AI Provider Match
NLP provider search that replaced brittle filter logic with meaning.
Overview
An NLP provider search across clinical networks that indexes 80+ variables per provider, replacing deterministic search and filter logic with semantic matching.
The problem
Deterministic filters missed the right provider whenever the query did not match the taxonomy exactly.
Approach
Replaced rigid filters with semantic matching over a richer provider index.
- Indexed 80+ variables per provider.
- Modeled queries by meaning, not keywords.
- Tuned against real referral outcomes.
Results
Lifted match rates by roughly 80% against the previous deterministic approach.
~80%Match-rate lift
80+Variables indexed
Stack
NLPPythonVector search