How we work

AI statement

Last updated September 2026

We use AI in marketing every day, on most of what we make, for almost every client we have. A person is accountable for every line of it.

That is the policy. The rest of this page is the detail, because clients ask, and we would rather answer it once, in public, than a different way on every call.

Where it sits in the work

  • Research and synthesis. Reading a market, a competitor set, a stack of call transcripts, a year of analytics. AI is fast at the reading. We are the ones who decide what the finding means and whether it is worth acting on.
  • Drafting. First cuts of documents, decks, copy and code. Most of what we hand you started as a draft AI helped produce. None of it reaches you in that state.
  • Build work. Sites, automations, internal tools, the plumbing behind a marketing function. This is where our output has changed the most, and it is the easiest place to check, because the thing either runs or it doesn't.
  • Analysis. Modeling and forecasting, and reading performance data for the patterns worth chasing.
  • Running TRCH. Notes, summaries, scheduling, the internal admin that used to eat a day a week.

What a person always does

Every deliverable goes through a team review and a tone pass, and nothing leaves the building without executive approval.

No claim, number, date or source goes out without a human checking it against the original. AI gets facts wrong with total confidence, and that is the failure mode that would cost you money, so it is the one we spend the most time on.

What we don't give it

No strategic recommendation. No positioning or brand language that carries your name. No judgment about a person's performance. And no decision you are paying us to make, because the judgment is the thing you hired, and a machine that has never sat in a room with your customers cannot make it.

Your information

Client material goes into these tools. Briefs, call transcripts, analytics, financials, whatever the work needs to be any good.

The tools we use for client work run on paid business accounts. The content we put through them is not used to train the underlying models.

Your NDA covers it. The confidentiality we owe you applies to every part of how we work, and adding a drafting tool to that chain does not loosen it.

We do not offer an opt-out. AI is part of how this firm works now, the same way a laptop is, and a client who needs a guarantee that no machine ever touched their file is a client we would be wrong for. We would rather put that here than have you find out in month three.

What it costs you

AI has made us faster. We did not cut our prices when it did, and we have not raised them either.

Here is why. The efficiency is table stakes in our category now, and the cost of using and learning these tools sits on our side of the ledger. What you are paying for is the judgment layer, and that hasn't gotten cheaper.

What we have actually seen from AI in marketing

We reviewed more than 20 client engagements at the twelve-month mark of AI being in their workflows. Every one of them uses it daily and pays for it properly. All of them say the work got faster and better. Not one could point to revenue it directly produced.

Both of those are true at once, and you should hear it from us before you budget for it.

If we are not willing to put this in writing, what exactly are we asking you to trust?