If I could give you only one takeaway from Legalweek in New York last week, it would be that more mature and thoughtful conversations have emerged from the froth and chaos of the post GenAI years. Discussion topics included trust, workflow, cost, ROI, and shadow IT. Obviously agentic AI was prevalent, and there are some exciting advances on this front. But so too were increasingly urgent thoughts around what technological change means for legal professional’s roles, training, and how legal work is delivered.
Trust and workflow evolution
Speaking during a press briefing, Sean Fitzpatrick, CEO of legal research giant LexisNexis, said: “We’re seeing an inflection point. We’re moving from testing to use and efficiency, and we’re seeing a move to trust.”
Like any technology, user experience dictates adoption, and LexisNexis has rolled out a new ChatGPT-style interface for Lexis+ Protege that “people love.” AI assistant Protege includes over 10,000 pre-built workflows and if your process doesn’t map to one of those, Lexis’ white glove service can build it for you.
The markets were recently spooked (to put it mildly) by the release of Anthropic’s Claude legal plugin, which is now integrated within Protege. Chief technology officer Greg Dickason said that Lexis was working with Anthropic long before the launch and that Cowork is great at many things, including automating complex tasks. “But it’s not grounded and it can’t verify sources,” Dickason said. The combination inside of Protege means an agentic framework grounded in Lexis’ content, giving a “final polished outcome,” he said.
Training your people
Workflow was also a heavy focal point for Actionstep, which in January launched automated time capture tool Trace and, CEO Early Stephens says, is working on the agentification of AI for delivery. The workflow can “manage any task”: the agent reads a request and gives you details on what happens next.
There has been slow adoption of cloud technology in the law firm practice management space, and Stephens says, the business continues to grow as firms take that leap.
What is the hardest truth about ROI that firms aren’t ready to hear, I ask. Saunders says: “I don’t know if firms are thinking enough about how they train their new hires in an AI world.”
Reimagining how work is delivered
Thinking about training neatly segues into questions, and maybe less answers, around how work product will be delivered going forward. For my part, this is a source of fascination. Technology is leaping ahead and it will change who delivers what tasks within a matter, but few firms are yet re-imagining their processes.
To give an example or two, at Legalweek, Thomson Reuters gave me a demo of the next version of CoCounsel Legal, which includes more autonomous agents.
What this means, is that a partner or senior associate can potentially conduct a piece of work quickly without handing it off to a junior. Whether they should or not, is another question.
In another example, Litera announced an integration with Midpage, an AI-powered legal research platform, to bring U.S. case law and statutes directly into Lito, Litera’s AI legal agent. It means that US statutes and case law are available within Lito, and within Micosoft 365, where lawyers are working. How will this impact the current law firm division of labour, I wonder.
There are examples of where AI may enable more junior lawyers to ‘play up’, for argument’s sakes. NetDocuments Smart Answers, launched in March, enables legal professionals to ask natural language questions and receive answers grounded in their firm’s documents, with citations.
“It’s not just a legal issue: you need to take a step back and say, ‘how is the business going to reengineer itself?’” QuisLex recently launched an advisory service to help legal teams integrate AI but also optimise operations and Gummaregula says: “If you focus too narrowly on the problem you’re trying to solve, you might miss the opportunity.”
The ecosystem is changing
For in-house teams, one of the big questions is how much work you should be giving to outside counsel and how much you should keep in-house.
At Epiq, president Roger Pilc says that corporate legal teams in addition to using outside counsel and/or using an ALSP such as Epiq, can build their own technology. At Legalweek they hosted a session on building your own agent, with Pilc observing: “The ecosystem is changing. We see that the percentage of work done by ALSPs and in-house teams will grow.”
I have reiterated many times already this year that many in-house teams are planning to keep more work internally. There are limitations, though. Are already stretched in-house teams set up to and capable of bringing more work in-house? “Maybe a small minority,” says Pilc, “but most don’t have the support.” For this reason, Epiq launched Agent Labs, to help teams turn their ideas into agentic AI solutions.
Shadow IT
One thing all organisations need to be careful of in this AI race is shadow IT, which is a growing problem. At Control Risks, partner Brad Kolacinski said: “There is such an arms race and people are keeping up with the Joneses, but there is a big risk of shadow IT and also of building a Frankenstein tech stack.
“Where are you accessing data from and are you falling foul of data privacy laws? If you’re in a verein, are you falling foul of what you can share? It’s really important to set expectations,” he said.
While companies may have a policy, staff often find a way to instal other systems and applications and, Kolacinski says, it’s worth bearing in mind that in an eDiscovery process, this is where any offending text is likely to lie.
AI for good
While focusing on the risks of AI going wrong, it is only fair to mention the conversations I had around using AI for good. Two in particular stand out.
The first is the news from Everlaw that its Everlaw for Good Program has, over the past year, supported more than 675 active cases across 235 organisations, and expanded its support to a growing network of non-profit organisations.
The program extends Everlaw’s technology to organisations working to advance access to justice. In a recent survey by Everlaw, 88% of legal aid professionals said they are optimistic about AI’s potential to help narrow the justice gap.
“Mission-driven organizations are increasingly handling complex investigations and litigation with limited resources,” said Joanne Sprague, head of Everlaw for Good. “Expanding access to powerful, easy-to-use technology helps level the playing field so these teams can uncover critical evidence, take on more complex matters, and yield stronger results for the communities they serve.”
A final takeaway
I had conversations at Legalweek around whether AI will mean the end of lawyers. Some of the smartest people I know are quietly convinced that this is the direction of travel for a large chunk of legal work. These conversations, for another time, get quite heated, in my experience.
We are nowhere near there yet, but legal teams need to be very strategic in their thinking now. Technology is advancing. Trust is growing. Agentic workflows are taking – or will take – a bunch of tasks off lawyers’ hands. What does the next stage of practice look like? The time for reengineering your processes – at least making a start – is now.
The post Legalweek: It’s time to re-engineer how legal work is delivered appeared first on Legal IT Insider.