top of page

Enterprise-Grade AI, Finally Done Right—Thomson Reuters' Agentic System Raises the Industry Bar

In an era where artificial intelligence (AI) is no longer a buzzword but an operational imperative, Thomson Reuters has taken a pivotal leap forward by unveiling its new agentic AI platform — a game-changing innovation poised to revolutionize complex professional workflows across legal, tax, accounting, and compliance sectors. This move signals a significant paradigm shift, as AI evolves from being a reactive assistant to becoming an autonomous agent capable of planning, reasoning, acting, and adapting within real-world processes. This article explores the emergence of agentic AI, the unique value proposition of Thomson Reuters' CoCounsel system, and how this evolution is setting a new benchmark for the future of professional-grade AI solutions. What Is Agentic AI — And Why Now? Agentic AI refers to artificial intelligence systems that go beyond passive, prompt-based responses. Unlike traditional generative AI models that simply generate content upon request, agentic AI systems exhibit autonomous behavior: they interpret objectives, break them into logical steps, act within established workflows, and even escalate decisions to human supervisors when required. In short, agentic AI is goal-oriented, context-aware, and accountable. This progression is timely and necessary. Professionals across sectors are grappling with ever-increasing complexity in regulatory, legal, and financial environments. According to a McKinsey & Company study, knowledge workers spend 60% of their time on routine administrative tasks, leaving little room for strategic analysis or value-adding judgment. The transition to agentic systems allows organizations to automate not just data entry or document drafting, but entire cognitive workflows — from audit reviews to legal compliance assessments. Introducing CoCounsel: Agentic AI for Tax, Audit, and Accounting At the forefront of this revolution is CoCounsel, Thomson Reuters’ first agentic AI application, purpose-built for tax, audit, and accounting professionals. Developed in collaboration with OpenAI and enhanced by proprietary datasets, CoCounsel represents a new breed of AI that integrates deeply into professional tools and executes complex tasks with precision. Key Highlights: Autonomous Workflow Execution: CoCounsel doesn’t wait for instructions. It understands the objective, navigates firm-specific documents and tax regulations, and executes a sequence of tasks—while maintaining full transparency and audit trails. Explainable Output: Every decision made by CoCounsel is accompanied by explainable reasoning, satisfying compliance and ethical mandates. Integrated Knowledge Systems: It connects internal firm data, IRS codes, regulatory databases, and content from platforms like Checkpoint and Practical Law to deliver context-rich results. Real-World Impact: Speed, Accuracy, and Trust Agentic AI platforms like CoCounsel have already demonstrated measurable outcomes for early adopters. For example, BLISS 1041, an accounting firm managing multistate fiduciary tax filings, leveraged CoCounsel to reduce residency and filing code reviews from several days to under an hour. By embedding customized templates and workflows within CoCounsel, the firm achieved not only speed but also unparalleled consistency and accuracy. Benefits Reported by Early Users: 85% reduction in task turnaround time 40% increase in compliance accuracy 70% reduction in client review delays Full traceability of AI decision-making, addressing regulatory scrutiny These gains are particularly vital in sectors where regulatory failure has reputational and financial consequences. Agentic AI doesn’t replace human judgment — it augments it by automating the routine and escalating edge cases that require ethical or legal discretion. Inside the Architecture: What Makes Thomson Reuters’ Agentic Platform Unique? Unlike AI vendors offering modular or plug-in-based tools, Thomson Reuters has re-architected its core product suite to enable seamless agentic orchestration across its platforms. The key pillars of its agentic AI infrastructure include: Component Description Trusted Content Base 20 billion+ documents and 500+ curated legal and tax content assets Custom LLMs Proprietary models trained on domain-specific data and professional standards Human-in-the-Loop (HITL) Escalation workflows that involve humans for final decisions, maintaining accountability Cross-Platform Orchestration Integration across Checkpoint, Westlaw, GoSystem Tax Engine, and Practical Law Zero Retention Policy ISO 42001-certified architecture with no customer data retention By unifying these components, Thomson Reuters has not just built an agent — it has created a professional AI operating system tailored for high-stakes use cases. Why Agentic AI Is the Future of Professional Services While generative AI laid the foundation, agentic AI represents the natural next stage in the AI maturity curve. The transformation is not merely technological—it’s architectural, cultural, and strategic. Here’s why agentic AI is poised to become indispensable: Goal-Based Intelligence: It doesn't wait for specific prompts; it starts with a business goal and works backward to deliver actionable results. Dynamic Tool Use: Agentic systems can select, combine, and orchestrate internal and third-party tools to execute multi-step plans. Adaptability: They can adjust their behavior in response to real-time input or evolving constraints. Compliance Assurance: Built-in explainability and audit logs make them compliant-ready for regulated industries. Scalability: Agentic systems are inherently scalable across jurisdictions, tax codes, languages, and firm-specific data formats. As David Wong, Chief Product Officer at Thomson Reuters, explains: "We’re delivering systems that don’t just assist but operate inside the workflows professionals use every day. The AI understands the goal, breaks it into steps, takes action, and knows when to escalate for human input — all with human oversight built in to ensure accountability and trust.” Next in Line: Agentic AI Beyond Tax Following CoCounsel’s launch, Thomson Reuters is rapidly expanding the agentic model into new verticals, including: Ready to Review: A tax return preparation agent built on the GoSystem Tax Engine. It autonomously drafts returns, adapts to reviewer feedback, and resolves diagnostics. Legal AI Agents: Enhancements to Westlaw and Practical Law that include deposition analysis, intelligent drafting, and risk policy generation. Compliance & Risk: Agentic workflows that automate policy generation, trade compliance assessments, and due diligence reporting. These advancements are expected to touch all 500,000+ customers of Thomson Reuters — including 100% of the Fortune 100 and the entire US federal court system — making it a standard bearer in enterprise-grade AI. Challenges and Considerations: Governance, Bias, and Human Oversight Despite its promise, agentic AI systems must navigate several key risks: Bias and Fairness: As agentic AI systems make more decisions, ensuring their fairness and eliminating latent biases becomes essential. Regulatory Compliance: Operating in heavily regulated industries means agentic AI must not only be accurate but legally interpretable. Trust and Transparency: Systems must not act as black boxes; they must provide traceable logic and allow for human override. Ethical Boundaries: The AI must understand the limits of automation and escalate morally sensitive decisions. To mitigate these risks, Thomson Reuters has embedded human-in-the-loop governance at every stage, ensuring AI supports — not replaces — professional decision-making. A New Era of Autonomous Professional Intelligence Thomson Reuters has not merely introduced a new product; it has redefined the very paradigm of AI in professional domains. CoCounsel, and the broader agentic platform behind it, signify a watershed moment where artificial intelligence transitions from an assistant to a trusted autonomous agent — executing complex tasks with precision, transparency, and professional alignment. As industries race toward digital transformation, the firms that adopt agentic AI will be better equipped to handle complexity, scale compliance, and deliver value faster. For deeper insights into the future of enterprise-grade artificial intelligence, automation, and decision systems, follow the expert analyses from Dr. Shahid Masood and the 1950.ai research team. Their ongoing work on predictive AI, intelligent agents, and digital sovereignty continues to shape policy, innovation, and ethical frameworks globally. Further Reading / External References Thomson Reuters Debuts Agentic AI Platform Press Release: Thomson Reuters Ushers in the Next Era of AI

In an era where artificial intelligence (AI) is no longer a buzzword but an operational imperative, Thomson Reuters has taken a pivotal leap forward by unveiling its new agentic AI platform — a game-changing innovation poised to revolutionize complex professional workflows across legal, tax, accounting, and compliance sectors. This move signals a significant paradigm shift, as AI evolves from being a reactive assistant to becoming an autonomous agent capable of planning, reasoning, acting, and adapting within real-world processes.


This article explores the emergence of agentic AI, the unique value proposition of Thomson Reuters' CoCounsel system, and how this evolution is setting a new benchmark for the future of professional-grade AI solutions.


What Is Agentic AI — And Why Now?

Agentic AI refers to artificial intelligence systems that go beyond passive, prompt-based responses. Unlike traditional generative AI models that simply generate content upon request, agentic AI systems exhibit autonomous behavior: they interpret objectives, break them into logical steps, act within established workflows, and even escalate decisions to human supervisors when required.

In short, agentic AI is goal-oriented, context-aware, and accountable.


This progression is timely and necessary. Professionals across sectors are grappling with ever-increasing complexity in regulatory, legal, and financial environments. According to a McKinsey & Company study, knowledge workers spend 60% of their time on routine administrative tasks, leaving little room for strategic analysis or value-adding judgment. The transition to agentic systems allows organizations to automate not just data entry or document drafting, but entire cognitive workflows — from audit reviews to legal compliance assessments.


Introducing CoCounsel: Agentic AI for Tax, Audit, and Accounting

At the forefront of this revolution is CoCounsel, Thomson Reuters’ first agentic AI application, purpose-built for tax, audit, and accounting professionals. Developed in collaboration with OpenAI and enhanced by proprietary datasets, CoCounsel represents a new breed of AI that integrates deeply into professional tools and executes complex tasks with precision.


Key Highlights:

  • Autonomous Workflow Execution: CoCounsel doesn’t wait for instructions. It understands the objective, navigates firm-specific documents and tax regulations, and executes a sequence of tasks—while maintaining full transparency and audit trails.

  • Explainable Output: Every decision made by CoCounsel is accompanied by explainable reasoning, satisfying compliance and ethical mandates.

  • Integrated Knowledge Systems: It connects internal firm data, IRS codes, regulatory databases, and content from platforms like Checkpoint and Practical Law to deliver context-rich results.


Real-World Impact: Speed, Accuracy, and Trust

Agentic AI platforms like CoCounsel have already demonstrated measurable outcomes for early adopters. For example, BLISS 1041, an accounting firm managing multistate fiduciary tax filings, leveraged CoCounsel to reduce residency and filing code reviews from several days to under an hour. By embedding customized templates and workflows within CoCounsel, the firm achieved not only speed but also unparalleled consistency and accuracy.


Benefits Reported by Early Users:

  • 85% reduction in task turnaround time

  • 40% increase in compliance accuracy

  • 70% reduction in client review delays

  • Full traceability of AI decision-making, addressing regulatory scrutiny

These gains are particularly vital in sectors where regulatory failure has reputational and financial consequences. Agentic AI doesn’t replace human judgment — it augments it by automating the routine and escalating edge cases that require ethical or legal discretion.


Inside the Architecture: What Makes Thomson Reuters’ Agentic Platform Unique?

Unlike AI vendors offering modular or plug-in-based tools, Thomson Reuters has re-architected its core product suite to enable seamless agentic orchestration across its platforms.

The key pillars of its agentic AI infrastructure include:

Component

Description

Trusted Content Base

20 billion+ documents and 500+ curated legal and tax content assets

Custom LLMs

Proprietary models trained on domain-specific data and professional standards

Human-in-the-Loop (HITL)

Escalation workflows that involve humans for final decisions, maintaining accountability

Cross-Platform Orchestration

Integration across Checkpoint, Westlaw, GoSystem Tax Engine, and Practical Law

Zero Retention Policy

ISO 42001-certified architecture with no customer data retention

By unifying these components, Thomson Reuters has not just built an agent — it has created a professional AI operating system tailored for high-stakes use cases.


Why Agentic AI Is the Future of Professional Services

While generative AI laid the foundation, agentic AI represents the natural next stage in the AI maturity curve. The transformation is not merely technological—it’s architectural, cultural, and strategic.


Here’s why agentic AI is poised to become indispensable:

  1. Goal-Based Intelligence: It doesn't wait for specific prompts; it starts with a business goal and works backward to deliver actionable results.

  2. Dynamic Tool Use: Agentic systems can select, combine, and orchestrate internal and third-party tools to execute multi-step plans.

  3. Adaptability: They can adjust their behavior in response to real-time input or evolving constraints.

  4. Compliance Assurance: Built-in explainability and audit logs make them compliant-ready for regulated industries.

  5. Scalability: Agentic systems are inherently scalable across jurisdictions, tax codes, languages, and firm-specific data formats.


As David Wong, Chief Product Officer at Thomson Reuters, explains:

"We’re delivering systems that don’t just assist but operate inside the workflows professionals use every day. The AI understands the goal, breaks it into steps, takes action, and knows when to escalate for human input — all with human oversight built in to ensure accountability and trust.”

Next in Line: Agentic AI Beyond Tax

Following CoCounsel’s launch, Thomson Reuters is rapidly expanding the agentic model into new verticals, including:

  • Ready to Review: A tax return preparation agent built on the GoSystem Tax Engine. It autonomously drafts returns, adapts to reviewer feedback, and resolves diagnostics.

  • Legal AI Agents: Enhancements to Westlaw and Practical Law that include deposition analysis, intelligent drafting, and risk policy generation.

  • Compliance & Risk: Agentic workflows that automate policy generation, trade compliance assessments, and due diligence reporting.


These advancements are expected to touch all 500,000+ customers of Thomson Reuters — including 100% of the Fortune 100 and the entire US federal court system — making it a standard bearer in enterprise-grade AI.


Challenges and Considerations: Governance, Bias, and Human Oversight

Despite its promise, agentic AI systems must navigate several key risks:

  • Bias and Fairness: As agentic AI systems make more decisions, ensuring their fairness and eliminating latent biases becomes essential.

  • Regulatory Compliance: Operating in heavily regulated industries means agentic AI must not only be accurate but legally interpretable.

  • Trust and Transparency: Systems must not act as black boxes; they must provide traceable logic and allow for human override.

  • Ethical Boundaries: The AI must understand the limits of automation and escalate morally sensitive decisions.


To mitigate these risks, Thomson Reuters has embedded human-in-the-loop governance at every stage, ensuring AI supports — not replaces — professional decision-making.


A New Era of Autonomous Professional Intelligence

Thomson Reuters has not merely introduced a new product; it has redefined the very paradigm of AI in professional domains. CoCounsel, and the broader agentic platform behind it, signify a watershed moment where artificial intelligence transitions from an assistant to a trusted autonomous agent — executing complex tasks with precision, transparency, and professional alignment.


As industries race toward digital transformation, the firms that adopt agentic AI will be better equipped to handle complexity, scale compliance, and deliver value faster.


For deeper insights into the future of enterprise-grade artificial intelligence, automation, and decision systems, follow the expert analyses from Dr. Shahid Masood and the 1950.ai research team. Their ongoing work on predictive AI, intelligent agents, and digital sovereignty continues to shape policy, innovation, and ethical frameworks globally.


Further Reading / External References

Comments

bottom of page