Usage & Consumption Telemetry
Usage telemetry describes observed consumption, not intrinsic Product quality.
Typical signals include:
- active consumers or entitled consumers;
- port/API invocations;
- batch or event consumption volume;
- agent/tool invocations;
- physical-control sessions where applicable;
- successful/failed consumption attempts;
- retention or repeated use;
- consumption cost and resource use.
Context is mandatory
A usage metric should include:
- Product version;
- observation window;
- source or telemetry reference;
- aggregation method;
- consumer segment or context where policy permits;
- unit and metric definition.
Adoption is not ontology
High usage does not make a Product:
- more intrinsically "AI";
- Primary rather than Derived;
- lower risk;
- universally fit for purpose.
Likewise, low adoption does not invalidate AI Product qualification.
Usage may indicate that Product intent, discoverability, economics, usability, or fitness should be reviewed, but those conclusions require additional evidence.