All your marketing data in one database — and an AI that works directly on it: as a chat for you and your clients, as a weekly report with the numbers behind every claim, as an ad briefing built from real conversions.
The feature clients tell other people about: their report link answers questions.
Because clicks, purchases, calls and ad spend all sit in one database, an AI can answer straight from it. As a chat in the dashboard — and even inside the shareable client report link, with no login and no training.
Example dialogue — tap a question
Google Ads is at 4.2×, Microsoft Ads at 3.2×, Meta at 2.5×. Google carries the largest share of revenue; Meta is below average but above break-even.
Evidence: Ad cost per channel (7 days) · revenue per channel from your own orders
Of 38 calls, 21 came from paid clicks, 14 of them longer than 30 seconds. The other 17 were organic or direct.
Evidence: Call events with source · threshold for qualified calls
Four items produced clicks in 30 days and sold nothing. Two others sell well but barely get budget.
Evidence: Ad cost per product · order lines per product
A scripted example with example figures, so you can see the shape of an answer. In the product it runs on your own data, and every answer opens down to the raw query.
Search Console tells you what Google shows your pages for. We link that to your measured enquiries and purchases: measured per page, projected per search term by its share of clicks. Ask in the chat: “Which search terms make me money?”
The AI reads the weekly aggregates of your measurement data and writes a short report: what moved, where revenue is being left on the table, what to check — every statement with the number behind it. Including the conversions platform dashboards never see: calls, order values added later, recovered attributions.
One click produces copy drafts for Google search ads, Meta and TikTok — grounded in what demonstrably converts for you: revenue per channel, lead mix, call behavior. No generic text AI can do that, because it never sees your conversions.
The drafts come out of your real signals — and the platforms' character limits are enforced hard on the server.
On request the system puts the draft straight into the ad account — always paused. It goes live by hand, never automatically. Changes to the ad account from the chat work the same way: Steering via AI chat →
Via the MCP standard you connect Claude or ChatGPT directly: campaign performance, return report, Google search data, business profile — one question instead of five dashboards.
Account analysis, campaign numbers, a tracking reconciliation and recommendations — laid out the way a managing director expects it: executive summary, clear tables, an honest platform-vs-reality comparison. In German and English, in a corporate look, without anyone having to open a dashboard.
For context: the AI weekly report and the data chat are already included in tracking from €299/month. The management report as a PDF belongs to the Copilot tier from €749/month — together with the account analysis and prepared implementation. What's in the Copilot →
Executive summary with a verdict, a setup check of the ad account, placement and audience numbers, creative ranking — recommendations sorted by leverage, with the expected effect in euros.
The section no ad dashboard can deliver: what the platform claims for itself, next to what the shop actually booked — with an honest lower and upper bound instead of a single flattering number.
All numbers are read out of the ad APIs and your first-party database — nothing in the account is changed. Every number in the report names its source; the methodology is in the footer.
Bring a real question — we'll ask it live against a running account during the call.
Book a demoBring a real question about your marketing data — we'll answer it live in the call, by chat, against a running account.