As businesses give AI a greater role in their work, they need to understand how much control they retain over what it can do. An AI system may recommend decisions, communicate with customers, access information or take actions without somebody approving each one beforehand. A business needs to know when it would need to intervene and whether it would still be able to act when that point is reached.
The Business Intervention Test examines whether a business has retained the practical ability to intervene in its use of AI. Intervention may mean stopping an action, rejecting a recommendation, restricting access or requiring additional approval. It may also involve suspending a process, returning to another way of working or ending the relationship with a provider.
The central question is simple: if the way an AI is being used becomes unacceptable, could the business recognise the problem and act before the consequences become unmanageable?
The test begins by asking what authority the AI has been given and why the business needs to retain the ability to intervene. It then examines the boundaries the business has set, the circumstances that would require action, whether those circumstances could be detected and whether there would be enough time for somebody with the necessary authority to respond.
The checklist contains 11 YES/NO questions. A YES means that the business has a clear answer and can explain what supports it. A NO identifies a limitation, an uncertainty or something that has not yet been established. The questions also consider whether intervention would disrupt essential work, whether it would contain the problem, what evidence would remain afterwards and whether action depends on the provider.
The test does not produce a score, rating or simple pass/fail result. The answers need to be considered together. Having authority to stop an AI offers limited protection if the business cannot detect a problem in time. Equally, a control may be difficult to use if doing so would interrupt essential operations or require cooperation from a provider.
The benefit is a clearer understanding of where the business can act with confidence and where its control may be weaker than it appears. The checklist helps organisations identify those limitations before a problem forces them to rely on arrangements they have never properly examined.
The Business Intervention Test is a checklist, not an implementation framework. It does not prescribe particular controls or tell a business when to stop using AI. Its purpose is to help the organisation understand whether the ability to intervene remains a practical choice when it is needed.
A public reference standard that defines AI governance as a measurable structural condition, moving beyond policy statements to test whether control, accountability, traceability and integrity can be demonstrated in practice across the AI lifecycle.
A quantitative AI governance metric designed to measure how much of an AI system remains provable when models, APIs, vendors or other upstream dependencies change. ERR focuses on traceability, reconstructability, version fidelity and dependency proof to assess whether governance can survive change.
A practical governance test examining whether a business can maintain critical services if access to an external AI provider is suddenly restricted or withdrawn because of provider decisions, government action, sanctions, trade restrictions or geopolitical disruption.
A public reference framework for testing whether important AI protections actually work in practice, focusing on control, accountability, business dependence and effective oversight rather than relying on policies, promises or formal safeguards alone.
A practical business checklist examining why intervention in AI use may become necessary, what it is intended to protect and whether the organisation retains the authority, information and practical ability to act before consequences become unmanageable.