Application
AI implementation checkpoint
Before AI changes the work, understand the knowledge the work depends on.
Before implementation
- What knowledge currently makes the process possible?
- Who understands difficult exceptions?
- Where are the authoritative sources?
- Which human capabilities remain consequential?
- Which dependencies already exist?
After implementation
- What knowledge is now generated or transformed by the system?
- Which human skills are being used less often?
- Can users still explain and challenge consequential outputs?
- Is provenance sufficient for the intended use?
- Could the organisation continue or recover if the system or supplier became unavailable?
This checkpoint is a knowledge-assurance lens. It is not a complete AI governance, legal, safety, privacy or conformity framework.