Thomson Reuters just unveiled the next step in the evolution of CoCounsel, and the future is agentic! At least that’s the pitch. Of course it sort of has to be the pitch given the agent conversation dominates the industry zeitgeist and no one can risk being left behind. But while McConaughey cuts commercials about turning over dinner reservations to SkyNet, the question remains just how much do lawyers need or even want an agent?
Based on TR’s announcement, it feels as though they know the answer is “not much,” but can’t duck the branding anyway.
“While today’s most advanced AI assistants can generate results when prompted, agentic AI goes beyond simply responding under a pre-defined sequence of actions,” declares this morning’s press release. What does it mean to go beyond? “It plans, reasons, acts, and even reacts — operating inside real workflows to complete complex, multi-step assignments with the transparency, precision, and accountability professionals require.”
In an industry that routinely marks down work conducted by elite law school graduates, it’s hard to imagine there’s a real hankering for a bot that “plans, reasons, acts, and even reacts.” If we’re not trusting a third-year from Harvard Law to make decisions on their own, it’s not clear why we’d want an algorithm.
The good news is that the agents that Thomson Reuters showed off appeared a lot less autonomous than the sales copy might suggest.
The product announcement for law dovetails with an announcement aimed at tax, audit, and accounting professionals. While admittedly not an expert in those worlds, the product looks useful for those folks — automating compliance reviews, memo drafting, and regulatory checks. It taps into Checkpoint, IRS code, and internal documents to deliver the right output based on the situation.
The good news for lawyers, is that the legal product seemed less an autonomous tool deciding whether or not to open the pod bay doors and more of an AI-enabled workflow. Which, I suppose, is some level of autonomy and has a definite use case, but there’s a lot more human oversight. And that matters in legal because, not to spoil a 60-year-old movie but HAL 9000 didn’t try to kill everyone because he broke with his programming — he tried to kill everybody because he was following his instructions with the power to “plan, reason, act, and even react.” Trying to do the job correctly was the problem.
The rollout of agentic systems continues this year with expanded capabilities across legal, risk and trade, and compliance domains — including intelligent workflows for intelligent drafting, employment policy generation, deposition analysis, and compliance risk assessments. Many of these experiences already exist within CoCounsel, Westlaw, and Practical Law, but are now being upgraded with full agentic orchestration, where agents not only generate output but plan, execute, and adapt across tools in real time.
That comports with what I saw. In a nutshell, the user can feed the product a file and it will start pursuing a vetted workflow. For example, it can scour the material and start brainstorming claims and pulling up relevant material from Practical Law and drafting a complaint. That’s useful! Would I call it an agent? Maybe… but that’s not exactly an Ari Gold level of agent.
Nor do I think lawyers would want more. There’s a reason why Thomson Reuters paid big money for CoCounsel and not AutomaticAttorney. A co-counsel is there to help, an agent is there to take care of the problem for you. They want tools that can fill in templates, check citations, pull filings, summarize discovery. To the extent lawyers have embraced AI at all, they’re looking for an ally they can skeptically consult and not something trying to think on its own.
The lawyers who’ve tried the latter haven’t fared well.
Which is why it’s such a conundrum to cover this announcement because what Thomson Reuters is calling an agent is not what most of the marketing hype would consider an agent… and that’s GOOD! This is a next-level automated workflow constructed on the back of the incredibly deep Thomson Reuters knowledge base and crafted by subject matter experts. The offerings — which they could only say were coming at some point this year — are “governed by human-in-the-loop oversight for safety, accuracy, and accountability.” That’s as it should be, but there’s obvious tension between “plans, reasons, acts, and even reacts” and “we guarantee there’s human judgment all over this!”
The latter sounds less cool to a tech conference, but a lot more heartwarming to a malpractice carrier.
Thomson Reuters obviously has a deep customer base and gather a lot of feedback, but I’m just not buying that lawyers are going to get hyped about “agentic.” Silicon Valley might be racing to build models that book flights, order lunch, and eventually plot the downfall of mankind, but what gets the legal industry in a buying mood is going to be something much more boring. The “human-in-the-loop” part matters a lot more.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news. Joe also serves as a Managing Director at RPN Executive Search.
The post ‘Agentic’ AI Is The Hot Buzzword… But Do Lawyers Actually Want An Agent? appeared first on Above the Law.
Thomson Reuters just unveiled the next step in the evolution of CoCounsel, and the future is agentic! At least that’s the pitch. Of course it sort of has to be the pitch given the agent conversation dominates the industry zeitgeist and no one can risk being left behind. But while McConaughey cuts commercials about turning over dinner reservations to SkyNet, the question remains just how much do lawyers need or even want an agent?
Based on TR’s announcement, it feels as though they know the answer is “not much,” but can’t duck the branding anyway.
“While today’s most advanced AI assistants can generate results when prompted, agentic AI goes beyond simply responding under a pre-defined sequence of actions,” declares this morning’s press release. What does it mean to go beyond? “It plans, reasons, acts, and even reacts — operating inside real workflows to complete complex, multi-step assignments with the transparency, precision, and accountability professionals require.”
In an industry that routinely marks down work conducted by elite law school graduates, it’s hard to imagine there’s a real hankering for a bot that “plans, reasons, acts, and even reacts.” If we’re not trusting a third-year from Harvard Law to make decisions on their own, it’s not clear why we’d want an algorithm.
The good news is that the agents that Thomson Reuters showed off appeared a lot less autonomous than the sales copy might suggest.
The product announcement for law dovetails with an announcement aimed at tax, audit, and accounting professionals. While admittedly not an expert in those worlds, the product looks useful for those folks — automating compliance reviews, memo drafting, and regulatory checks. It taps into Checkpoint, IRS code, and internal documents to deliver the right output based on the situation.
The good news for lawyers, is that the legal product seemed less an autonomous tool deciding whether or not to open the pod bay doors and more of an AI-enabled workflow. Which, I suppose, is some level of autonomy and has a definite use case, but there’s a lot more human oversight. And that matters in legal because, not to spoil a 60-year-old movie but HAL 9000 didn’t try to kill everyone because he broke with his programming — he tried to kill everybody because he was following his instructions with the power to “plan, reason, act, and even react.” Trying to do the job correctly was the problem.
The rollout of agentic systems continues this year with expanded capabilities across legal, risk and trade, and compliance domains — including intelligent workflows for intelligent drafting, employment policy generation, deposition analysis, and compliance risk assessments. Many of these experiences already exist within CoCounsel, Westlaw, and Practical Law, but are now being upgraded with full agentic orchestration, where agents not only generate output but plan, execute, and adapt across tools in real time.
That comports with what I saw. In a nutshell, the user can feed the product a file and it will start pursuing a vetted workflow. For example, it can scour the material and start brainstorming claims and pulling up relevant material from Practical Law and drafting a complaint. That’s useful! Would I call it an agent? Maybe… but that’s not exactly an Ari Gold level of agent.
Nor do I think lawyers would want more. There’s a reason why Thomson Reuters paid big money for CoCounsel and not AutomaticAttorney. A co-counsel is there to help, an agent is there to take care of the problem for you. They want tools that can fill in templates, check citations, pull filings, summarize discovery. To the extent lawyers have embraced AI at all, they’re looking for an ally they can skeptically consult and not something trying to think on its own.
The lawyers who’ve tried the latter haven’t fared well.
Which is why it’s such a conundrum to cover this announcement because what Thomson Reuters is calling an agent is not what most of the marketing hype would consider an agent… and that’s GOOD! This is a next-level automated workflow constructed on the back of the incredibly deep Thomson Reuters knowledge base and crafted by subject matter experts. The offerings — which they could only say were coming at some point this year — are “governed by human-in-the-loop oversight for safety, accuracy, and accountability.” That’s as it should be, but there’s obvious tension between “plans, reasons, acts, and even reacts” and “we guarantee there’s human judgment all over this!”
The latter sounds less cool to a tech conference, but a lot more heartwarming to a malpractice carrier.
Thomson Reuters obviously has a deep customer base and gather a lot of feedback, but I’m just not buying that lawyers are going to get hyped about “agentic.” Silicon Valley might be racing to build models that book flights, order lunch, and eventually plot the downfall of mankind, but what gets the legal industry in a buying mood is going to be something much more boring. The “human-in-the-loop” part matters a lot more.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news. Joe also serves as a Managing Director at RPN Executive Search.
The post ‘Agentic’ AI Is The Hot Buzzword… But Do Lawyers Actually Want An Agent? appeared first on Above the Law.
