Expert matching is the core capability of the platform. When a request comes in, the system runs a multi-stage pipeline to find, score, and rank the best expert witnesses from the Salesforce-synced database.
The matching pipeline
Stage 1: Hard filters
Every search starts by eliminating experts who don't meet non-negotiable criteria:
Filter | Requirement |
|---|---|
Active status | Expert must be active in the system |
Expert witness | Must be flagged as an expert witness |
Geography | Must be within 75 miles of the case location (see The 75-Mile Rule article) |
Stage 2: Candidate retrieval
After filtering, the system retrieves up to 500 candidates for detailed scoring. This cap keeps response times fast without sacrificing match quality.
Stage 3: Candidate Matching
Each candidate is scored using three independent signals, keyword matching, semantic similarity, and structured field comparisons, which are fused into a single ranking using Reciprocal Rank Fusion (RRF). Business rules then apply multipliers for factors like subspecialty alignment, exclusivity, and testimony experience.
The top 20 results are returned, each with a score tier (Excellent, Good, Fair, or Low) so you can quickly prioritize your review.
For a detailed breakdown of how scores are calculated, see Scoring & Ranking.