Thomson Reuters just launched agentic AI for its CoCounsel assistant that plans and executes multi-step workflows autonomously. Early results show tax jurisdiction reviews that took half a week now finish in under an hour.
Key Points:
- CoCounsel now features agentic AI, enabling autonomous, multi-step task execution in tax and accounting workflows.
- Thomson Reuters plans to extend agentic AI to legal, risk, and compliance domains in the coming months.
The new agentic version of CoCounsel launched Monday for tax, audit, and accounting professionals, with legal and compliance versions coming this summer. Unlike typical AI assistants that only respond to prompts, this agentic AI version plans multi-step tasks and executes them across Thomson Reuters' platforms — Checkpoint, Westlaw, and Practical Law.
"Before CoCounsel, we were manually comparing residency and filing codes across 36 states. Each jurisdiction used to take us half a week to fully review—now it takes under an hour," said Rich Marlatt, CIO at BLISS 1041.
David Wong, Chief Product Officer at Thomson Reuters, emphasized the transformative nature of this development: “Agentic AI isn’t a marketing buzzword. It’s a new blueprint for how complex work gets done.” He highlighted that these systems are designed to operate within existing professional workflows, understanding goals, breaking them into actionable steps, and knowing when to seek human input .
Thomson Reuters claims they rebuilt their core software around AI that can actually operate autonomously. The system connects firm knowledge, IRS code, and internal documents into one workspace where the AI handles complete workflows, not just individual queries.
Thomson Reuters has some serious advantages here: 20B+ documents, 15+ petabytes of data, and 500+ trusted content assets. Plus 4,500 subject matter experts and 180+ AI engineers working side-by-side. When you're training AI systems for high-stakes professional work, having that much domain expertise matters.
The system also tackles one of the biggest problems with AI in professional settings: trust. Their agents are fine-tuned by legal and tax, audit, and accounting experts to reason in alignment with professional standards and best practices, while ensuring that human expertise remains in the loop to guide judgment, validate outputs, and make final decisions.
The company is planning to roll out agentic capabilities across their entire ecosystem this year with expanded capabilities across legal, risk and trade, and compliance domains. This will include new workflows for intelligent drafting, employment policy generation, deposition analysis, and compliance risk assessments.
They're also launching something called "Ready to Review" — an agentic tax prep application that's redefining what professional-grade AI can do. Built on the GoSystem Tax Engine, it will draft returns, adapt to system feedback, and resolve diagnostics on its own.
The real test isn't whether the technology is impressive — it's whether busy professionals will trust it with work that actually matters. "As agentic AI emerges as a predominant theme in 2025 – marking a fundamental shift from traditional AI tools to proactive agents and teams of agents – so too will questions around accountability and control of these increasingly autonomous systems".