Notes from Legal Tech Talk 2026

This week I headed to London for Legal Tech Talk, Europe’s largest legal tech conference. This was my first LTT since launching the Best Practice Newsletter back in September 2025, so I’m dedicating this week's edition to the trends and tech that caught my eye.

The Case for Staying Grounded

Both Thomson Reuters and LexisNexis went all-in on grounding their tech in hard facts—a message that couldn't have come at a better time. Both tools plug into their databases and link to the underlying source.

This is particularly relevant having recently seen that Damien Charlotin’s database of court filings tainted by fabricated AI citations has now crossed 1,300 cases worldwide, in addition to the recent press that has faced multiple other firms over hallucinated AI cases.

All-in on Agents

On the development front, much of the chatter between everyone has shifted from "chatting" with the AI tool to "working with the tool”.

Legora showcased its Agentic Operating System, while Harvey put its own agents front and centre. Many of the speakers on the panels confirmed this too, how we are moving past simply asking a model a question, and are now handing a matter over to a system and letting it run—research, drafting, and review all handled seamlessly between specialised agents.The term “agentic” is also the new buzzword for 2026, although being truly agentic can mean different things for different vendors.

The market is backing this shift with plenty of of capital into the space. Just months ago in March, Harvey secured a $200 million growth round co-led by Sequoia and GIC, catapulting its valuation to $11 billion specifically to scale these multi-step agentic workflows.

Highlights from the Startup Stage

The startup stage drew a solid crowd, proving there is still plenty of room for more Legal AI startups. Other than a few Harvey and Legora wannabe dupes, one company doing something about different was Courtroom AI, co-founded by ex Kirkland associate Elizabeth Grabowski Parikh.

They have built a proprietary data set trained on real people, built to to simulate jury panels and judges. They claim it will help companies predict potential trial outcomes (a massive sector in the US, where the vast majority of civil cases are decided by jury trials).

It's exciting to see so many novel new startups targeting a specific problem, many of them being built by ex-laywers who have previously experienced it in their practice area.

AI and the Unloved Timesheet

The time-billing corner was also noticeably busy.

Both PointOne and Laurel were both there pushing their AI-driven timekeeping products. Automating time capture might not be glamorous, but it is exactly the kind of unglamorous, high-frequency friction point where AI can help out. PointOne in particular has momentum behind it with a $16m Series A in April this year, led by 8VC, takes its total to roughly $20m - not bad for a small team.

A few final thoughts

What struck me most walking the floors at LTT was not only the amazing atmospheric buzz, but also the sheer volume of capital driving Legal AI forward. Yes, you can see it in the massive exhibition stands, but the real indicator is the talent.

Much of this capital is flowing into a steady, distinct stream of lawyers walking away from their traditional roles and TC’s to build their careers directly inside legal AI companies, helping them to build the best possible tools on the market.

I think there is plenty of opportunity left in Legal AI. The real runway has moved on from chatbots and now lies in niche vertical areas, including back office opportunities such as business development, operations, and billing.

Not everything has to be a client-facing tool to be revolutionary. Simply automating the high-frequency friction points of running a firm is where the value lies.

I'm incredibly excited to see what the next year brings for Legal Al. I hope you'll continue to follow along and keep Best Practice as your home for legal AI insights - l have a fantastic lineup of exciting guests coming up on the podcast that you won't want to miss.

See you next week,

-George