From Google Sheet to Live Dashboard in Under Two Minutes
There is a spreadsheet in your organization that should be a dashboard. It gets updated every Monday morning by someone on the ops team. Someone exports it to CSV, uploads it to a BI tool, and produces a chart that is already out of date by the time it lands in the meeting.
Phase 3 of AI Insights 365 eliminates that workflow entirely.
The two-minute connection
- Go to Settings → Datasources → New datasource and set type to Google Sheets.
- Click Connect with Google — approve two read-only permissions.
- Pick the spreadsheet from the dropdown, pick the tab, and click Test connection.
That is it. You now have a live datasource. Your Google password is never seen or stored — AI Insights holds an encrypted OAuth token with read-only scope.
What "live" actually means
Every time a chart renders — page load, manual refresh, scheduled email, live refresh interval — AI Insights calls the Google Sheets API and reads the current cell values. There is no cache older than 60 seconds. If a colleague updates a number in the sheet, the chart shows the new number within a minute.
Power BI requires a scheduled gateway refresh with a minimum 15-minute interval (longer on most plans). AI Insights has no concept of a scheduled refresh for Google Sheets because the data is always current.
The AI layer kicks in automatically
Every chart built on a Google Sheet gets:
- AI narrative — 2–4 bullet-point observations generated after each render
- Anomaly detection — Z-score analysis flags statistical outliers
- Forecast overlay — one checkbox extends the chart three periods forward
- Executive summary — the entire dashboard summarised in 3–5 sentences
None of these require extra configuration.
The organizational pattern that makes this powerful
The highest-value use is not replacing a SQL datasource — it is augmenting one. One dashboard can pull from a SQL datasource (transactional data), a Google Sheet (finance team's quarterly forecasts), and a second Google Sheet (sales VP's monthly targets) simultaneously. When the finance team updates their forecast, the dashboard updates. No one needs to export anything.
That is the gap Phase 3 closes. Not a feature gap — a workflow gap.