What it does
A reporting automation for a marketing agency with many retainer clients. On the first working day, each client's data is pulled from Meta Ads, Google Ads, Google Analytics and Search Console into one format. Every number is compared with last month, the same month last year and the client's targets. AI drafts the commentary: what moved, the likely reason, and what the agency will do next, using only those numbers. The account manager edits and approves, and the branded report goes out with a short email.
The problem
Each monthly report took most of a day per client: screenshots from four tools, numbers copied into slides, and commentary written from scratch. Reports went out late and looked different for every client.
Approach
The numbers are pulled and checked by code, never typed. The commentary may only cite figures that are in the report, and it marks anything unusual, such as tracking that stopped or spend that doubled. The account manager approves every report, and their edits are kept so next month's draft learns the client's tone.
- Input
- AI step
- Rules or code
- Person decides
- Result
Screens
Swipe sideways to see the whole screen.
Client work is shown under confidentiality agreements: no client names, and screens are recreated with sample data.
Stack
- Meta, Google Ads and Analytics APIs
- Python
- Gemini API
- Branded PDF reports
Impact
Typical for systems like this.
