A claimant fills your form. Three AI models score it in parallel. Your attorney sees a clean dossier with the consensus score, variance, and citations — before the prospect refreshes the thank-you page.
Branching logic adapts to the practice area. Eight rubrics, dynamic follow-ups, document upload — all white-labelled to your firm and embedded on your domain or hosted on ours. PHI is classified at intake and routed through pre-signed S3 before it touches an app server.
Claude Sonnet, GPT-5.2 and Gemini each return a 0–100 score with cited evidence — strengths, watch-items, statute checks. We weight the consensus by confidence and flag anything with variance above σ 15 for an attorney, never auto-decided. Every score carries its receipts.
Strengths, watch-items, statute citations, damages estimate, and the routing decision land in the right intake queue before the prospect refreshes the thank-you page. Accept and the matter syncs to Clio with a Cal.com booking link in the auto-reply. Decline, with a reason.
