Anthropic adds a gallery of ready-made legal artifacts to Claude for Legal (cite check, contract tables, regulatory scans), with tools from Harvey, PwC, Goodwin and Foley

Your AI and this story

Key facts

• Anthropic’s six artifacts: cite check (flags quotes that don’t match, wrong pages, unsupported sources); jurisdiction survey matrix (e.g. breach-notice rules across states, with links to the law); matter board; playbook deviation board (routes non-standard clauses to the approver); regulatory horizon scan (timeline ranked by impact); tabular review (terms such as severance or non-competes across a stack of contracts, linked to the exact language)
• Partner artifacts: AI Governance Benchmark (Goodwin), Harvey Intake (routes business users’ legal questions, answers routine ones from Legal’s guidance in Harvey), Obligations Checker (Foley & Lardner), Subpoena Response Tool (PwC); more coming from Legora and Orrick
• Also announced: funding for public-interest groups building access-to-justice artifacts and for legal-education programmes; a Legal Ambassadors Program; more than 60 legal connectors added since May (Bloomberg Law, Filevine, Ivo, Legora, Patlytics and others); sharing settings that support ethical walls
• Customer quote: Spellbook says it is ‘approaching 500,000 contracts per month’ on Claude; Eve says Claude Opus 5.5 ‘made a real leap’ on its PlaintiffBench
• Reach: @claudeai post (Oct 28, 15:34 UTC) ~4.44M views by Oct 29 (fxtwitter)
• Same day, Claude Enterprise: smart reports (Claude analyses a sample of a team’s transcripts and groups the work into workstreams with spend, deliverables, friction points and recommendations such as skills to create; ‘aren’t for evaluating individual performance or making employment decisions’), in-app surveys and pooled group budgets, all generally available (TechCrunch: ‘tools to show companies what their AI spending is buying’)
• Smart reports and privacy (Anthropic help centre): the beta Enterprise analytics report’s Inefficiencies section counts sessions that were personal or off-topic, with their cost, as totals only (no summaries or drill-down), and removes them before analysis; Mashable’s Oct 29 headline framed it as reports that ‘tell bosses when workers are off-topic’