Legora CTO on building legal AI agents: $100M ARR, evals without customer data
Latent Space · rss · 2026-11-03
Latent Space interviews Legora CTO Jacob Lauritzen after its $550M Series E at $8.8B valuation. Key points:
- Business: crossed $100M ARR in April within 18 months, 1,800+ customers, five acquisitions this year (Wexler latest).
- Billable hour upended: routine reviews shift to flat fees; junior lawyers are worth more than two years ago.
- Evals without traces: 3 engineers + 3 lawyers build RL environments (e.g., a 10,000-file data room due diligence run), 100 legal engineers on call, failures classified into 10-20 known modes, everything on Azure because law firms wouldn't accept AWS/GCP.
- Models: 60% of legal work already handled well; post-training open-weight models for narrow tasks; every output cited with bounding-box highlights to fight hallucinations.
- Primitives: projects, agent, Tabular Review (up to 1M cells), and Notion-style lists power all products; legal engineers configure without code, FDEs write production integrations.
More from coding & agent
- Five manual checks before trusting an agent actually finished its job — Kodenstudio · 2026-11-03
- Open-source AI pit crew for iRacing runs 5 agents on Kimi, checks every number against telemetry — JeremyCMorgan · 2026-11-03
- Agent control plane benchmark: Smalltalk-80 cascades hit 100% compliance vs 2.5% for Gherkin — RandalSchwartz · 2026-11-03
- Karpathy: one prompt to Claude Code deploys a YouTube thumbnail-swap browser extension — amaarora · 2026-11-03
- Cherry Pi proposes layered harness architecture: agent loops compose with governed workspaces — sull · 2026-11-03
- $200 Claude Max runs overnight multi-subagent Opus 5.5 job, burns only 6% of weekly quota — lxfater · 2026-11-03