AI Ads-Analysis Dashboards
A web app that turns a keyword list and a client brief into a shareable, interactive ads-analysis dashboard, generated by an LLM.
A web app that produces a full ads-analysis dashboard from two inputs: a seed keyword list and a short client brief. The analysis itself is written by a language model, so the output is a finished report rather than a spreadsheet to interpret.
The problem
Turning raw keyword data and a brief into a clear, client-ready ads analysis is expert work, and it gets repeated for every account. The thinking is consistent enough to encode, but doing it by hand each time is slow.
What it does
- Takes a brief and keywords. A React front end collects the seed keywords and the client context.
- Runs the analysis. An Express backend feeds the inputs to Google Gemini, using a dedicated ads-analysis prompt as the system instruction.
- Returns a dashboard. The output is a self-contained HTML dashboard, stored per project and shareable through an unlisted URL.
The outcome
A consistent ads analysis on demand, packaged as something a client can actually open and read. It follows a pattern I reused from an earlier internal forecasting app, swapped to LLM-generated output instead of client-side calculation.