AI is a tool. The job is cutting through the noise so you can see where it helps your brand and where it doesn't.

There's never been more talk about AI in ecommerce, and there's never been more confusion about what to actually do with it. Every agency's selling an AI service. Every platform's bolting AI onto itself. And every founder's being told she's already behind.

Some of it helps. A lot of it's for show. So the work is telling the difference, and then putting the useful part somewhere it earns its keep.

And this is the part of the work I care about most.

What’s in it for you

Done properly, AI compounds. It takes on the work humans shouldn't be doing. It finds patterns in your data you'd never go looking for. It shortens the gap between a question and an answer, and that gap's where most of your week goes. A small advantage, every week, and it adds up.

Done badly, it's just an expensive distraction. Tools nobody opens. Outputs nobody trusts. A process change that put more friction in than it took out, and that's the one I see most, actually. Money spent on AI because everyone else is spending money on AI.

And here's the bit that really annoys me. Everyone will happily tell you what AI can do for you. How many of them will tell you what it can't? So there's a bit of a double-dip going on there. And I suppose that's most of the reason I do this work.

How I think about AI

So the whole idea is AI has to fix a real problem, one you can put a dollar figure on, inside a process you've already got. At the end of the day that's the test.

A few principles.

  • I'd start with the problem. What's slow? What's manual? What gets done twice? What's sitting in the data that nobody ever looks at? I'd answer those first. AI's often the answer. Sometimes it's a spreadsheet and a better process.
  • I'd want a number against everything. Worth $20K a year. Worth $80K a year. If it hasn't got a number attached, I wouldn't build it.
  • I'd automate the boring stuff and keep the judgement. AI's very good at the stuff that's got rules around it. It isn't going to make the call on your brand, your customer, or your next move. Nor should it.
  • I'd build it inside the business. AI living in a separate tool nobody opens is just dead money. AI running inside the workflow the team already uses actually gets used.
  • And I'd give each one ninety days. Treat every one of them like an experiment. If it's paying for itself, it stays. If it isn't, out it goes.

Now look, I'm by no means an expert on this stuff. It moves every week and I'm learning it every day, same as everyone else. What I have got is twenty years of knowing which problems in an ecommerce business are actually worth solving. And what I always find is the principles are always exactly the same.

What working with me looks like

So where do you start? Most of this starts with an AI Diagnosis. Fixed scope, fixed price. I map how the business runs now, find where the time and the money are going, and cost each opportunity in dollars per year.

And what you get back is an action plan where every line's got a number against it. Here's the $40K a year win. Here's the $200K mistake you're about to make. Here are the three things everyone's telling you to buy that I'd ignore.

So then implementation runs as its own engagement. Embedded, in there doing it, built into the systems the team already uses. Measured. Killing what doesn't work, keeping what does.

And it goes in with the operations, the CRO, the forecasting, the platform work. It's all one thing. Change something here and it'll change something over there. It doesn't need its own sign over the door.

But before we go any further, the fundamentals underneath all this goes back twenty years. The dollar-costed AI case studies are newer, and I'm building them right now inside live client work.

Who it's for

Premium retail brands, $3M to $20M. And people who want someone independent in the room helping them work out what's worth doing and what isn't. Founders tired of being sold AI as the answer to everything by people who can't put a value on it. And brands who'd rather build something small that keeps paying than something loud.

Who it isn't for

Anyone wanting AI bolted on quickly. Brands who want AI as marketing material. And anyone hoping AI will save a business that hasn't been designed properly yet. It won't. It'll just get the mess to you faster.

Let's have a look at where AI helps

If you're trying to work out what to do about AI and you're not getting a straight answer from anyone, get in touch. The first conversation's free.

It usually involves walking through how the business runs now and pulling out two or three places I'd want to look at properly.

But before you spend a dollar with anybody, mine included, go and ask them one question. What's this worth to my business per year, and how will we know? If they can't answer it, you've learned something useful for nothing.

Contact me

Key points

  • Fixed-scope, fixed-price AI Diagnosis, with every opportunity costed per year
  • Implementation runs as its own separate engagement
  • Built into the workflows the team already uses
  • Ninety days on everything that goes in. If it's paying for itself it stays
  • Twenty years of knowing what's worth solving underneath it, and honest about what's still being proven