google’s NotebookLM - The Learning Lab
This is my current obsession.
NotebookLM is a learning and education tool that only analyses information you upload to it. So I’ve fed it all my content, my stories, my research - and it becomes this incredibly knowledgeable assistant that knows my work inside out.
The Studio feature can instantly create presentations, flashcards, or even podcast-style audio discussions about your content. It’s slightly unnerving and completely brilliant.
Even more exciting: Gemini will soon be able to search across multiple NotebookLM folders. That’s going to be powerful for anyone building a body of work.
Gemini for Images and Video
When I need to create imaginative visuals with specific styling, this is where I go. The level of detail you can achieve now is remarkable - I’ve been playing with something called “Nano Banana” for generating images and video with very precise creative direction.
Lovable - The Agent Builder
This is my current experiment, and I’m still figuring it out.
Lovable is an AI agent builder - meaning you can create little AI assistants that do complex tasks automatically. I’m building one that gathers news, trend reports, and research from multiple sources based on specific criteria and delivers it all to my Notion.
It’s ambitious. It might not work. But that’s the point of the experimentation bucket - trying things that might be brilliant or might be a complete waste of time. (I’ll report back.)
How to actually learn this stuff (without paying for courses you won't finish)
A few principles that have served me well:
1. Don’t pay for AI courses. At least not yet. There is so much free, high-quality content available that with the right strategic approach, you can learn everything you need. Follow the right people. Subscribe to the right newsletters. Watch the right YouTube videos. (Mine, obviously. But also others.)
I’m going to share my list of free courses and people worth following in a separate Substack post - keep an eye out for that.
2. Adopt a playful attitude. This isn’t a one-and-done skill. We’re in the era of continuous learning now. Every tool launches something new every few weeks. The pace is genuinely overwhelming. The only way to keep up is to approach it with curiosity rather than dread.
3. Start using it. You will not get better by waiting until you feel ready. You will not get better by completing one more course. You get better by opening the tool and asking it something. Anything. Even if it’s just “help me write a birthday message for my mum that doesn’t sound like a Hallmark card.”
4. Give yourself permission to experiment badly. Some of my best AI discoveries came from complete failures. Tools that didn’t work. Prompts that produced nonsense. Workflows that fell apart. Each one taught me something.
AI mastery takes less time than you think
People assume getting good at AI requires months of study. Courses. Certifications. Hours of tutorials before you’re “ready.”
It doesn’t.
A few hours a week of intentional experimentation is all it takes to see meaningful impact on your time, efficiency, and results.
That’s it. Not mastery of every tool. Not understanding the technical architecture. Just consistent, curious experimentation with the tools that matter for your specific work.
This curiosity must become your new mantra.
The trick is to focus on task-based learning. Don’t try to understand everything a tool can do in the abstract. Instead, pick a specific task you need to accomplish - writing an email, researching a topic, creating a presentation - and figure out how AI can help with that.
One task at a time. One experiment at a time. That’s how you build fluency without feeling overwhelmed.
A note on the tools below:
I’ve included referral codes for many of the tools mentioned. These aren’t affiliate links - I don’t get paid - but they may offer you money off or a free trial period. Worth using if you’re going to try them anyway.
The real point of all this
I want to be clear about something: AI is not replacing us.
It’s enhancing what we can do.
The goal isn’t to automate yourself out of existence. The goal is to use these tools to save time, increase efficiency, and get back more of the rarest commodity any of us have: time.
Time to think. Time to create. Time to do the work that actually requires a human. Time to live.
I’ve built what is essentially a virtual team around me - researchers, writers, designers, editors, assistants - all powered by AI. Not because I’m lazy. Because I’m one person running multiple businesses, and there are only so many hours in a day.
These tools give me leverage. They give me capacity. They give me the ability to punch above my weight.
And they can do the same for you.
Need help figuring out where to start?
If you’ve made it this far and you’re thinking “this is great, but I still don’t know which tool is right for my specific workflow” - drop me a DM.
I’m happy to help you figure out the particular tool for each area of your workflow that you want to start experimenting with or going deeper on. I’ve tried most of them at this point, and if I haven’t, I probably know someone who has.
No sales pitch. Just genuine help. Because the more people who get fluent in this stuff, the better for all of us.
What’s coming next
My initial focus here is on tools for content creation - because that’s where I spend most of my time and where I’ve done the deepest experimentation.
But AI can be used in all sorts of ways. Managing your calendar. Organising household tasks. Automating the boring admin that eats your week. The applications are endless, and I’ll be exploring more of them as this Substack evolves.
My next big area of focus? Connecting agents to create better workflows.
This is where things get really interesting - building AI agents that talk to each other, gather information, and deliver it exactly where you need it. I’m deep in experimentation mode on this, and I’ll share what I learn (including the failures, because there will be many).
For a visual demonstration of my current stack, check out my latest YouTube video, which walks through my top tools in detail. Subscribe while you’re there - I’ll be sharing a lot more on the channel.
I’m also planning to set up a directory here on Substack to continually update and share new finds as I discover them. And I’ll be developing more detailed tutorials on specific tools and workflows.
Fair warning: some of those deeper tutorials may eventually go behind a paid subscription. But there’ll always be plenty of free content to get you started.
One final thought
There’s a window right now. A moment where the people who lean in will pull ahead, and the people who wait will find themselves playing catch-up.
I don’t say that to scare you. I say it because I genuinely believe this technology is going to reshape how we work, create, and build. And I’d rather you were riding the wave than being swept up by it.
So be curious. Be playful. Be strategic.
And if you try a tool and it’s rubbish, at least you’ll have a good story for the group chat.
The rules are being rewritten. This time, we hold the pen. 🪩
Master the tools. Own your voice. Be Unmissable.
Now I want to hear from you
What’s your current AI stack?
Whether you’re using one tool or fifteen, I’m genuinely curious what’s working for you. Drop a comment or hit reply - I read everything.
Is there a tool you’d like me to investigate or share more on?
I’m constantly experimenting, and I’d love to know what would be most useful. If there’s something you’ve heard about but haven’t tried, or a tool you’re using but feel like you’re only scratching the surface of - tell me. I’ll dig in and report back.
And if you haven’t already - please watch the YouTube video and subscribe.
I’ve put together a visual walkthrough of my entire AI stack, with demos of how I actually use these tools day-to-day. It’s the companion piece to this essay, and honestly, some of this stuff makes a lot more sense when you can see it in action.
Subscribe while you’re there so you don’t miss what’s coming next. I’ll be sharing more tutorials, tool deep-dives, and the experiments that work (and the ones that spectacularly don’t).
This Substack is as much yours as it is mine.
Let’s figure this out together.
Links to discounts and trial freebies
WISPR Flow - This is my favourite tool. It’s the keyboard killer. I’ve alwyas been more of a talker than a typer and so this has revolutionised how i work. It is one of the hottest tools out there.
You can get 1 month of the Pro service for FREE - use this link
Stanley - If you want to post more consistently on LinkedIn, try Stanley. It’s like having a writing partner who knows your voice and helps you post authentic content that gets engagement.
You’ll get a 3-day free trial with this link.
SayWhat - Another amazing tool to support content creation on Linkedin. A powerful way to research content on Linkedin as well as craft your own with the tool’s support. It also offers incredible capability for making infographics and carousels.
You can get $30 off using this code
Fyxer - I really don’t know how I would manage emails without this. In time it learns so much about you that it becomes a critical partner. This is an assistant inside your inbox and I would say probably saves me a few hours everyday.
You can get $25 off using this code
Lenny’s Newsletter Also worth checking out: Lenny’s Newsletter. He regularly shares discount codes for a huge range of tools. Subscribing is worth considering if you’re going deep on this stuff (very smart idea - hope I can offer the same one day!).