Why does Employee Read Tracking matter for AI governance?

That shift is not a problem in itself — it is usually an improvement. It becomes a problem when the organization's assurance still rests on read evidence that no longer reflects behaviour. A population showing near-zero opens and heavy assistant usage on the same policy has moved its consumption, and the governance question moves with it: what version is the assistant retrieving, and is the answer consistent with the text. Centralpoint holds both datasets, so the shift is measurable rather than invisible.