Data
The Monday metrics email that somebody writes on Sunday night
The numbers pulled and the paragraph written, delivered as a file rather than a link.
What the worker does
Reporting automation with an AI worker covers the writing, not only the pulling. The worker gathers the figures through the Supabase and Stripe connections, writes the paragraph stating what moved and by how much against the previous period, attaches the report as a file, and posts it as a delivery comment on the task. It states movements. Claiming what caused them is a human addition.
- Runs
- Weekly or monthly
- Connections
- Supabase, Stripe, Google Drive
- Output
- A written report file
A dashboard link is not a report
Everyone has a dashboard. Almost nobody opens it, because opening it requires knowing which of the eleven numbers changed this week, and the dashboard is equally happy to show you all eleven whether or not any of them moved.
What people read is the paragraph. Two hundred words saying signups are up nine percent week over week, driven by one channel, and the conversion rate held. Writing that paragraph is the part that takes forty minutes and the part that gets skipped when the week is busy.
What separates a written report from a generated one
It says what moved, and by how much
Both the absolute and the relative change, against the same period last time and against the same period last year where the data goes back that far.
It says what did not move
A metric that has been flat for six weeks is a finding. Reports that only mention movement make stability invisible until it becomes a problem.
It stops before the causation
The worker reports that a number moved. Saying why it moved requires knowing about the campaign that launched on Tuesday, and unless it was briefed to look, it should not guess.
It arrives as a file
Deliveries in Polaris are comments that can carry generated documents, so the report is an artefact with a date on it rather than a link whose contents change under you.
Where the numbers come from
| Connection | Typical contribution |
|---|---|
| Supabase | Product usage, signups, activation, retention cohorts |
| Stripe | Revenue, refunds, failed payments, subscription movements |
| Google Drive | Where the finished report file is filed for the record |
| Gmail | Sending the report on, when it goes to people outside the workspace |
| Slack | Posting the summary where the team reads on Monday morning |
Why the recurring shape is cheap
The same report every week is the ideal shape for metered work. The searching is minimal because the sources are known, the prose is short, and the estimate lands near the bottom of the range. Every run is itemised on the worker's work log with the effort that produced it.
If a week's report costs more than usual, the log shows which part took the work, which is a more useful conversation than a line item on an invoice.
Questions people ask
+Can it email the report to people outside our workspace?
With the Gmail connection authorised, yes. Most teams keep the delivery inside Polaris as a comment with the file attached, and send it on themselves, because that keeps a human between the numbers and the recipient.
+Will it explain why a number changed?
Only if you brief it to look, and only from sources it can reach. A worker told to check the campaign calendar and the deploy history before writing can offer candidate explanations, and it should mark them as candidates rather than conclusions.
+What if the numbers are wrong?
Reject the delivery and say why in a reply. Nothing was published, because the report arrives as a comment on an open task, and the task stays open until a person closes and rates it.
+How is this different from a scheduled dashboard export?
An export ships the same table every week regardless of whether anything happened. This delivers prose that names what changed, what held and what is missing, which is what people actually read on a Monday morning.
Related
One analyst, thirty requests, and an AI worker in between
Five jobs that stand between a data team and the analysis they were hired to do.
The null rate tripled in March and the board slide was already wrong
Row counts, null rates, orphans and freshness, checked every morning. Silence means clean.
Three teams, three definitions of active user, one meeting
Every definition of a metric found and listed side by side, with the source of each.
The weekly review, written before the meeting starts
A meeting that begins with everyone reading the same document is a different meeting.
Connect Stripe to Polaris
This is the one connection where the key you choose matters more than anything on this page.
Connect Supabase to Polaris
Supabase is the connection where a careless key choice does the most damage, and the only one Polaris cannot yet check for you.
When the deliverable is a file
A chat reply is not a deliverable if the thing you needed was a document somebody can open.