Problem and mission

The Innocence Center - a pro bono law firm freeing the wrongfully convicted - faces a critical challenge: potential clients who maintain their innocence may never reach the Center due to:

  1. self-selection from confusion regarding eligibility
  2. simply not knowing the Innocence Center exists

Meanwhile, strong innocence claims sit buried in lengthy parole hearing transcripts that take hours to review.

Our mission is to find the innocent and accelerate attorney review so the Innocence Center can reach more people in need.

What we built

Justice AI is a proactive client identification and case analysis tool built specifically for parole hearing transcripts. It:

  • detects explicit and implicit innocence claims
  • generates 1-page briefs, can export as PDF
  • extracts critical case factors
  • line-level citations for all insights

Every factual statement links directly back to the source transcript so attorneys can fact-check every claim.

Attorneys have everything they need to:

  1. identify eligible clients
  2. approve cases
  3. begin building legal cases for exoneration

How it works in one flow

  1. Upload a parole hearing transcript.
  2. Justice AI parses and indexes with page and line references. It detects and flags innocence claims - both explicit denials and contextual patterns. It extracts parole factors, contradictions, programming evidence, and supporting quotes.
  3. The summary and structured fields are saved to the case view for attorney review.
  4. Attorneys may export the one-page case summary as a PDF to import into their case management system.

Why this matters

Proactive reach: The Innocence Center can now identify potential clients with strong evidence for innocence but may not know they qualify for representation.

Verifiable insights: Line-level citations mean attorneys and volunteers can trust the analysis and verify every fact in seconds.

Faster triage: Attorneys spend minutes reviewing cases instead of hours, enabling the lean team to scale impact.

What is next

  • Integration with Lawmatics: Prefill client intake data for Lawmatics, the Innocence Center's CRM, to accelerate onboarding and early-stage case building.

  • Public source monitoring: Automatically detect newly posted CDCR hearing transcripts to eliminate manual uploads

  • Multi-transcript consolidation: Merge multiple hearings per client into unified case files with deduplicated citations

Built With

Share this project:

Updates