Mar 10, 2026 | AI and product innovation
Rebuilding for the Agent Era: The Next Generation of CoCounsel Legal
Joel Hron, Chief Technology Officer
When we first brought generative AI into Thomson听Reuters听products, the goal was practical. Help professionals complete real work faster, with systems they could trust.听
One million users now听have听access听to听CoCounsel. That scale made something clear. The opportunity is larger than a collection of AI-powered skills.听听
We are at an architectural inflection point. Generative AI is not another feature layer in professional software. It听enables automation by simulating human judgment in the execution of a workflow. When that happens, incremental iteration is not enough.听Core systems must be rebuilt.听
Every major professional software platform will face this decision.听听Over the past year, we made听the decision to rebuild听CoCounsel听Legal for the听agent听era.听
This is not a feature update. It is a听foundational听transformation.听
From Delegation to听Planning听
The first generation of听CoCounsel听Legal was built around specialized skills and guided workflows. That model worked听well听for clearly defined use cases. But professional work does not stay within predefined lanes.听
As new needs听emerged, we added more skills. Routing grew more complex. Multi-step work requires听chaining workflows together.听The system was more听capable,听but the burden on users grew 鈥� to learn what skills were available, how to听call them, and听how to听combine them听into a workflow.听
The next version of听CoCounsel听removes that burden while simultaneously handling a much听wider听range of legal tasks.听The system understands听objectives听as the user describes them,听and it听comes up with听the plan for users to adjust if they wish.听
Building a Legal AI Platform听听
The next generation of听CoCounsel听Legal is being built as a听highly capable, customizable听legal AI platform.听
Instead of routing questions to isolated skills, the听system designs听a solution specific to each request. It plans. It selects tools. It retrieves authoritative content. It adapts as听new information听emerges. In the agent era, AI will not simply sit beside professional workflows. It will听guide听them.听This is not a听feature听enhancement. It is a platform re-architecture.听
The first wave of generative AI in SaaS听attached models听to existing workflows. The next wave requires redesigning the听system听around reasoning, orchestration, and verification. That is the shift we made.听
Our听objective听is to build an extensible operating system for legal work, one that allows new tools, workflows, and models to evolve on top of it. We are听leveraging听modern agent frameworks听and investing heavily into domain specific听differentiators听鈥� grounding responses in authoritative content, easing the task of fiduciary-grade verification, and听leveraging听the full breadth of听domain听expertise听from practical law.听听
The New System
The new system is content native by design. It directly accesses Westlaw, Practical Law,听KeyCite, and customer documents,听where听permissioned,听through structured content tools built for agentic reasoning.听
We are applying the same taxonomy, citation, and metadata discipline to customer content听that听is brought into the workflow for a specific customer, with听appropriate permissions, as听we do for authoritative content. That enables the agent to reason not only over authoritative external sources, but also over听that customer鈥檚听own precedents and institutional knowledge听within a clearly defined and governed boundary.听
Language models do not store up-to-date,听accurate, or verifiable knowledge of the law. Trust in high-stakes domains requires grounding in authoritative content and customer context.听
This is what makes AI professional-grade, and with this next evolution, it moves even further toward fiduciary-grade AI,听designed to meet the highest standards of responsibility and client duty.听
We听remain听model agnostic. The model layer will evolve. Our responsibility is to ensure that whichever models we use are paired with strong content intelligence, evaluation discipline, and enterprise guardrails.听
Measuring the Shift听
Architectural ambition must be matched with disciplined evaluation.
We constructed an internal evaluation set听representing听hundreds of real-world legal tasks across research, drafting, and analysis.听These reflect the work professionals actually attempt to complete.听
Thousands of evaluation runs have been reviewed by subject matter experts,听representing听tens of thousands of hours of expert review and annotation. In high-stakes domains, evaluation听isn鈥檛听a quality gate for听release;听it鈥檚听how you build the product.听
Where听we previously measured the success rate of individual skills,听we now measure听whether听the system听can听complete听a听natural听task听end to end, even when there is no prebuilt听skill designed for it. This evaluation is intentionally unforgiving: unconstrained, multi-step requests that mirror real work, reviewed by subject matter experts. Under听this听stricter听methodology,听CoCounsel听completes听roughly three听quarters听of听these听workflows,听76% in the latest evaluation runs,听with听virtually no听learning curve for users.听This is a step change in capability听that will feel听substantially different听to users on day one.听
What This Signals听听
We believe professional-grade AI platforms require four elements:听
- Advanced reasoning models and听an agentic harness to maximize their capabilities听
- Domain-specific content and intelligence听
- Deep customer context and workflow integration听
- Human听expertise, evaluation, and governance听
At听scale, this is not theoretical. It is already supporting more than one million users and evaluated across hundreds of natural, end-to-end workflows under structured subject matter expert review.听
Models alone are not enough. Content alone is not enough. Without all four working together, systems fail under听real professional听pressure.听
The organizations that define this era will not be those that add AI features fastest. They will be those that build听well-governed, extensible platforms capable of听harnessing听model improvements while听ensuring听seamless听human oversight and built-in verification听
What Comes Next听
We are expanding structured content tools, deepening听the integration between customer context and authoritative content听for complex workflows, building matter-based workspaces, enabling firm-level customization, and extending听third-party听integrations so the agent can听operate听across the systems听our customers use听every day.听
Broader availability听of听CoCounsel听Legal in its next generation听is planned for later this year.听听
A Different Standard听
AI will not replace professional judgment. But platforms architected for real work, grounded in authoritative content, and evaluated with rigor can materially change how听and where听that judgment is applied.听
For a company with more than 170 years of history, transformation is not about chasing trends.听听Rearchitecting听CoCounsel听for the agent era sets the foundation not just for our next release, but for how professional work will be executed in the听years听ahead.听
We believe this architectural shift will define the next generation of enterprise software. The question is not whether agents will听participate听in complex work, but which platforms will be trusted to run听them.
Sign Up Now听
The new听CoCounsel听Legal is entering beta in the听United States听soon, with general availability planned for later this year. More regions and territories will follow.听
听for early beta access to the fully reimagined听CoCounsel听Legal experience.听
Fewer Steps. Bigger Outcomes. A fundamentally better way to practice law.
Read more: CoCounsel Legal, Reimagined. Fewer Steps. Bigger Outcomes
