Status
As of 27 July 2026, several prototypes have been built, and full-scale product development for production environments is progressing in parallel. Pilot validation projects with selected enterprise customers are planned to test how well the product fits real AI development and operations, and to assess its impact on risk assessment and response workflows. Findings from these pilots will be used to refine the product, strengthen the underlying risk and action libraries, and prepare the governance workflow for wider production use.
FLAGSHIP PROGRAMME
Continuous AI risk governance
Summary
Enterprises are scaling AI across business functions, but few can clearly show which AI systems are in use, what risks they create, and how those risks are being managed over time.
The AI Risk Governance Product is an enterprise platform co‑created with professors from the University of Oxford, global leaders in AI research and governance, that uses a broad incident and risk library together with practical, expert‑validated response actions to guide AI use cases toward safe, continuous operation. It connects information capture, risk analysis, recommended actions, and ongoing review in a single workflow, so governance keeps pace with the speed of AI adoption.
The problem
As AI development and deployment accelerate, many organisations struggle to maintain a reliable, up-to-date view of all AI systems in use. Critical details are scattered across documents, tools, and teams, making information gathering slow and manual. Risk assessments take time and specialist effort, and even when key risks are identified, it is often unclear what to prioritise and how to respond in a practical way. As AI systems change and new technologies appear, it is hard to keep risk assessments up to date when processes are ad hoc or rely on a few individuals.
Solution
The AI Risk Governance Product provides an endtoend workflow for AI risk management. It uses large language model capabilities to capture and structure AI development and operations information, then draws on a dedicated risk library built from global AI incident data and usercontributed cases to generate risk scenarios tailored to each system. It also offers a curated library of recommended actions and checks, developed jointly with professors at the University of Oxford, recognised as AI leaders in both research and governance, so teams can move from identifying risks to taking practical next steps.
All evaluations and actions are stored in one environment, supporting periodic reassessment as systems evolve and helping organisations shift from oneoff reviews to continuous governance.