Tom Saunders, RSGI
Kirkland & Ellis announced today (28 May) that it would spend $500m over the next three to four years developing its own custom AI tools and services. The story broke as an exclusive in the Financial Times and while the firm has been tight on details so far, here’s what we know and how it compares to what else is going on in the market.
The firm will pay for this investment out of its revenue, which was $10.6bn last year and the AI platform will be broad in its use, with lawyers able to use it across their work rather than relying on multiple tools. Undisclosed outside companies have been working with Kirkland & Ellis to build the technology but will not be able to sell it to other law firms. This distinguishes Kirkland’s approach from Freshfields’ recent deal with Anthropic, which does allow the AI company to sell products developed to rival firms.
While legal AI sometimes feels like a two-horse race between Harvey and Legora at the moment, Kirkland and Freshfields are not the only firms leaning into the “build” approach. In our FT Innovative Lawyers research, we have come across several examples of firms building their own proprietary AI solutions. Two of the most comprehensive examples are from Simmons & Simmons and Allen & Gledhill. Simmons & Simmons built Percy, a generative AI platform, entirely in-house using its own LLM team, launching it in late 2023 and reaching 87% adoption among fee earners within a year. Percy runs on a purpose-built “legal inference engine” and sits entirely within Simmons’ own network. Singapore firm Allen & Gledhill took a similarly uncompromising approach with A&GEL, a custom large language model platform hosted entirely on-premise, built that way specifically to meet client confidentiality requirements in Singapore’s financial services market. Dentons also announced a partnership with OpenAI at the end of 2025 which gives the firm early access to the latest models and technical assistance from the AI company when creating new tools.
What makes the Kirkland investment stand out is not just the size, it is what the firm is betting on. By owning the technology outright, Kirkland controls its roadmap, its data, and its future decisions about how and whether to commercialise it. That is a significant strategic position to be in. This is also consistent with Kirkland’s previous moves around technology and data which have arguably helped make it the largest law firm in the world by revenue. RSGI covered the firm’s CTRAN database almost 10 years ago in the FT Innovative Lawyers reports, a proprietary system for collecting data on past M&A transactions that allowed the firm to spot trends in deal terms and deploy that intelligence for clients. That was a data advantage that competitors couldn’t easily replicate. This feels like the same instinct at a much larger scale. If the bet pays off and AI genuinely transforms how the firm handles complex mandates end to end, it could be a durable source of competitive advantage in a market where many firms are using the same tools as each other. If it doesn’t, $500m is a very expensive lesson. And for legal AI vendors, the implicit message is uncomfortable: the world’s highest-revenue law firm has just decided that best-in-class bought software is not good enough.
