Approach

Why is a lipstick founder telling you about AI?

Fair question. The honest answer is that they're the same job. Both are about making a complicated thing work in public, where it's obvious when it doesn't and somebody is waiting on you.

There's also a pattern I only noticed afterwards: I keep building the software I couldn't buy while I was doing the work. Nine years running a product company became AI Discovery Group. Years of producing events, as an IATAN-accredited producer, became Planidex. Neither started as a business idea. Both started as a Tuesday I was sick of.

Produce
Live experiences
Events, still. Everything happens in the right order, in front of people, or not at all.
Build
Brands & growth
Brands. Why people notice you, buy, come back, and tell a friend.
Operate
Founder leadership
My own company. Product, staff, margin, and every consequence.
Connect
Technology
Getting systems to talk to each other, and to the people using them.
Advance
Practical AI
AI that helps, without handing over the decisions that matter.

I learned this by being the one it landed on.

Nine years, one company, no safety net
How I work

What I actually do.

The order matters more than any single part of it. Almost every failed project I've seen did these backwards — picked the software first, then went looking for whose job it was.

01

I follow the customer, not the org chart

Your org chart shows who reports to whom. It doesn't show where people fall through. I trace what actually happens to an enquiry from the moment it arrives, and those two maps are almost never the same shape.

02

We decide what stays human

Most of this job is working out what not to automate. Nobody sells you that decision, because there's no software to invoice for it. It's usually the one that saves you the most money.

03

Then we actually build it

A strategy that stays a PDF is just an expense. Everything ends with something running and somebody's name against it, so it survives the week after I leave.

04

You can see inside it

If you can't see what a system decided and why, you haven't automated anything. You've moved the risk somewhere you can't watch it.

Principles
Someone owns it firstYou can see inside it Plan for the mess, not the demoA tool isn't a system
Before you hire me

Four things I'll tell you either way.

01

Someone has to own it before we automate it

AI can't fix a job nobody owns. Before anything gets built we write down who owns what — and more usefully, where nobody does. Half the time that list is the answer, and you don't need me after that.
On order
02

You should be able to see what it did

I won't hand you something that makes decisions you can't check. Everything ships with a screen showing what it did and why, and a written list of what it's not allowed to touch.
On trust
03

The mess is where the money goes

Every demo shows the version where nothing goes wrong. Your budget gets spent on everything the demo skipped. So we design for the mess first and write down exactly when a human takes over.
On cost
04

A tool isn't a system

A tool is something your team has to remember to use. A system is something that just happens. You find out which one you bought on your busiest Tuesday.
On the difference

Every demo shows you the easy version.

Field notes · On implementation
Your move

Let's talk.

Write me four lines about what's breaking. I read everything myself, usually the same day. If I'm not the right person, I'll say so and point you at someone who is.

Tell me what's breaking