AI connector · any desk
Data Doughnut inside your AI tool
A hospital strategy analyst
A hospital strategy analyst asks their AI assistant one question and gets back a board-ready memo. The memo mixes public data with the hospital's own numbers, and every public figure links to the page it came from.
The question
The board wants to know whether to add a cardiology clinic in a neighboring county. The analyst has the hospital's own discharge spreadsheet and last year's market study. They don't have the public picture. They connect the Data Doughnut connector to Claude once. From then on Claude can use the desks the account pays for.
The work
The analyst attaches the spreadsheet and the study, and asks: “Should we open a cardiology clinic in the next county over? Use our discharge data and the public data.” Claude:
- get_health_market and chronic_disease_report. Heart disease rates, Medicare use, and who already provides care in the county.
- search_care_providers. Nearby cardiology practices and hospitals.
- compare_places. The county side by side with the hospital's home county.
- The analyst's own files. How many of the hospital's heart patients already drive in from that county.
The result
A two-page memo covering demand, competition and the hospital's own patients, ending in a recommendation. Every public number links to a datadoughnut.com page, so the board can check it.
Why it matters
The analyst never learned a new website or exported a file. The connector keeps agency facts and our own estimates apart. If the account's plan doesn't include something, it says so instead of guessing.
See it with real data: the free Polk County, Iowa sample, or what's inside each desk. Every plan starts with a 7-day trial.
The people, places and numbers in these stories are made up to show how the pages work. The pages and the public data behind them are real.
Next case study: From “we need a truck” to a filed grant →