Anthropic and law firm Freshfields sign a deal to develop specialized legal AI tools for document drafting, contract review, due diligence, and more
Tech group will tap ‘magic circle’ law firm's expertise as it builds products that can be sold to rivals
Context & Ripple Effects
Legal teams and large firms have been testing AI on work traditionally performed by junior lawyers, while Harvey, Legora and Anthropic are putting pressure on established legal-information platforms. The competition is no longer only about general-purpose models; it is moving into workflow-specific products.
Freshfields’ participation gives Anthropic access to practitioner input in a market where another leading firm, Kirkland & Ellis, is pursuing its own platform rather than adopting tools shared with rivals. That contrast highlights a strategic choice for firms: help shape a commercial product or retain proprietary AI capabilities.
First-order effects
- Anthropic can use Freshfields’ legal expertise to tailor tools for drafting, review and due-diligence workflows, while Freshfields gains influence over product design and access to the resulting capabilities.
- The arrangement creates the prospect of a product built with one elite firm’s expertise being offered to competing firms, potentially widening Anthropic’s legal-market reach.
Second-order effects
- Legal-tech incumbents such as Thomson Reuters and LexisNexis face added pressure to improve workflow-specific AI features, not merely provide legal research or underlying content.
- Law firms must weigh the benefits of vendor tools shaped by peer-firm expertise against the control and differentiation offered by building proprietary platforms, as Kirkland is attempting.
Third-order effects
- If such collaborations proliferate, specialist professional services may become a key channel through which frontier AI companies turn general models into vertical software products.
- The legal sector could split between firms that help define shared vendor platforms and firms that treat AI workflow systems as proprietary operational infrastructure; the durability of either model will depend on whether shared tools deliver enough advantage without eroding differentiation.
The trend: This is part of the institutionalization of frontier AI: model providers are partnering with domain leaders to package general-purpose capabilities into professional workflows and compete with incumbent vertical software.