Think of a tailor cutting a suit. The questions, the measuring, the cloth, the cutting, the fitting, the adjustments. You can follow every step of it — and at no point does anyone expect you to start cutting your own cloth.
That’s the arrangement here. I find the work AI can genuinely carry in your business and I fit it there, explaining every step as I go. By the end, AI is doing real work for you, and nothing in the getting there was a mystery.
Teaching yourself AI is no longer the only way in.
The claims and the fearmongering are everywhere. So are the off-the-peg tools. And the courses, which take your problem and hand it straight back to you with homework attached. What remains rare is the useful version: AI quietly and competently carrying part of a real business. The technology has been ready for a while now. What isn’t ready, in most companies, is the description of the company itself — and that turns out to be the harder half.
That is what I do, and it’s all I do.
What this looks like in practice
Every company is different, so the work is different. But it tends to look mundane, which is rather the point.
Chasing the invoices nobody has time to chase. Turning an enquiry into a quote without someone retyping it. Assembling the weekly numbers before the Monday meeting instead of during it. Onboarding a new client without the same eleven emails. Reading the long documents nobody reads and telling you what’s in them.
None of it is glamorous. All of it is hours, every week, that somebody good is currently spending on work beneath them.
Each one sounds simple. Each one turns out to rest on a question only your company can answer — what counts as overdue, which enquiries are worth quoting, when a client is properly onboarded.
What has to exist first
on·tol·o·gy
/ɒnˈtɒlədʒi/ · noun
1. the branch of metaphysics concerned with the nature of being.
2. a set of concepts and categories in a subject area or domain, showing their properties and the relations between them.
ORIGIN — EARLY 18TH CENTURY, FROM MODERN LATIN ONTOLOGIA, FROM GREEK ŌN, ONT- ‘BEING’, PLUS -LOGIA
The second sense is the one that matters here, and it is worth knowing. An ontology is a written account of what exists in a particular world and how those things relate to one another. Your business is such a world.
AI, left to itself, guesses. That is precisely what it does — it predicts what ought to come next, fluently enough that a wrong answer sounds exactly like a right one. Perfectly good when you are drafting a letter. Not good when it is approving a payment.
But it does not guess evenly, and this is the part worth understanding. On everything the world has already written down — the law, the language, the textbooks, the whole public internet — AI is now genuinely expert, and that expertise costs almost nothing. Step outside it, into what is true only of your company, and the same system drops from expert to something nearer a bright child. It keeps the confident voice the whole way down.
Which is where the value went. Knowledge that is common has become close to free. Knowledge that is yours alone has become the scarce thing — and it is precisely the thing AI is worst at, because nobody ever wrote it down.
Businesses do not run on what is likely. They run on definitions. What counts as a customer. When an enquiry becomes a quote. Which discount needs a second signature. What ‘overdue’ means here, which is not what it means at the company down the road. Every business has its answers, and in most of them the answers have never been written down anywhere — they live in the heads of whoever has been there longest.
So before AI can carry the work, somebody has to write them down. Not the technology’s job, and not something an off-the-peg tool can do for you, because the answers are yours and nobody else’s.
You cannot automate what you cannot articulate. The articulating is, more or less, the whole job — and it is the part almost nobody sells you.
How I work
No two engagements are the same, but the shape stays the same. The flow and order is critical. It's the aspect of the process that generic will never give you.
1The opening conversation
We talk. You tell me about the business, I tell you honestly whether I think AI can do useful work in it. No proprietary detail, nothing to sign, nothing to pay.
2The NDA
Before you share anything proprietary, I sign a non-disclosure agreement you control. The work ahead depends on your willingness to show me how your business actually runs, so your comfort with that comes first.
3The scope
We agree, in writing, exactly what the project covers, what it costs, and how long it lives. Details will evolve — much of what gets built emerges from Discovery — but you’ll know what you’re buying before we start.
4The discovery
Concentrated, immersive and thorough: what you do, in what order, under what rules, where the opportunity is, where the pain is, and which work can move onto AI’s shoulders. It ends in a full written report — your company, set down properly. Everything after this stands on it.your company, not for an average of companies. Accompanied by a full report.
5The build
What gets built follows from what Discovery found and sits on top of it: your company’s way of working, done by your company’s tools, held to the definitions you agreed. Explained before, during and after, so by the end you know what you own, what it does, and why.
6The retainer
Businesses move. Definitions change, people leave, a new product doesn’t fit the old categories — and an account of a company that nobody maintains quietly stops matching the company. Because AI fails politely rather than loudly, that can go unnoticed for months. I stay on to keep it true, and to keep what I built effective as your company and the technology move. The NDA stays in force.
My story
I spent thirty-five years in brand and positioning for global corporations — the long discipline of working out what a business really is, and what it’s worth to the people it serves.
Then AI arrived in my industry as a working force, and what struck me wasn’t the disruption. It was the relief. Here was a way to take the repetitive work off people’s desks and leave them doing the thing they’re actually good at. I saw it in my own work first, and I couldn’t unsee it.
The people around me weren’t feeling relief, though. Owners and executives, genuinely good at what they do, quietly worried about falling behind — because keeping up had become a full-time job nobody hired them to do. The limiting factor in all of this is no longer which AI system you use. Those are close to interchangeable, and they improve without anyone’s help. The limiting factor is whether a company can say clearly what it does, what its terms mean, and where its lines are. That is not an engineering problem. It is the problem I have spent thirty-five years solving, under an older name.
That’s what The Context Forge is: a long career spent understanding businesses, applied to the newest problem they have.
— Trevor Gilchrist
Start with Discovery
You don’t have to commit to the whole thing to begin.
Discovery is available on its own, at a fixed price. I spend concentrated time inside your business and give you a written report: your company set down properly — what you do, in what order, under what rules — and then where the repetitive load sits, what could move onto AI’s shoulders, what it would take, in what order, and what it’s worth to you.
What that report describes is the one thing about your company that cannot be bought, copied or downloaded. Every competitor can hire the same AI you can, on the same terms, tomorrow. None of them can hire your company’s own account of itself. It is closer to a thumbprint than to a document, and it is the part of all this that holds its value.
So it is worth having even if nothing ever gets built. It’s yours either way, whether you go on with me or with anyone else. Most people go on to the build. Some take the report and act on it themselves. Both are legitimate.
And sometimes the honest finding is that there’s little here worth building yet. If that’s what Discovery shows, you’ll have it in writing with the reasoning, and we’ll stop.
A clear no is worth more than an expensive maybe.
Whether this fits
This suits an established company with work that repeats, and people too valuable to be spending their weeks on it. It helps if you can stand a fortnight of being asked why you do things the way you do them.
It doesn’t suit a business hoping AI will take over a function it hasn’t yet defined — that is the wrong way round, and no amount of technology fixes it. Nor one looking for the cheapest available version of this. Off-the-peg exists, and for some companies it is genuinely fine. What I do is the other thing.
Where your information goes
Three fair questions, all asked less often than they should be.
Does what I show you end up inside somebody’s model?
No. Your material is used to build your system and nothing else. Nothing you give me is shared with another client, and the NDA you control covers all of it. When an engagement ends, everything can be returned or destroyed — your choice.
How would I know if it were quietly getting things wrong?
Because it runs on definitions you gave me and signed off, because it is held to what you said it may and may not do, and because it shows its working. AI that is wrong loudly is an annoyance. AI that is wrong quietly is the real risk, and it is the one this is built against.
If that account of my company is so valuable, does building it hand it to you?
The sharpest version of the question, and the fair answer is that it doesn’t. What gets written down is a description of your business, made for your business, and it stays in your systems under your control. I don’t keep a copy to reuse, I don’t carry what I learn into another company’s work, and the NDA you hold covers every part of it. You can have everything returned or destroyed whenever you choose. It is your thumbprint. I am the person who helps you take it.
Where I am
I’m based in Los Angeles and work in person wherever that’s practical. Further afield we can do it properly over video, but I’d rather sit in your building at least once. You learn things about a company by being inside it — usually the things nobody thought to mention.
What it costs
DISCOVERY, ON ITS OWN — [DISCOVERY FEE]
BUILD ENGAGEMENTS BEGIN AT — [BUILD FROM]
ONGOING RETAINER — [RETAINER]
You’ll have the number in the first conversation, rather than having to ask three times.
A fifteen minute chat. Pick a time that suits you:
Telephone
Booking asks for your name and email so the meeting can reach your calendar — that is all they are used for. Never sold, never shared.