RSGI x LITI: Recent Insights

In early January, Tom Saunders unpacked Litera’s headline claim of a record-breaking year for AI adoption, and the detail behind it is telling. Litera reports 10x growth in access to its generative AI-enabled cloud drafting tools since Q3 2025, as more customers transition onto its cloud environment.
Out of roughly 15,000 customers, around 7,000 currently have access to these capabilities. The piece situates Litera’s numbers against newer entrants like Harvey and Legora, highlighting the advantage incumbents can have when distribution, workflow integration, and installed customer bases are already in place.
Read the full piece here: Litera’s AI adoption in numbers: Customers, usage, strategy and moat.

Later in January, Saunders covered Luminance’s most significant product update to date: “institutional memory” that retains negotiation history and decision making across a company’s contracts.
The announcement is a useful marker of how quickly the category is evolving beyond point-in-time clause analysis toward systems that can learn context over time. Luminance attributes the shift to advances in model reasoning and context length, paired with what it calls “recursive legal context engineering”, a way for the system to pull in the right historical or organisational information when a query demands it, rather than treating every document as an isolated artifact. The article notes the different practical constraints vendors face as they work through privacy, governance, and rollout readiness.
Read the article here: Luminance unveils major product update with institutional memory

The focus then shifts from vendors to the people who have to operationalise all of this. In early February, LITI & RSGI reported from CLOC, where a senior in-house crowd leaned into a practical question that keeps resurfacing: how legal teams communicate value to the business, especially as they build data-driven models to support legal operations and AI adoption.
One of the strongest threads from the day was that “value” rarely travels on metrics alone. Storytelling matters, but it needs credible data behind it; champions outside legal ops can sometimes carry the message further than the team itself; and even the familiar “no” from legal can land differently when framed as “no, but here’s what we can do instead.”
See LITI’s coverage of CLOC here: CLOC Europe: practical takeaways on legal ops, AI value, and cultural transformation

In another piece, Reena SenGupta pulled together reactions to Anthropic’s legal plug-in, an announcement that jolted the legal tech market and sparked loud predictions about whether domain specific tools are at risk. Some voices, particularly those closer to enterprise buying dynamics, see increased competition as inevitable and potentially good for consumers, especially for in-house teams already standardised on a given model provider. Others argue that the real moat for legal platforms is not just model access, but the reliability layers that sit around it: workflow fit, security posture, accuracy expectations, governance, and deep domain training that reflects how legal work actually gets done.
View the article here: The Anthropic legal plug-in palaver: A round-up of views

Adoption at scale was also underscored by Ashwin Gohil’s coverage of Thomson Reuters’ announcement that its AI platform CoCounsel is now accessible to one million professionals worldwide, just two years after launch. The article unpacks what sits behind that headline number, from CoCounsel’s reach across legal, tax, accounting and compliance roles to the company’s heavy ongoing investment in AI as competition intensifies.
Read the full article here: CoCounsel now available to one million professionals, says Thomson Reuters

Finally, the most data heavy piece in this set brings the conversation back to foundations. LITI reported on iManage’s February benchmarking study, which argues that knowledge maturity is a key differentiator in AI success. While a large majority of organisations report that they are piloting, implementing, or using AI, only a much smaller fraction have embedded it into daily operations. In iManage’s view, the gap isn’t primarily about intent; it’s about data quality, governance, and the ability to translate experimentation into sustained, everyday value.
See the article here: Knowledge maturity identified as key differentiator in AI success