- Robert Mahari is Anthropic’s first Head of Claude for Legal and will lead deployment across the legal industry. [1] - Anthropic launched 20-plus legal connectors and 12 practice-area plugins in May, with integrations spanning Thomson Reuters, Relativity, LexisNexis, document-management systems and e-discovery platforms. [2] - Enterprise customers can configure permissions, audit logs, retention controls and no-training defaults, but Anthropic acknowledges Claude can still produce incorrect or fabricated legal content.
[3][4] Anthropic has hired Robert Mahari to lead Claude’s expansion into law firms and legal departments, creating what the company describes as its first Head of Claude for Legal role. Mahari holds a PhD in legal AI and founded Akiva AI, a startup focused on helping legal teams adopt artificial-intelligence systems. [1]
The Decoder reported the appointment on August 12, 2026. Anthropic has not publicly specified Mahari’s start date or provided a fuller description of the role’s reporting line and decision-making authority. His remit, according to the report, is to deploy and expand Claude across the legal industry. [1]
A growing legal product line #
The appointment follows Anthropic’s May launch of more than 20 Model Context Protocol connectors and 12 legal plugins. The tools connect Claude with systems including Thomson Reuters’ CoCounsel, Relativity, iManage, NetDocuments, DocuSign and Everlaw, while plugins target work such as contract review, corporate transactions, privacy compliance, regulatory monitoring and litigation preparation. [2]
Anthropic’s public safeguards include permission-bound access to connected repositories, audit trails, configurable retention and a default commitment not to train models on commercial customer inputs and outputs. The company also says legal teams should verify outputs against authoritative sources and keep professional judgment in the loop. [3][4]
Law firms are moving past pilots #
The move comes as legal AI adoption accelerates. Bloomberg Law reported that all 40 large firms in its 2025 survey that disclosed technology usage had adopted legal-specific AI tools. Freshfields said in April that it was deploying Claude across 33 offices and giving access to 5,700 employees, while Hanson Bridgett announced a firmwide Claude rollout in June. [5]
Anthropic’s own documentation cautions that Claude can produce authoritative-sounding but false citations and should not be treated as a single source of truth. It does not publicly assign responsibility for legal errors beyond emphasizing customer configuration, verification and attorney oversight, leaving liability questions to contracts, professional rules and individual deployments. [4]
Companies mentioned #
Further sources #
[1] The Decoder reported that Robert Mahari is Anthropic’s first Head of Claude for… ↗
[[2] Anthropic’s May 12, 2026 product announcement described more than 20 legal MCP … ↗](https://claude.com/blog/claude-for-the-legal-industry)
[[3] Anthropic’s legal product materials describe permission-bound connectors, audit… ↗](https://claude.com/solutions/legal?success=true)
[4] Anthropic’s help materials acknowledge that Claude can generate incorrect or mi… ↗
[5] Bloomberg Law reported that all 40 large firms in its survey that disclosed tec… ↗
The stories that matter, in one email. Free — unsubscribe anytime.