Automation is a commodity. Judgement is not.
Last week, on 12 June, we put something into production that we’d been building for the better part of a year. We call it the Performance Engine, and it answers a question every media and communication leader is wrestling with right now: if AI can run a campaign end to end, what exactly are people still for?
Our answer, after testing it rather than debating it, is that people are for the part that matters most, and the Performance Engine exists to make that part scale.
The market has moved past the hype. AI can now set up, run and optimise a campaign with very little hand-holding. That’s real, and it’s not slowing down. But “capable of execution” is not the same as “makes the right call.”
We proved this to ourselves with a controlled experiment. Two identical instances of the same AI model, same brief word for word, same budget, same live campaign driving donations for a humanitarian project. The only difference: one ran on its own, the other had our client context, our methodology and a human analyst validating its moves.
After seven days, the standard instance had spent its budget with zero conversions — and then blamed the landing page, exactly as a weak agency would. The human-infused instance delivered a return close to three times spend. Same model. Same brief. The entire gap was context, expertise and human judgment.
That’s the thesis in one number. Automation is becoming a commodity: anyone can spin one up. Compounded judgment is the scarce thing.
The Performance Engine is how we make that judgment structured, available, self-improving and at scale.
The core idea: amplification of human expertise.
The problem with expertise is that it’s trapped. It lives in people’s heads, in old decks, in one analyst’s memory of a campaign from two years ago. When that person is in a meeting, on holiday or gone, the expertise goes with them.
The Performance Engine takes that expertise and makes it structured, queryable and reusable — available at the moment of every decision, on every account. Concretely, it puts human know-how to work in four directions at once:
- Cross-account — patterns from similar challenges inform yours, with evidence rather than copy-paste.
- Cross-platform — one unified view across Google, Meta, TikTok and programmatic.
- Cross-time — an insight from two years ago resurfaces the moment a new brief resembles it.
- Cross-brain — human and machine challenging each other, with every decision encoded.
The machine proposes; a human validates; the decision is logged, the action is executed by an agent. Nothing goes live without a person accountable for it.
Why it gets stronger every time a human touches it
This is the part we care about most.
The Performance Engine compounds. Every time an analyst corrects a recommendation, overrides a call or adds context, that decision is captured and fed back into the system. The gap between what the machine proposed and what the human decided is the single most valuable input we have. The Engine won’t repeat the mistake — not on this account, and not on any similar one.
So the more people use it, the sharper it gets. Human expertise sits at the centre, and the system amplifies it rather than replacing it. That’s the flywheel.
Responsible by design
For a media or comms team, “powerful AI on our data” raises an immediate question: is this safe and is it compliant? We built the answer in from the start.
The infrastructure runs on enterprise cloud with European hosting, fully under control. Data never trains third-party models. No personal data enters the system by default. Client contexts are fully isolated from one another. There is human judgment on every decision, with two validators required for high-impact changes — no action without validation.
Two ways to get it
We’re offering the Performance Engine in two models.
As part of our service. If you work with CLICKTRUST, the Engine will sit behind the work — senior judgment applied to every account, not just the ones with a senior name on them, with faster and more consistent execution across every channel. The first clients are being onboarded now.
As your own platform. For organisations with internal media teams, we build and implement the entire Engine on your premises, on your stack, fully owned by you. You decide who runs it — our operators, your people, or both. You get the architecture, the methodology and the compounding expertise layer as an asset that stays yours. We help you setup and maintain it for as long as you need.
Where we are now
The Performance Engine went live on 12 June and is rolling out to our first two clients.
“For years, the honest answer to ‘why hire an agency’ has been judgment — the kind you only earn by running hundreds of campaigns across dozens of verticals. AI doesn’t change that answer; it just makes it louder. The Performance Engine takes the expertise that used to be trapped in people’s heads and old decks and makes it available at the moment of every decision. The machine proposes, a human decides, and every one of those decisions makes the whole system sharper. We’re not automating our people away — we’re amplifying them.”
Fabian Van De Wiele, CEO, CLICKTRUST
If you’re rethinking how your media operation works in the age of AI — whether you want it run for you or built for you to own — this is the conversation we’d like to have.
