Skip to Main Content

VeriSight AI

Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

VeriSight AI — Demo video

VeriSight AI is a multimodal forensic and legal audit agent that helps investigators identify potential document fraud, verify evidence, and quantify contractual financial exposure. Traditional forensic and legal investigations often require extensive manual review of scanned documents, handwritten signatures, contractual clauses, and financial records. This process can be time-consuming and difficult to audit. VeriSight AI addresses this challenge through three integrated capabilities: 1. Multimodal Signature Forensics: Uses Meta SAM 3.1 to segment handwritten ink strokes and extract signature regions from scanned evidence. Meta Muse Spark 1.3 then analyzes visual characteristics such as stroke loops, slant, and rhythm to compare reference and questioned signatures. 2. Contractual Damages Quantification: An agentic financial tool calculates liability exposure under contractual fee caps versus uncapped fraud or wilful misconduct claims, grounding its calculations in evidentiary citations. 3. Verifiable Forensic Reporting: Generates an interactive HTML forensic dossier combining evidence citations, ground-truth checklist coverage, signature analysis, and financial exposure breakdowns. In our Northbridge v. Caldera demonstration case, the system identifies a suspicious signature, connects findings to supporting evidence, and quantifies a potential $4.94 million increase in financial exposure. Our goal is to make complex forensic investigations more systematic, transparent, and evidence-driven.

Team