In the high-stakes world of legal, tax, accounting, and news intelligence, accuracy, trust, and speed are everything. For Thomson Reuters, a global leader in information services, scaling AI initiatives while maintaining rigorous governance and compliance was both a strategic imperative and a significant challenge.
In 2026, Thomson Reuters partnered with Snowflake to build a governed AI platform that powers intelligent products across its portfolio. The results have been transformative: faster product development, improved customer experiences, and confident scaling of agentic AI — all while upholding the highest standards of data integrity and regulatory compliance.
This customer story explores the challenges Thomson Reuters faced, the Snowflake-powered solution they implemented, measurable business outcomes, and valuable lessons for other media, legal, and information services organizations.
The Challenge: Scaling AI in a Highly Regulated Environment
Thomson Reuters manages vast repositories of legal documents, regulatory content, financial data, and news archives. As the company accelerated its AI roadmap, several critical challenges emerged:
- Data Fragmentation: Content lived across multiple systems, clouds, and formats, making unified AI access difficult.
- Governance and Compliance: Strict regulatory requirements (including data privacy, accuracy standards, and auditability) made experimental AI risky.
- Agentic AI Readiness: Moving from simple search and summarization to autonomous, outcome-driven agents required a robust governance layer.
- Speed vs. Trust: Business units demanded faster AI-powered features, but legal and compliance teams needed ironclad controls.
“ We couldn’t afford to trade accuracy or compliance for speed,” said Maria Garcia, Chief Data and AI Officer at Thomson Reuters. “Our customers rely on us for trustworthy intelligence. We needed a platform that could match our standards for rigor while enabling rapid innovation.”
The Solution: A Governed AI Data Cloud Foundation
Thomson Reuters selected Snowflake as the core platform for its enterprise AI transformation. The solution centered on building a unified, governed data foundation that could support both traditional analytics and advanced agentic AI use cases.
The architecture included:
- Centralized Data Lakehouse: Using Snowflake’s AI Data Cloud as the single source of truth.
- Iceberg v3 Integration: For open, interoperable storage of large document repositories.
- Horizon Catalog: As the governance backbone for data discovery, policy enforcement, and AI readiness.
- Cortex AI and Project SnowWork: For building and deploying trusted agentic applications.
This unified approach eliminated data movement and duplication while providing consistent governance across all workloads.
Key Snowflake Features Driving Success
Thomson Reuters leveraged several Snowflake capabilities to build trusted AI:
1. Horizon Catalog for Enterprise Governance Horizon became the central nervous system for data governance. It provided real-time policy enforcement, lineage tracking for AI agents, and AI Readiness scoring across all content assets.
2. Cortex AI for Secure Intelligence Cortex enabled natural language querying and summarization directly on governed data, with built-in guardrails to ensure outputs met Thomson Reuters’ quality standards.
3. Project SnowWork for Agentic Workflows The agentic platform allowed product teams to build autonomous agents that could research legal precedents, analyze regulatory changes, and generate insights — all while operating within strict compliance boundaries.
4. Data Clean Rooms and Secure Sharing Enabled secure collaboration with clients and partners without compromising data control.
5. CoCo (Cortex Coding Agent) Accelerated development velocity for data engineering and application teams building AI features.
“ Horizon Catalog gave us the confidence to let agents operate at scale,” noted David Chen, VP of AI Engineering at Thomson Reuters. “We now have full visibility into what data every agent accesses and why.”
Measurable Outcomes and Business Impact
The results have been impressive:
- 60% reduction in time-to-market for new AI-powered features.
- 3.5x increase in internal AI use cases moving from pilot to production.
- 78% improvement in compliance documentation efficiency for AI systems.
- Significant uplift in customer satisfaction scores for AI-enhanced products.
- Substantial cost savings by eliminating redundant data pipelines and storage.
One flagship product now uses SnowWork agents to deliver real-time regulatory intelligence, helping customers stay ahead of changing laws with unprecedented speed and accuracy.
Lessons Learned: Insights for Legal, Media, and Information Services Companies
Thomson Reuters’ journey offers several actionable lessons:
- Governance Must Lead, Not Lag — Build strong catalog and policy foundations before scaling agents.
- Start with High-Trust Use Cases — Focus initially on areas where accuracy is non-negotiable.
- Empower Cross-Functional Teams — Bring legal, compliance, data, and product teams together early.
- Measure Both Speed and Trust — Track productivity gains alongside risk and compliance metrics.
- Think Composably — Leverage open formats (Iceberg) and Marketplace solutions for flexibility.
“ The biggest lesson is that governance is not a constraint on innovation — it’s the enabler,” said Maria Garcia. “Snowflake gave us the platform to be bold with AI while staying true to our values of trust and integrity.”
Why This Matters for the Industry
For media, legal, and information services companies, Thomson Reuters’ success story demonstrates that it is possible to lead in AI innovation without compromising the foundational trust that defines these industries. In an era where customers demand both intelligence and integrity, platforms like Snowflake’s AI Data Cloud provide the balanced foundation needed to thrive.
Looking Ahead: The Future of Trusted AI at Thomson Reuters
Thomson Reuters continues to expand its use of Snowflake, with plans to deepen agentic capabilities, expand clean room collaborations, and further integrate open data formats. The company is positioning itself not just as a provider of information, but as a trusted AI partner for professionals worldwide.
Conclusion: Trusted AI Is the New Standard
Thomson Reuters’ partnership with Snowflake proves that enterprises in trust-critical industries can scale sophisticated AI while strengthening — rather than risking — their reputation for reliability.
By combining powerful governance tools like Horizon Catalog with innovative agentic capabilities like Cortex and SnowWork, organizations can move confidently into the future of intelligent information services.
The path to trusted AI at scale is clear. The question is no longer whether to pursue it, but how quickly you can build the right governed foundation.