Signal SentrySearchUpdated Nov 20, 05:30 UTCAM Drop
Agents · industry

Agents in Legal

Agent deployments and launches in Legal, from the Drops.

From the Drops

AgentsLegalTestingHuman in loop: unclear

Preprint: delegating to worker agents found more conflicting facts but flagged fewer

A preprint reports that on Harvey LAB document-review tasks across three models, forced delegation raised the share retrieving both conflicting claims from 50.2% to 68.7%, while conflict detection fell from 16.2% to 13.2%.

Why it matters: Splitting research across sub-agents can hide contradictions at the handoff. Test whether your lead agent sees the conflicting evidence workers found, not only their summaries.

On the Frontier watch →From Drop #094 · Wed, Nov 18, 2026 · Evening
AgentsLegalLiveHuman in loop: unclear

Legora's CTO on testing legal agents without seeing customer data

On Latent Space, Legora's Jacob Lauritzen says three engineers and three lawyers build simulated tasks, like a 10,000-file data room, since client data is off limits; production failures are sorted into 10 to 20 known types.

Why it matters: If privacy blocks you from reading agent traces, Legora shows a path: rebuild the work in test environments and track results by task type, since a new model can regress on some work.

From Drop #063 · Tue, Nov 3, 2026 · Morning
AgentsLegalTestingHuman in loop: unclear

Preprint: with a strong harness, the weakest Claude model beat the strongest alone

Researchers turning US state bills into strict XML crossed three harnesses with three Claude models on 153 gold bills. The weakest model in the full multi-agent pipeline scored 0.948, versus 0.806 for the strongest model alone.

Why it matters: The tooling around a model can matter more than the model, the authors report; a deterministic lookup tool supplied citation exactness that a larger model did not.

From Drop #026 · Thu, Oct 15, 2026 · Evening