Thomson Reuters launches CoCounsel for tax & audit

Thomson Reuters has launched a new version of CoCounsel Tax, Audit & Accounting for professionals in those fields. The product is now available to customers in the United States.

The platform is designed to take users from source documents and client data to completed work within a single workflow. It is aimed at tax, audit and accounting teams that need calculations, analysis and draft deliverables tied to relevant authority and open to professional review.

The launch comes as professional services firms test how far generative artificial intelligence can move beyond chat-based answers into regulated work. In tax and audit, the central issue is not only whether a system can generate an output, but whether the reasoning behind that output can be traced to the right rule, period and client facts.

According to Thomson Reuters, the latest version of CoCounsel can plan work, perform calculations, apply authority and produce finished material for human review. The company positions the system as a way to reduce manual effort while keeping professional judgment in the review process.

How it works

Thomson Reuters highlighted functions such as code execution and verified calculations for tasks like depreciation, amortisation, required minimum distributions, qualified business income, threshold checks, and multi-scenario comparisons. The platform also brings together client information, engagement activity and workflow context in one place, the company said.

It can produce multi-tab Excel workbooks with live formulae and an assumptions tab, as well as draft memos and other documents in Word and PDF formats. It can also update spreadsheets and documents professionals upload while following a firm's existing workpaper templates, Thomson Reuters said.

The system supports MCP connectivity, which allows it to operate within other applications and systems and incorporate information from them into its working context, the company said. Thomson Reuters added that it has expanded integrations with products including UltraTax, Engagement Manager and Guided Assurance.

Content base

The product sits on Thomson Reuters' broader CoCounsel platform and draws on the group's tax and accounting tools, including Ready to Review and Ready to Advise. According to the company, it is built on a multi-model approach that uses different models for different tasks.

Its underlying research base includes Checkpoint, Thomson Reuters' accounting research engine. The repository contains more than 12 million primary source and proprietary analytical documents maintained by hundreds of specialists, including certified public accountants and tax attorneys, the company said.

That content base is central to the company's pitch. In professional compliance work, vendors are increasingly trying to distinguish between general-purpose AI systems that can draft text or write code and specialist systems that can link conclusions back to recognised source material.

Richard Lynch, Managing Principal at Sikich, described the impact on his teams in a customer endorsement. "The capabilities emerging from tools like Thomson Reuters CoCounsel Tax, Audit and Accounting represent some of the most significant productivity advancements in my teams' careers," Lynch said.

"By streamlining standard tasks, CoCounsel has helped Sikich tax and audit professionals become more agile by improving efficiency and enhancing consistency across deliverables," he added.

Thomson Reuters executives said the product was built to address a common weakness in AI systems used for professional work: the gap between generating an answer and demonstrating why that answer is correct.

David Wong, Chief Product Officer at Thomson Reuters, said the company sees the product as changing the standard for AI in the sector.

"The next generation of CoCounsel Tax, Audit & Accounting changes what professionals should expect from AI. It doesn't just answer a question; it plans the work, runs the calculation, applies the right authority, and delivers a finished result a professional can review and sign their name to," Wong said.

"Building that took decades of authoritative content, credentialed expertise and real-world integrations, the kind of foundation that turns AI that answers questions into AI that does the work," he added.

Joel Hron, Chief Technology Officer at Thomson Reuters, said the technical challenge lies less in doing calculations than in selecting the right facts and legal or accounting basis before any calculation begins. "We didn't just build a tool that does math. We built one that knows why the math is right," Hron said.

"The code execution piece is table stakes now; every company in this space has it. What's hard is making sure the system is reasoning from the right facts and the right authority before it ever runs a calculation. We built that judgment in, grounded in content that's kept current and tested by the people who actually do this work, so a professional can check it themselves rather than take our word for it," he added.