I did not set out to build a productised service this year. It came out of noticing that I kept doing the same three things, in the same order, for businesses that had very little in common — different sectors, different sizes, and very different levels of comfort with AI. Eventually the shape was obvious enough that I started planning for it rather than arriving at it.
Understand
Before anything technical, I need to know what the business is actually trying to do — the real strategy rather than the mission statement version, which means what they sell, who they sell it to, and where the margin sits.
Then the layer underneath, which is how the work really happens. Who does what. Which processes are written down and which live in someone's head. Where the time goes.
Those two pictures rarely match. The gap between them is usually where the opportunity is.
Analyse
Once I can see how a business creates value, I can work out where technology changes that equation. Knowing the tools matters here, but this is mostly not a technology question. It is a question of which changes are worth the disruption, which ones the team will actually adopt, and which ones look good in a demo and quietly die in month two.
What comes out is a ranked view with reasons attached, rather than a list of everything that is possible.
Decide
Then the business decides, and the honest answer is that it goes three different ways.
Sometimes I build what I recommended. Sometimes we build it together, with their team learning as we go so they can carry it afterwards. Sometimes they take the whole thing in house and run with it themselves. All three happened this year, and all three were the right call for the business that made them.
The part that created the value
Looking back, the technology was rarely the hard part.
What made the difference was knowing what to focus on. The businesses I work with are usually unsure which AI ideas are actually worth pursuing, and what they need is a defensible view on which ones are worth money, in what order, and why. Just as often, the piece they are missing is the plumbing underneath — the context layer that lets AI produce consistent, usable output without endless prompting back and forth and a slow slide into an iteration death spiral.
The other half of the work is helping people frame it, and that is a deliberate part of what I offer rather than a side effect. Designing processes and ways of working whose value to the business is obvious. Explaining to a team why a change is happening and what it means for their job. Explaining to clients why the firm is investing in this and what they get from it. None of that is a technical deliverable, and all of it decides whether the work survives contact with a busy month.
Why package it
Once you have run the same sequence several times and it has held up each time, treating it as a bespoke engagement starts to look like waste. The sequence is stable. The questions are mostly the same. What differs is the answers, which is how it should be.
Packaging it does two useful things. The price and the timeframe become knowable before we start, which is a real problem for anyone buying consulting and rarely solved. And it forces me to be specific about scope, because a fixed price with a vague scope is a bad idea for both sides.
That is what the Service Business AI Gameplan is. Understand, analyse, decide. Fixed price, three to six weeks, ending in a decision the business owns either way.