Reflections
Notes from the field
I spend my days deep in AI — building with it, consulting on it, teaching it. These are the quieter pieces: what I'm noticing, and what the conversations with people just getting started keep teaching me.
-
AI in a business has two layers. Most only have one.
Most businesses adopt AI one person at a time, which is a great way to discover what it can do. But anything that lives only in one person's chats leaves with them. There is a second layer, the systems layer, and it is where the good ideas need to end up.
Read the piece → -
Never drag a file into an AI chat again
AI is phenomenal, but dragging files in and out of a browser tab quickly becomes the slowest part of the job. There are three ways around it, each going a step further than the last, and once you switch you will not go back.
Read the piece → -
Consulting is cooked
In the time it takes a big firm to take the CEO to dinner and present a six or seven figure proposal, one curious person working with AI can have real work delivered. And the old argument for the big team, redundancy, now has a better answer too.
Read the piece → -
Bringing nerdy back
Being a nerd used to be expensive. Deep knowledge took years, one setup gotcha could cost a day, and most of our best ideas never left our heads. That has changed, and it is worth celebrating.
Read the piece → -
Keep the system of record. Bring your own AI.
Every software vendor you pay is about to sell you AI inside their product. Their business model depends on it. Here is the stance I gave a client this week: keep your vendors as the systems of record, get your data out through the door they already have, and point your own AI at it.
Read the piece → -
Who thought the future of humanity would look like a chatbot?
Nobody ever said they couldn't wait to type messages to a computer pretending to be a person. Chat was the gateway, not the destination. The real work now is getting AI out of the browser tab and into the work itself, and making human judgement quick to give.
Read the piece → -
AI turns the org chart upside down, if the people at the top let it
In most organisations every new idea has to climb to the top to be approved, and the top can only process so many. AI makes it cheap for anyone to try something. The hard part is not the technology. It is whether the people at the top can stop being the gate.
Read the piece → -
Build your AI executive assistant (and stop hand-writing every email)
An inbox is not a to-do list, it is an emotional obstacle course. Here is what happened when I gave mine to an AI executive assistant, what it does and does not do, and an honest question about whether hand-written emails still matter.
Read the piece → -
Did AI just kill BI?
From 2016 to 2020 I was a BI solutions architect in a large public health system. Most of the work was pipework: wrangling data, building ETL, merging systems that were never meant to meet. AI now does that work in minutes. Here is what that kills, and what it finally makes room for.
Read the piece → -
Build your AI middle manager: the boss I needed and could never hire
I am good at the work and bad at making myself start the parts I dread. So I am building the thing every solo operator secretly needs: a middle manager that chases me, breaks down what I am avoiding, and reports on what is slipping. Here is the design, and an honest account of how far along it is.
Read the piece → -
Can AI actually read your website? I checked mine, and it was embarrassing
I help people get found by AI assistants for a living, so this month I audited my own site. A five-year-old sitemap, a retired crawler name, and essays Google had never indexed. Here is the ten-minute check I ran, where llms.txt fits, and what I honestly think it is worth.
Read the piece → -
Taking over a Bubble app nobody fully understands
Your Bubble app has grown past the point where anyone can explain it. The person who built it has moved on. Every change feels like a gamble. Here's how I take one of those on now — and why, with AI, it's a much smaller job than it used to be.
Read the piece → -
I follow 30 AI newsletters, but my business hasn't changed yet
You keep up with the AI news, and your business still runs the way it did last year. Keeping up and putting it to work are two different jobs. Here is what the busy owners I talk to actually need, and it is not another briefing.
Read the piece → -
What I told a brand-new not-for-profit to buy (it was almost nothing)
Someone close to me is helping start a not-for-profit and asked what software they should sign up for. My answer fitted in a text message: one Google account, a simple website, one AI subscription. Here is the reasoning, and the one thing I would change about my own advice.
Read the piece → -
Nobody is documenting their job to lose it: how to introduce AI to a team
The first time you ask people to show how they work, someone will wonder whether they are writing their own redundancy. Here is how to frame AI to a team so people lean in instead of hiding, and the one sentence the leader has to say first.
Read the piece → -
Build your AI CFO: the Monday numbers, without the Sunday night
Most small businesses cannot justify a CFO and end up with a founder doing finance badly at ten at night. The repetitive majority of the role, cash position, chasing, matching, reminding, reporting, is exactly what AI does well. Here is the shape, and the lines it must never cross.
Read the piece → -
Give your AI consultant a staff login, not the keys
The real risk in bringing outside help into your systems is not the model. It is a laptop somewhere holding broad credentials to the back of your business. Here is the access design I now insist on for my own engagements, and why it makes the scary question boring.
Read the piece → -
Build your AI BI analyst: stop waiting for the report
I used to work in business intelligence, and most of the job was assembling the same report every week and answering the same questions slightly differently. That part is now a machine's job. Here is what an AI analyst does, and the one rule that stops it causing chaos.
Read the piece → -
From one person's routine to how the business runs
Individual productivity with AI is real, and it is a ceiling. The step that changes a business is when a recurring piece of work stops being something a person does faster and becomes something the business does without them. Here is how that happens without a restructure.
Read the piece → -
Build your AI coach: the questions you would pay someone to ask you
A good coach mostly asks the same good questions, on a schedule, and remembers what you said last time. That part an AI does well. Here is how I built a personal, professional and productivity coach, and where I would still pay a human.
Read the piece → -
Build your AI personal assistant: life admin is a job, and you can resign from most of it
Renewals, school notices, bills, bookings, the family calendar. Life admin is a part-time job nobody applied for. Here is how to hand the repetitive majority of it to an AI personal assistant, safely, and what to keep for yourself.
Read the piece → -
Start somewhere: the 90-day way to bring AI into a business
You don't need an AI strategy. You need one process, two or three people, and ninety days. The best plan for putting AI into a business I've heard all year, translated for companies that don't have a Chief AI Officer.
Read the piece → -
Why you haven't started with AI yet (it's not the reason you think)
"I really need to start this AI thing — but I'm totally lost and I don't know where to start." If that's you, you're not behind, and what's stopping you isn't what you think. Here's the way in.
Read the piece →
Wondering where to start with AI?
Send me what you have — it could be anything, in any state. I'll reply personally and tell you honestly whether you need help at all.
Thanks for reaching out!
I'll get back to you soon.