How we work
The platforms will run the campaign. Someone still has to be right about it.
Our method
We interpret, steer and take accountability for what AI-driven marketing actually produces.
Interpret — Read what actually happened
Before anything else, we establish whether your numbers are real. Broken conversion tracking, double-counted events and self-reported platform attribution mean most businesses are optimising against fiction. Interpretation starts with verification, then works out what the movement in the numbers is caused by — which is rarely what the dashboard suggests.
Steer — Decide what changes
Automation is very good at optimising within the boundaries you set and completely indifferent to whether the boundaries are right. Steering is choosing the offer, the audience, the objective and the budget split — then changing them when the evidence says to. These are the decisions that determine whether the spend works, and no platform will make them for you.
Own it — Answer for the outcome
We agree what the programme is accountable for before it starts, and we report against that rather than against whatever looked good this month. When we get it wrong we say so early, with what we are changing. An agency that never reports a miss is an agency editing its own scorecard.
What changed, and why it matters to you
Ten years ago a competent agency earned its fee on execution. Building a well-structured campaign was genuinely difficult, and doing it badly cost you money. That is no longer where the difficulty is. The platforms now build, target and optimise competently by default, and they will do it for anyone who opens an account.
What automation has not absorbed is judgement. It will optimise relentlessly toward the objective you selected, using the conversion signal you gave it, for the audience you defined, against the offer you wrote. Every one of those is a human decision, every one of them determines whether the money works, and none of them are made by the algorithm.
So the value moved. It is no longer in doing the work — it is in deciding what the work should be, reading honestly whether it worked, and being answerable when it did not. That is a harder thing to sell than a list of deliverables, and it is the only thing worth paying an agency for now.
What accountability actually means here
It means agreeing the outcome before we start, reporting against that rather than against whatever metric flattered us this month, and telling you early when we are wrong. It also means saying no — to work we do not think will pay for itself, to budget increases the account cannot absorb, and to clients we cannot help.
The uncomfortable version: an agency that has never told you a campaign failed is an agency marking its own homework. We would rather have that conversation in month two than lose you in month nine.
Where AI fits
Two places, and they are different. Inside campaigns, AI is the thing doing the optimising — our job is to feed it real conversion data instead of platform defaults, and to notice when it is confidently wrong. Outside campaigns, AI is increasingly how your buyers form a shortlist before they ever search, which is a visibility problem rather than an advertising one. We cover both under AI optimisation.
Talk to a strategist
Let's find out what's actually holding you back
Fifteen minutes on the phone, an honest read on your situation, and a number. If we're not the right fit we'll say so and point you somewhere better.
