AI Summaries Are Now in Metrics Watch
Metrics Watch now writes the executive summary for your client reports. It is live today: add the AI summary block to a report, and each run writes that paragraph from the report's own numbers and sends it inside the report email with everything else.
That is the paragraph at the top of every client report that used to be yours to write. Same client, same charts, new month, and the part the client actually reads was the part you retyped by hand.
What the AI writes
An executive summary for your client: what moved and by how much, written from the numbers in the report, with a few highlights beneath it. It is a block you place on the report like any chart, so you decide where it sits.
Email addresses, phone numbers, and similar personal details are stripped out of the data before it reaches the model.
What it will not say
A connector that came back empty and a month with genuinely zero spend look identical in a spreadsheet, and they mean opposite things to a client. The summary treats them differently. A source that returned nothing is named once, plainly, and left out. A real zero is reported as a number.
Every summary is checked for wording that breaks that rule before it goes out. If it fails, it goes back to be rewritten once. If the rewrite fails too, the AI paragraph is dropped and the report goes out without a summary. A missing paragraph is a smaller problem than a wrong one.
How a report gets made
- Connect your sources. Google Analytics 4, Google Ads, Search Console, Facebook and the rest of the sources on our integrations page.
- Build the report. Pick the metrics and charts, then add the AI summary block where you want it.
- Schedule it. Each run pulls that period's numbers, and the summary is written from them.
- Preview it before it goes out. If you want to read a report before your client does, it can wait for you.
- Your client gets the whole report in the email. The summary, every chart and every table are in the email itself, not behind a link. If you want one, a footer link lets them view it in a browser or download it as a PDF.
How it compares with AgencyAnalytics, Whatagraph, and DashThis
This comparison lives on the Metrics Watch blog, so you know whose product this is. The facts below come from each vendor's own pages, checked on 21 October 2026.
All three write AI summaries. AgencyAnalytics has an AI Summary widget you add to a report, which you can guide with your own prompt. Whatagraph generates summaries, wins, recommendations, and trends in 18 languages, for you to edit before sharing. DashThis has an AI Insights tab on the dashboard with four views: summary, opportunities, wins, and issues.
Where they win today is breadth. DashThis says its AI Insights work with all 30+ of its native integrations; we connect fewer. If your agency needs a source we do not connect, they may be the more complete home for it, and we would rather say so here than have you find it out later.
Where we differ is when the summary is written. AgencyAnalytics' help centre says the summary is generated when you add the widget, refreshes when someone changes the date range or opens the report, and on a scheduled report with auto-refresh on, refreshes the first time it is opened after it is sent. DashThis describes its insights as something you open in the dashboard, and says its AI only talks about data present in the dashboard and tells you when you ask about a period you have not loaded. In Metrics Watch the summary is written when the report runs, from that run's numbers, and it arrives in the email with the rest of the report.
On missing data, we could not find anything on AgencyAnalytics' or Whatagraph's AI feature pages about how the summary words a source that came back empty. AgencyAnalytics does notify your team when an integration disconnects.
Try it on your next report
If you already send reports with Metrics Watch, open a report, add the AI summary block where you want it, and preview the next run before it goes out. Questions or something that reads wrong: email Jon.
