From AI Induction to AI Accountability

AI is intimidating not as a technology but as unexplored territory, and the real difficulty is not complexity but excess: in the digital world almost anything is possible, so projects drown in options. The frame has to come from your business, and you are not required to understand code, only your own services, prices and rules. This joint paper connects two halves of one accountability problem. Amanda Cunningham examines what must happen inside an organisation before AI meets the customer: AI usually arrives incrementally, through accounts and updates, without any formal moment of decision, so organisations acquire AI capability without ever defining AI authority. Her answer is AI induction, treating the system the way you would a new employee, with an eight-question readiness check for leadership: purpose, context, access, authority, boundaries, accountability, evidence, review. Sergei Ponomarev takes over at the moment the AI becomes customer-facing: Article 50 of the EU AI Act (applicable from 2 August 2026) requires telling people they are talking to an AI. That is the legal floor, not the customer proposition. A customer really has three questions: how does this company use AI at all, what can this particular bot do, and how do I prove my case if it goes wrong. Three documents answer them, an AI Policy, an AI Service Passport and an AI Receipt, and a periodic test purchase checks that the answers are true. The paper closes with an executive checklist and one claim: responsible AI is a loop that never stops turning. Intent, authority, experience, evidence, review.