Selected work
Thomson Reuters
Enterprise AI decision architecture
Context
Enterprise AI must support revenue-generating information products and internal operations without treating accuracy, privacy, or human review as an afterthought.
Executive mandate
Create a platform and operating model that can scale reusable AI capability while keeping the reasoning and review path visible to the people accountable for the outcome.
Architecture approach
Designed a federated enterprise AI architecture with human verification at high-stakes gates, structured challenge for ambiguous cases, and governance signals that turn correction into continuous improvement.
Resulting capability
- Reusable capabilities designed for product and operational contexts
- Human review retained where expertise and accountability matter most
- A clearer path from AI output to evidence, exception, and final decision