Survey-based read on everyday GenAI usage. Most users stay in the shallow end (emails, summaries), power users stretch further, and firms with templates and training see materially higher adoption depth.
TR's internal benchmark on how frontier LLMs actually cope with long legal documents. Useful reality check if you're buying based on context window size alone.
Adoption up across legal, tax and compliance (~70% using GenAI) but trust still a drag (~65% flagging reliability). Fewer than 10% feel ready for client AI demands.
New reasoning model generation. o3 is the flagship for complex multi-step reasoning, o4-mini is a lower-cost version aimed at agentic workloads.
Official guidance on getting the best out of GPT-4.1. Covers the quirks of the new model's instruction-following, agentic workflows and tool use.
Tight piece on the specific reasons Word keeps the legal market - all-or-nothing adoption, no commit atomicity, and weak version / approvals tracking. Bleedingly obvious once pointed out.
Skills.law's NPS-style survey of ~100 firms. Top 15 includes DeepJudge, Centari, Syntheia, Lega, Legora, vLex, DraftWise, Definely and StructureFlow.
Mollick on the new reasoning wave. Capabilities are superhuman in places and fragile in others - "jagged AGI" - but agentic behaviour now changes the integration picture.
The Lawyer's seven-profile framework for UK firm innovation (Accelerators, Developers, Experimenters, Full Service, Going Steady, Pivoters, Specialists). AG appears as an Experimenter.
Commentary arguing firms conflate activity with achievement. Budget, announcements and pilots don't equal business-model change.
Detailed profile of how White & Case runs innovation across practice areas. Useful benchmark for firms building internal innovation functions.
Wharton's empirical finding that "best" prompts vary wildly by task, model and context. Strong evidence against one-size-fits-all prompt libraries.
LN's Chief AI Officer on a multi-model strategy: 30+ models tested, fine-tuned SLMs, knowledge graphs and agents, quality over cost. Clear view of where a major legal publisher is betting.
Annual WTI based on 31,000 respondents plus M365 telemetry. "Frontier firms" with AI deeply embedded are 60% more likely to be seeing productivity gains.
UK feature on the shift toward voice-first legal AI and "digital colleague" framing. Covers Kingsley Napley's Knowledge Exchange and other firm-vendor partnerships.
FT on how the Big Three consultancies are rewiring their delivery model around AI. Direct read-across for how large professional services businesses (including law) restructure billable work.
Argues the core lawyer skill is shifting from writing clauses to codifying judgment as structured prompts. Makes the case that this actually raises the bar on "why" reasoning.
Kingsley Napley with Let's Think launch The Knowledge Exchange, a conversational tool that captures senior lawyer expertise and surfaces it to juniors. Interesting approach to the institutional knowledge problem.
OpenAI Agent Guidance (PDF)
[Internal AG resource] OpenAI's practitioner guidance on building agents - planning, tool use, evaluation, safety. Good reading for anyone designing agentic workflows.
OpenAI's free learning platform: workshops, digital content and events covering AI literacy through to developer integration. Worth pointing staff towards as a foundational resource.