Most shop owners get a monthly P&L from their accountant and a dashboard from their software. Neither tells them what actually happened.
The monthly story report does. It's a short narrative recap of your shop's month — written by Claude from your own numbers, on demand, whenever you want to read one.
What's in it
It's a tight recap — a few sentences, not a report — comparing this month against last month on:
- Revenue, both collected and still outstanding
- Appointment counts, broken out by status
- Quote conversion rate
- New vs. returning customer mix
- Your top service category by revenue
The value is in the deltas being called out in plain language, with the direction stated rather than left for you to work out from a table. Don't expect a full operational review: it's this month vs. last month on those five things, not a seasonality analysis, a bay-utilization audit, or a 30-day action plan.
How it works
Claude reads those numbers straight from your shop's own records and writes the paragraph. Every figure it cites comes from the same data behind your other reports, so if a number surprises you, it's drillable — open the matching report and the detail is there.
How to run one
Open Reports and look for the Story-mode recap card on the dashboard view, then click Generate. It takes a few seconds to read your numbers and write the paragraph.
There's nothing to configure — no toggle, no schedule, no recipient list, and it isn't emailed anywhere. It's a button you press when you want it. Reading it on the 1st of the month is a good habit, but that habit is yours to keep; nothing fires on its own.
Two things gate the card: your plan has to include the AI report features, and your account needs permission to view financial reports. If you don't see the card at all, it's one of those two rather than a bug.
Privacy
No customer records leave your shop for this. SalesThumb computes the totals first — revenue, counts, rates, your top category — and sends only those finished numbers to be narrated. Customer names, vehicle VINs, and phone numbers are never part of that payload, so there's nothing to anonymize. It also means the AI can't invent a figure: it's writing sentences around arithmetic that was already done. We never train models on your data.