Thomson Reuters today announced the launch of the \u201cnext gen\u201d of its CoCounsel artificial intelligence platform, marking what the company describes as a fundamental shift from AI assistants that respond to prompts to intelligent agentic systems that can plan, reason and execute complex multi-step workflows within professional environments. Today\u2019s announcement centered on the immediate availability [\u2026]<\/p>\n
Thomson Reuters today announced the launch of the \u201cnext gen\u201d of its CoCounsel artificial intelligence platform, marking what the company describes as a fundamental shift from AI assistants that respond to prompts to intelligent agentic systems that can plan, reason and execute complex multi-step workflows within professional environments.<\/p>\n
Today\u2019s announcement centered on the immediate availability of CoCounsel for tax, audit and accounting professionals, while previewing similar agentic capabilities for legal professionals coming sometime this summer.<\/p>\n
The legal industry rollout will include agentic workflows for a variety of tasks such as document drafting, employment policy generation, deposition analysis, and compliance risk assessments integrated across the Westlaw, Practical Law and CoCounsel platforms.<\/p>\n
During a media preview last week at Thomson Reuters\u2019 New York City headquarters, David Wong<\/a>, chief product officer at Thomson Reuters, said the company\u2019s agentic approach represents more than an incremental improvement over existing AI tools.<\/p>\n \u201cAgentic AI isn\u2019t a marketing buzzword. It\u2019s a new blueprint for how complex work gets done,\u201d Wong said. \u201cWe\u2019re delivering systems that don\u2019t just assist but operate inside the workflows professionals use every day.\u201d<\/p>\n The distinction, he said, lies in the systems\u2019 ability to break down complex tasks into individual steps, adapt based on user response and context, guide users proactively, and extend capabilities by using tools and other software in combination with AI reasoning.<\/p>\n Unlike traditional AI assistants that require specific prompts for each task, agentic systems can understand broader objectives and determine the necessary steps to achieve them.<\/p>\n \u201cThey do this with understanding of professional work and the objectives of professional work so they can seamlessly integrate into workflows that lawyers, accountants, tax professionals, and auditors expect and provide information, answers and solutions in real time,\u201d Wong said.<\/p>\n For legal professionals, Thomson Reuters is integrating CoCounsel\u2019s agentic capabilities directly into its existing workflow and knowledge tools rather than creating standalone applications.<\/p>\n At last week\u2019s briefing, Emily Colbert<\/a>, senior vice president of CoCounsel, demonstrated how the system will guide users through complex legal processes using what the company describes as \u201cguided workflows.\u201d<\/p>\n \u201cWith our agentic guided workflows, we go from just one single-shot task, answering one question, to actually getting to a work output,\u201d she said. \u201cAnd when we think about how we apply this new technology to CoCounsel, our ambition as to what we\u2019re designing is to be the most comprehensive AI solution for legal professionals.\u201d<\/p>\n (The video above was provided by Thomson Reuters.)<\/em><\/p>\n As an example, Colbert showed how lawyers will be able to initiate document creation processes \u2014 such as drafting demand letters or employment policies \u2014 through structured workflows that leverage Thomson Reuters\u2019 legal content and expertise. The system will analyze uploaded documents, summarize key details, create chronologies of events, and produce initial drafts, all while maintaining transparency about the sources it is using and its reasoning process.<\/p>\n The legal agentic system will integrate across Thomson Reuters\u2019 legal product suite, incorporating not just just Westlaw and Practical Law, but also tools such as Westlaw\u2019s Claims Explorer functionality, which can suggest potential legal claims based on provided facts and documents.<\/p>\n Colbert demonstrated one workflow showing the system drafting a paid sick leave policy for New York, with the AI automatically researching relevant Practical Law resources and footnoting sources throughout the document. More complex workflows, such as complaint drafting, integrate multiple research steps and can tap into Westlaw\u2019s generative AI capabilities for comprehensive legal research.<\/p>\n She said that CoCounsel will significantly reduce the time it takes to perform legal work, estimating that it will cut the time to review documents or to draft and review contracts by as much as 63%, while cutting the time of legal know-how tasks by 10%.<\/p>\n When the legal product is released sometime this summer, it will come with a number of these guided workflows, Colbert said, although she was not ready to specify what they will be. Although subject to change, among those listed in last week\u2019s preview were:<\/p>\n Colbert said a number of these workflows will be released this summer, with more to be added on an ongoing basis, covering all aspects of litigation, transactional and advisory legal work.<\/p>\n TR says its development of this agentic platform has been ongoing for over a year and was accelerated by its acquisition last year of Materia<\/a>, an AI startup specializing in agentic systems for tax and accounting.<\/p>\n Speaking at last week\u2019s briefing,\u00a0Ryan Walker<\/a>, vice president of engineering at Thomson Reuters, and formerly the chief technology officer at Casetext, which originally developed CoCounsel before its acquisition by Thomson Reuters in 2023<\/a>, said that the system is designed to be multimodal and multi-agent, capable of orchestrating across different models and agents rather than relying on a single tool for all tasks.<\/p>\n \u201cNo one tool is going to be the best tool for the job to be done,\u201d Walker explained. He also emphasized that the platform maintains human-in-the-loop functionality at two levels: users collaborate with the system to produce work product, and human experts are integral to building and training the tools.<\/p>\n The system integrates across Thomson Reuters\u2019 content, including more than 20 billion documents and 15 petabytes of data, while maintaining what the company describes as enterprise-grade security with ISO 42001 certification and zero-retention architecture for client data.<\/p>\n In developing this new version of CoCounsel, Thomson Reuters says, it has partnered with multiple AI providers, including OpenAI, whose models power some elements of CoCounsel.<\/p>\n While today\u2019s announcement only teased the coming launch of agentic AI for legal, it included the live release of agentic AI for tax and accounting professionals, offering a preview of what legal professionals can expect. Kevin Merlini<\/a>, vice president of product for CoCounsel Tax, Accounting and Audit and the former CEO of the aforementioned Materia, demonstrated how the system automates complex workflows while maintaining professional standards.<\/p>\n The tax system includes \u201cReady to Review,\u201d an agentic tax preparation application that can extract information from client documents, apply reasoning about where data should be placed in tax forms, input information into Thomson Reuters\u2019 GoSystem Tax engine, and respond to diagnostic messages from the tax software. When the system encounters issues it cannot resolve, it brings humans into the loop with context about the problem and potential solutions.<\/p>\n \u201cThis isn\u2019t GenAI in a prettier wrapper,\u201d Merlini said. \u201cIt\u2019s a fully integrated, intelligent system built to do the work. Now CoCounsel doesn\u2019t just assist \u2014 it acts with context, navigates complexity, and integrates directly into how professionals already operate. It\u2019s purpose-built for high-stakes work \u2014 and it\u2019s only the beginning.\u201d<\/p>\nLegal Implementation<\/h3>\n
\n
Multimodal and Multi-Agent<\/h3>\n
Live for Tax and Accounting<\/h3>\n