EvenUp builds AI for plaintiff-side personal injury practice. Its best-known output is the demand package — a drafted demand letter supported by medical evidence and comparable verdicts — but the platform now spans the case lifecycle: Express and Expert Demands, AI Drafts for litigation documents, MedChrons interactive medical chronologies, Communication Agents for client and carrier contact, a settlement repository, and firm analytics.
Founded in San Francisco in 2019 by Raymond Mieszaniec, Saam Mashhad, and Rami Karabibar, EvenUp reports more than 2,000 firms on the platform, including 20% of the top 100 U.S. personal injury firms, handling roughly 10,000 cases weekly. In October 2025 it raised a $150M Series E led by Bessemer Venture Partners — with participation from REV, the venture arm of RELX (owner of LexisNexis) — bringing total funding to $385M at a valuation above $2 billion.
EvenUp's differentiator is data rather than a general-purpose model: its proprietary Piai model is trained on hundreds of thousands of injury cases and millions of medical records, and demand packages cite comparable verdicts from a dataset the company reports at 250,000+ so attorneys can argue settlement value from precedent.
On data handling, EvenUp's trust center states that it does not allow third-party AI providers, including OpenAI and Anthropic, to use customer case files to train their models, and that it uses zero-retention settings where available. It also says its models are designed not to store or reproduce customer documents. However, the company separately describes improving its own product using experience across the cases it processes, so this listing marks the first-party training question as unclear rather than a clean no. EvenUp reports SOC 2 Type 2 examination and HIPAA attestation.
The platform is personal-injury-only — it is not a general practice tool — and is aimed at firms with steady case volume; third-party reviewers note that ROI is tighter for practices with small active caseloads.