Anthropic on Wednesday gave lawyers a new way to work inside Claude, with ready-made dashboards for legal teams’ daily jobs. The new legal gallery, announced on October 28, 2026, adds interactive artifacts for citation checking, contract review and regulatory tracking to Claude for Legal, the product Anthropic launched in May. Each tool opens with a worked example, links every output to its source, and leaves the final call to the lawyer.
From chat window to workbench
The pitch is a change of furniture, not just a change of features. Much of legal work is badly shaped for a chatbot: reviewing a stack of contracts is an inventory exercise, checking citations is an audit, and tracking regulation is a timeline. A scrolling conversation buries all three. Artifacts, Claude’s interactive mini-apps such as trackers and dashboards, sit beside the chat instead, so a team can edit, share and reuse the same structured view.
“A lot of legal work doesn’t fit very neatly in a back-and-forth chat. There’s a need for a different kind of interface — and artifacts provide that,” said Robert Mahari, head of Claude for Legal. The tools inherit the organisation’s existing permissions and guardrails, Reuters reports, and sharing settings support ethical walls between teams.
Six tools, one division of labour
Anthropic built six artifacts itself, covering legal teams’ most common requests since May. Cite check compares a document with its sources and flags quotations that do not match, wrong page numbers and claims the cited source does not support. The jurisdiction survey matrix compares one legal topic, such as data breach notice rules, across states or countries in a single table linked to the underlying law. The matter board turns engagement letters, emails and spreadsheets into one board of open matters, showing the owner, the next step and the next deadline.
The other three aim at the contract pile: a playbook deviation board flagging clauses that differ from standard terms, a regulatory horizon scan mapping upcoming rules on a timeline ranked by impact, and tabular review pulling chosen terms from many contracts into one table linked to the exact language.
Anthropic’s framing is deliberately modest: the lawyer brings the documents, Claude completes the first pass, and the lawyer applies judgment. Mark Pike, Anthropic’s associate general counsel, said his own team already uses artifacts internally on contracts, privacy impact assessments and patent issues. “It’s really shifted things away from chatbot-like experiences, where you ask Claude questions and get answers, to things that are more interactive and visual,” Pike said. The announcement publishes no accuracy figures for the tools, so firms must check the first pass like a junior associate’s work.
Built with firms, not just for them
The gallery is not only Anthropic’s work. Four artifacts come from firms and legal tech companies: Goodwin’s AI Governance Benchmark, Harvey Intake for business users’ legal questions, Foley & Lardner’s Obligations Checker, and PwC’s Subpoena Response Tool for audit-ready first drafts. More artifacts from Legora and Orrick are on the way.
Underneath the gallery sits a connector library that has grown well past its May launch: more than 60 legal connectors, including Bloomberg Law, Filevine, Ivo, Legora and Patlytics, so teams can pull in playbooks and research sources directly. That plumbing matters, because a dashboard is only as good as the documents feeding it.
The access-to-justice corner
Anthropic also funded work at the other end of the market. The gallery includes nine access-to-justice artifacts, five of them built by nonprofit and public interest organisations supported with Claude credits. They include a Criminal Charge Reality Check for New York defence teams, an Immigration Papers Reader, and rent ledger and record-sealing tools built on D.C. law. A new Legal Ambassadors Program lets professionals share what they build.
Why It Matters
Legal work is one of the most contested professional markets in AI right now: Google expanded its Gemini Enterprise platform with tools for lawyers and law firms in August, Thomson Reuters launched its own Thomson 1.0 model the same month, and Microsoft has rolled out legal-specific capabilities for its Copilot platform. Anthropic is preparing for a potentially record-setting public listing, and a credible professional product line is part of the case it makes to investors and enterprise buyers.
But the deeper test is trust, and it cuts both ways. Citation errors have already embarrassed courts, which is why a cite check leads the gallery. At the same time, a tool that makes the first pass faster also makes it easier to skip the second look, and Anthropic’s own language — first pass, lawyer’s judgment, links to sources — reads like a warning label as much as a sales pitch. The firms that benefit most will be the ones that treat these dashboards as junior staff: useful, fast, and never unsupervised.
Background
Claude for Legal launched in May 2026, and Anthropic has since been expanding it into a professional workbench, from drug discovery work to arguments over how risky AI models should be governed. The legal gallery is the latest step, and its reception among law firms will show whether AI assistants graduate from answering questions to doing structured work.
Source: Reuters report and Anthropic announcement
0 Comments