Service
AI security & governance assessment
Frame your use of generative AI: data, vendors, bias, traceability and accountability.
The problem we solve
Employees already use generative AI tools, often with company data and without a framework. Banning them does not work; doing nothing exposes you to data leaks, untraceable decisions and emerging regulatory obligations.
Who it is for
- Organizations rolling out Copilot, ChatGPT Enterprise or similar tools
- Software vendors building language models into their products
- Leadership teams wanting a clear, workable AI use policy
What is included
- Inventory of AI uses and tools in place
- Risk assessment per use case
- Acceptable use policy for AI
- Security review of language model integrations
- Governance and accountability framework
Our approach
We rank uses by the data involved and the impact of the decisions produced, then propose proportionate rules: allowed, supervised or prohibited. For products that embed a language model, we assess the risks specific to these systems, such as prompt injection.
Typical deliverables
Use register
Ranked by risk level
Use policy
Two pages, usable as soon as it is shared
Technical assessment
For each language model integration
Expected outcomes
- AI that is used, but governed
- Company data that stays under control
- An organization ready for upcoming regulatory requirements