Every period somebody rebuilds the same report by hand out of systems that do not talk to each other. This is the brief that hands that job to Claude. Copy it, paste it into a new Claude conversation, and attach last week's exports. It asks what your numbers mean before it builds anything.
On one run, the mapping showed that two delivery platforms reported gross while the third reported net of commission. The weekly delivery number the group had reported for months was the sum of two incompatible things. That was caught before a single figure was produced.
The numbers live in four places that disagree with each other, and the only thing able to reconcile them was somebody who knew what every column in every export was really called. Claude reads those files directly, asks what your columns mean before it combines anything, and holds the whole job at once. The morning becomes a paste, and the person who knew what the columns meant writes it down once instead of doing it every week.
It is an ordinary Claude conversation. You paste the brief, send it, and drag your export files into the same window when it asks. Nothing gets installed and nothing connects to your systems.
What it does not do. It runs on exports you attach by hand, each time. Making the files arrive on their own, or sending the finished report somewhere, is a build rather than a brief.
A marketing lead was spending about forty-five minutes a week assembling the delivery-sales report by hand from the Uber Eats, DoorDash and Skip emails. Rebuilt this way it takes about ten.
Your AI department, on demand. We plan, build, and deploy your company's AI strategy.