AI × Law

Major legal AI providers are moving away from simple prompt-driven chatbots to autonomous, agentic platforms that handle complex workflows. LexisNexis released Protégé capabilities in Lexis+, Thomson Reuters upgraded CoCounsel Legal, and Harvey launched Harvey II featuring personalized user memory.

Thomson Reuters has introduced its own custom AI model integrated into its flagship CoCounsel platform. The launch marks a direct competitive push against emerging legal AI specialized platforms like Harvey and Legora.

A survey of in-house legal leaders by Axiom revealed that among the 18% of legal departments yet to adopt AI, clarity on regulation is the top catalyst needed to begin exploring the technology. Lack of clear regulatory rules ranked far ahead of ROI concerns (19%) or budget constraints (11%) as the primary blocker.

AI compliance startup Norm Ai has raised $260 million and launched its own law firm, Norm Law, which employs 18 former Big Law partners to deliver legal services alongside automated AI agents. Institutional clients and investors like Blackstone, Bain Capital, and Khosla Ventures are using the firm for fund formations and transaction work at reported savings of 20% to 70%.

A federal judge in United States v. Heppner held that prompts and responses from a consumer AI chatbot were not protected by attorney-client privilege or the work-product doctrine. The court emphasized that chatbots are not attorneys and third-party consumer terms eliminate any reasonable expectation of confidentiality.

Thomson Reuters has partnered with Solve Intelligence to bring specialized, AI-driven patent workflows to its platform. The integration enables IP litigation and prosecution teams to draft patent applications, respond to office actions, and construct claim charts within Microsoft Word.

After buying back contract-analysis startup eBrevia from strategic investor Donnelley Financial Solutions, the original co-founders have layered generative AI onto the platform. The firm launched DraftPro, a new tool that uses firm precedents and playbooks to draft and evaluate agreements against preferred negotiation positions.

LexisNexis announced dynamic agentic AI capabilities for its Lexis+ With Protégé platform, moving beyond fixed workflows to integrate directly with client data repositories and internal Lexis resources. The legal publisher also revealed plans to potentially develop its own proprietary AI model to support these advanced capabilities.

National law firm Nelson Mullins is rolling out AI platform Harvey across all practice groups for its more than 1,300 attorneys and legal professionals. The firm reported that internal adoption of the tool dramatically outpaced any previous technology deployment in its history.

LexisNexis has expanded its Lexis+ with Protégé AI solution to Hong Kong, introducing guided workflow capabilities for complex legal tasks. The system integrates LexisNexis legal content, the local CaseBase citation service, and models from Anthropic to support automated drafting, review, and intake tasks.

Microsoft's Corporate, External, and Legal Affairs group has selected Harvey for its legal and compliance operations following a competitive selection process. Despite having Copilot in-house, Microsoft's 2,000-person legal group opted for the specialist legal AI tool over its own internal alternatives.

Thomson Reuters invested $40 million to build its first proprietary, domain-specific large language model based on open-source foundations and its extensive legal database. The new model initially powers Tabular Analysis in CoCounsel Legal and may eventually be licensed directly to law firms and corporate legal departments.

Major asset managers including Blackstone, Bain Capital, and Khosla Ventures are sending legal work to Norm Ai and its affiliated law firm, Norm Law, seeking lower costs and faster delivery. Valued at $1.2 billion after raising $260 million, the AI-native startup employs former Big Law attorneys to build AI agents for transaction and compliance workflows, challenging traditional law firm fee structures.

Legal tech startup Newcode has raised a Series A funding round led by OnDean Forward, bringing its total 2026 capital raised to $20 million. The company, which provides a configurable AI harness to help law firms build custom agentic workflows, will use the funds to support its U.S. expansion across the Am Law 200. Investors in the round also include The LegalTech Fund, Antiportfolio Ventures, and Relativity's Rel Labs.

In ANI v OpenAI, the Delhi High Court refused to issue an interim injunction against OpenAI, finding prima facie that storing news articles to train ChatGPT constitutes fair dealing under Indian copyright law. The court also ruled that while it has territorial jurisdiction over the US-based AI company, the output generated by ChatGPT did not amount to a substantial reproduction of ANI's news content.

The Colorado Supreme Court adopted Rule Change 2026(02), amending attorney ethics rules to clarify that AI use does not diminish a lawyer's duty of competence or independent judgment. The update follows high-profile federal court sanctions in the state over fabricated AI citations and misrepresented precedent.

In a 10 million lira promissory note lawsuit in Istanbul, the defendant alleged that the plaintiff's lawyer cited a non-existent Constitutional Court decision generated by AI. The plaintiff's lawyer denied using AI in the petition, and the court ultimately issued a jurisdictional ruling transferring the case to labor court.

Legal AI startup Harvey has unveiled Tenet, its first proprietary domain-specific model, post-trained on Moonshot's open-weight Kimi K3 model in partnership with Fireworks AI. The move shifts Harvey away from sole reliance on external foundation model providers like OpenAI, Anthropic, and Google to control costs and tailor performance for legal tasks.

Vals AI secured $40 million in Series A funding led by Andreessen Horowitz at a $400 million valuation. The startup provides independent benchmarking and evaluation infrastructure for AI models across domain-specific sectors, including the legal industry.

Enterprise AI startup Twin1 AI raised $20 million in a seed round co-led by Aramco Ventures, Bessemer Venture Partners, and Tribeca Venture Partners, with participation from law firm Orrick. The platform builds permission-aware 'digital twins' of professionals using work context from documents, emails, and workplace tools, serving initial customers across the legal, financial, and energy sectors.