Living with AI

You've been using AI for months, probably years. You have your doubts without being a luddite. You've done your research. And still there's a feeling of being slightly behind that never quite goes.

Maybe work felt easier to get your head around when it was more of w craft to learn and master. And probably it was more fulfilling as well. Now we've had this promise that AI can just "do it all" for us. The AI companies are falling over each other making grand promises, too.

Those are from research interviews I ran this year. Every one of them was further along than they thought. Every single one felt the pressure to "use more AI" at work, and yet somehow questioned either how do to this while retaining a sense of joy and pride in their work, or asking why in the first place. We had this technology for years now, it's the biggest investment humanity has ever made into any technology and try somehow things just don't add up.

The reason learning about AI doesn't compound isn't that you picked the wrong tool or found the wrong prompt. It's that framing that we were given: "Learn how to use AI, use it a lot, get more productive, and you'll be fine." What's missing are the questions like: What do you need to do your best work? What am I trying to do here? What does good look like?

Because something was taken from you twice, and only the second one had AI in it.

The first was process. Decades of standardisation, work broken into fractions, judgement centralised into oversight and sign-off. It was a rational trade at the time: distributed judgement is expensive to coordinate and standardisation is cheap. It was also the beginning of you not getting to decide much.

Then, AI got rolled out. AI arrived promised as the technology that could do anything. And it's genuinely very good at some of them, but it was handed over as a software update: Office 365 used to mean Word and Excel, now it's Word, Excel and Copilot, go and use it. You might even have a leaderboard showing who uses AI the most, and consequences if your number is low.

Nobody asked how AI should change the way the work actually runs, so it gets pushed into processes already fragmented and standardised half to death. And because saying I don't know what I'm doing with this is exposing at every level of seniority, nobody says it, and everybody assumes they're the only one.

Which is the hopeful part, oddly. For decades we shaped work by moving decisions away from the person doing the work, and this is the first technology to arrive that could move some of those decisions back to those who do the work. If we do it right.

Open a chat window and know what you're doing. Not confidence as a feeling, but the specific kind that comes from having decided what you want before you start typing.

Say what you want, what you need, and what good looks like. It's a lot easier to know when something doesn't look right than to put into words what would be, and with AI that gap is the whole job.

Have an opinion, because you have a practice. You'll be the person people bring their AI questions to — not because you follow the news, but because you've actually run the thing they're asking about.

Keep a practice that doesn't expire when a new model comes out. Not a template and not a prompt library. Your own way of working, which grows with you and with whatever arrives next.

Seven modules, self-paced, and deliberately tool-agnostic. It works whether you're on Claude, ChatGPT, Gemini or a Copilot licence you didn't choose. However, you'll get the most out of the course if you're on a paid plan.

Each module opens with a small rep in whatever tool you already use. Five to ten minutes hands-on, and then you learn about the thinking behind the exercise. Bring your reflections and questions to weekly office hours so you can learn from others and get help if you're stuck. And there will be optional exercises and stretch goals if you're hungry to dive deeper and take the principles into agentic workflows.

• The Primer The loops that don't change (free, available now)
• Module 1: The Vision What do you want work to be like in three years?
• Module 2: The Relationship The field you're already in, and the first move
• Module 3: The Pause Agency, intention, and reconnaissance
• Module 4: The Alignment Building shared understanding with your AI
• Module 5: The Test Going beyond proofreading
• Module 6: The Toolkit Document, share, and assemble your own kit

First run: Late 2026. Get on the waitlist to have the chance to lock in founder pricing.

It's for you if you've been working with AI for a few months, feel pretty competent in some areas, and somehow you still feel a bit trapped or like you need to 'get your head around it'.

For you that might look like:

• You avoid conversations about AI at work.
• You feel like you need to try every new tool, then lose interest.
• You worry about your job.

It isn't for you if you want a prompt library or a ranking of the best tools this month. Those exist, they're free, and they'll be wrong again in a couple of months.

It may be for you, and you still might want something else if you're a leader in your organisation and you're thinking about yourself and your team.

• If you're personally stuck between a mandate from above and a team that isn't talking about it, a course won't reach that. That's what 1:1 coaching is for.
• If your usage dashboard looks fine, your organisation is burning through tokens, but nothing has changed in the actual work, that's an organisational problem. That's a different piece of work.