Client case study · Healthcare analytics
Plain-English Answers From Pharmacy Data Across 50 Locations
Pharmacy staff ask questions in plain English and get answers from HIPAA-regulated data for an approximately $100M enterprise; every generated query is checked, authorized and audited before it runs.
The problem
Operators at an approximately $100M long-term-care pharmacy technology enterprise needed plain-English answers from operational data, without giving a language model open database access or losing HIPAA-grade auditability.
What was built
A natural-language analytics assistant in live use across 50 retail and private pharmacy locations and connected medication-management devices, alongside compliance guidance, safety alerts, inventory and demand-forecasting modules.
Results
- Staff get plain-English answers from operational pharmacy data without writing SQL.
- The model never gets open database access: every generated query is checked against the user’s permissions, validated and audited before it runs.
- User feedback goes to an administrator-reviewed correction loop, so answers improve under human review.
- The client expanded the work across four sequential engagements.