Thomson reuters unveils legal AI, greenberg traurig goes all-in

Thomson Reuters has strategically developed its proprietary large language model, Thomson, purpose-built for the legal domain by training on decades of proprietary content from Westlaw, Practical Law, and Reuters news archives. This $40 million investment, leveraging fewer than 400 GPUs, delivers a cost-efficient and high-performance alternative to frontier AI models, enabling specialized agentic AI capabilities such as autonomous tabular document analysis within the CoCounsel Legal platform. By integrating Thomson directly with trusted legal content and research tools, Thomson Reuters ensures outputs are transparent, verifiable, and aligned with professional workflows, embodying their concept of fiduciary-grade AI that meets the high-stakes accuracy demands of legal practice.

CoCounsel Legal exemplifies next-generation agentic AI by seamlessly embedding autonomous task execution within established legal workflows through deep integration with trusted systems like iManage, HighQ, and Noetica. The expanded partnership with iManage, including the innovative Model Context Protocol (MCP), allows CoCounsel to query AI-enhanced document management systems directly, preserving existing permissions and governance frameworks while enhancing access to matter-specific context. This integration not only streamlines complex legal workflows but also reinforces Thomson Reuters’ premium content positioning by enabling AI-driven automation that respects law firms’ institutional knowledge and ethical guardrails.

Thomson Reuters’ agentic AI approach in CoCounsel Legal leverages cascading autonomous workflows that outsource mental drudgery by self-generating prompts and iteratively refining outputs, significantly boosting productivity in high-volume legal tasks. While this method consumes substantially more computational resources—up to thirty times the tokens of a single query—it is carefully governed with ethical guardrails to maintain lawyer oversight and responsibility, ensuring that AI-generated work products remain transparent and verifiable. This balance addresses early skepticism among legal professionals about relinquishing control, as the platform’s design fosters trust by embedding explainability and citations directly into AI outputs.

By owning and continuously refining Thomson, built on the Snowdon open-weight foundation developed with Imperial College London’s FAIR Lab, Thomson Reuters gains a defensible technological edge that reduces dependence on third-party AI providers like Anthropic’s Claude. This autonomy enables ongoing model improvements using vast proprietary datasets—only a fraction of which have been leveraged so far—positioning Thomson to evolve with unique legal knowledge that frontier AI labs cannot easily replicate. The multi-model strategy within CoCounsel Legal intelligently deploys Thomson where it excels, such as in structured document review, while integrating other leading AI models for complementary tasks, optimizing both performance and operational costs in a competitive legal AI market.

Greenberg Traurig’s four-year partnership with Thomson Reuters was pivotal in co-developing and beta testing the next-generation CoCounsel Legal AI platform, culminating in a comprehensive global rollout across all 51 offices and 3,200 lawyers. This firmwide deployment reflects a strategic move beyond limited pilots, demonstrating confidence in the platform’s ability to seamlessly integrate into diverse legal workflows and enhance everyday practice on a global scale.

The CoCounsel Legal platform uniquely combines Thomson Reuters’ authoritative legal content with Greenberg Traurig’s internal knowledge management system, DeepJudge, enabling lawyers to access both firm-specific insights and trusted external resources within a single, agentic AI-powered workflow. Chief Pricing and Innovation Officer Matthew N. Beekhuizen highlights this integration as key to maintaining the professional judgment and quality clients expect, while the platform’s psychologically inviting interface encourages meaningful lawyer involvement throughout the AI-assisted process rather than relegating humans to end-stage reviewers.

Greenberg Traurig’s global rollout thoughtfully addresses the firm’s geographically and culturally diverse practice environments, adapting the AI platform to operate ethically and effectively across jurisdictions ranging from New York to regions with restrictive regimes. This is underscored by rigorous governance frameworks that mandate client transparency about AI use, lawyer review of AI-generated work, and strict protocols preventing AI models from training on client data, thereby safeguarding confidentiality and reinforcing the firm’s commitment to responsible, fiduciary-grade AI deployment.

To facilitate seamless adoption, CoCounsel Legal is embedded within the document management and communication tools already familiar to Greenberg Traurig’s lawyers, including iManage, NetDocuments, SharePoint, Microsoft Word, Outlook, and Teams. This integration strategy significantly lowers barriers to daily use, ensuring that the AI platform complements existing workflows and enhances productivity without disrupting established practice routines across the firm’s global footprint.

CoCounsel Legal revolutionizes global legal workflows by unifying operations across jurisdictions with consistent quality standards, leveraging guided workflows and expert prompt libraries to bridge diverse legal cultures and experience levels. This integration enables seamless cross-border litigation preparation, contract management, and regulatory analysis from a single platform, ensuring ethical accountability through authoritative Thomson Reuters content and human-in-the-loop oversight that upholds professional standards worldwide.

By empowering in-house legal teams with advanced multi-jurisdictional research and contract standardization tools, CoCounsel Legal reduces dependence on outside counsel, accelerating transactions and reshaping client relationships. This shift prompts law firms to adapt billing practices, moving from initial drafting to reviewing AI-generated memos, as one practitioner observed, 'We would no longer ever ask them for a first draft... now what we're asking them for is to review our own AI drafted memo,' highlighting a fundamental transformation in legal service delivery and cost structures.

The adoption of AI platforms like CoCounsel Legal drives a surge in legal technology investment, with law firms increasing the percentage of revenue allocated to tech to enhance productivity and service quality. This hybrid model, combining machine efficiency with human expertise, not only boosts demand across practice areas but also necessitates rigorous data governance and ethical safeguards, as firms and clients enforce strict contractual provisions to protect sensitive information from unauthorized AI training use.

Beyond efficiency gains, CoCounsel Legal transforms litigation preparation by enabling lawyers to simulate scenarios and receive AI-driven critiques, drastically reducing brief production times from 20-30 hours to 2-3 hours while elevating quality. This evolution displaces routine tasks rather than entire jobs, fundamentally altering how lawyers are measured and compensated, with an emphasis on empowering partners to take agency over their careers and client relationships, reinforcing the enduring value of human judgment and advocacy in legal practice.