An ad-management platform with a human in the loop

The challenge

Managing paid media well requires relentless, repetitive discipline: watching search terms, checking tracking health, spotting anomalies, keeping feeds clean — across multiple clients and multiple ad platforms. Done manually, the discipline slips. Done by black-box automation, trust does.

What we did

We built our own multi-client ad-management platform around a simple safety model: the system proposes, a human disposes. Each client gets a private warehouse fed by connectors to Google Ads, GA4, Search Console, Google Merchant Center and Microsoft Advertising, plus their own sales data.

On top of that data the platform runs audits — tracking health, account structure, search-term waste, product coverage gaps — and drafts change proposals with the supporting evidence attached. Nothing is applied until a human approves it. Approved changes are validated against the ad platform first, recorded in an append-only ledger, monitored daily, and formally evaluated after two weeks. That evaluation records what the account did afterwards — it is a check on the change, not proof the change caused the movement.

The outcome

Client accounts in retail and education now get the consistency of automation with full human accountability: a complete audit trail of what changed, why, who approved it, and whether it worked. The same platform produces the keyword research, competitor intelligence and strategy reviews clients see each month.