Stop summarising dashboards. Start investigating.
Most AI prompts just describe your dashboard back to you. These ten do the harder job. They help you find the real story in your data. Isolate the signal, compare it against a baseline, test what might be causing a change, then decide what to do next. Paired with Matomo’s AI Connector, you run them in plain language from a chat window, with no BI team in the loop.
The guide gives you ten prompt patterns drawn from real Matomo teams. Copy one, swap in your own site, dates and dimensions, and get a decision-ready answer in minutes.
Get the 10 prompts and run the first one on your own data today:
From “what happened” to “what do I do”
A standard AI prompt stops at a summary. The prompts in this guide push further. Each one is an investigation pattern that breaks a question down, weighs the evidence for and against each cause, and ends with a next step you can act on. You get the why behind the number, not just the number.
What you can do with the ten prompts
Each prompt targets a real job analytics teams face every week:
- Diagnose a conversion drop, broken down by country, source and segment
- Audit and grade your data quality, with the business impact of each risk
- Cross Matomo data with an external source, no BI pipeline needed
- Build segments, goals and custom reports tied to real business questions
- Separate a real business issue from seasonality, with a confidence level
Plus five more, each with a note on when to use it and why it works.
Built for people who keep judgement in the loop
Some prompts read your data. Others can create a segment, a goal or an annotation once you turn on write access. Nothing changes until you approve it. You review, you decide, and the AI simply gets you there faster. As Matomo recommends, check important numbers in Matomo before you share them, because AI tools can still get things wrong.
Before you start
These prompts assume Matomo’s AI Connector (MCP) is connected. On Matomo Cloud it is included and a Superuser switches it on. Pair it with a capable, recent model such as Claude or Codex, since these prompts rely on multi-step reasoning that lighter models struggle with. New to the AI Connector? Start with the beginner MCP guide, then come back here.
Get the guide and turn your next analytics question into a decision.
📚 Learn more
• Set up the Matomo AI Connector (MCP): https://matomo.org/guide/apis/mcp-model-context-protocol/
• Why teams use Matomo MCP: https://matomo.org/blog/2026/05/matomo-mcp-ai-now-available/