CASE STUDY
Rincon
AI-augmented finance intelligence for a multi-location clinic group
THE SITUATION
Off-the-shelf BI had become the bottleneck
Rincon was running on Tableau for reporting and Windsor for the data plumbing underneath. Each new question from leadership took days to answer: pull from QuickBooks, massage the spreadsheet, rebuild a chart, field the inevitable follow-up.
The team had outgrown the tools. Dashboards meant to save time had turned into maintenance overhead with every close.
BEFORE
Days to answer a question from leadership. Every close stretched by spreadsheet maintenance.
AFTER
The finance team answers in a click. Leadership asks in plain English.
WHAT I BUILT
A finance intelligence layer shaped around how the team actually works
I replaced the Tableau plus Windsor stack with a custom finance platform. Direct QuickBooks integration, typed schema, version-controlled dashboards, and the specific cuts the finance team asks for most often wired straight into the home view.
AI sits where it earns its place: anomaly detection on cash and receivables, plain-English queries against the warehouse for the questions leadership actually asks. Production from week one, weekly checkpoints with the finance lead.
QuickBooks ingestion
Reliable sync with retries, dedup, and an audit trail finance can actually trust.
Purpose-built views
Cash position, receivables aging, vendor spend by class. The views the team rebuilt in Excel every month are now one click.
Plain-English queries
An LLM layer on top of the warehouse. Leadership asks questions the way they think them, not by clicking through a filter UI.
THE OUTCOMES
Reporting stopped being a project
0
Tableau or Windsor licenses still in use
1
source of truth across the clinic group
Daily
executive view leadership runs on every morning
STACK
- Next.js
- TypeScript
- Postgres
- QuickBooks API
- Claude
- Railway
