See the controls behind every run.
Before a pilot, we agree who can act, what data a run can reach and what happens if something fails.
What we review with your team.
This is a working review, not a certification. We match each question to the checks and evidence agreed for the pilot.
Every person and agent acts through an authenticated identity with the right access.
Show sign-in, permissions and organisation administration working together.
Access reviewRequests, conversations and tool calls stay inside the active organisation.
Test two organisations, show the active context and stop on a scope mismatch.
Boundary testsJobs use named inputs and credentials, and leave logs and outputs the team can inspect.
Show the job definition, runtime boundary, logs, output location and failure path.
Runtime evidenceCustomers can see what enters Novem, where it is processed and what remains after a run.
Confirm the architecture, processors, retention and deletion process.
Data reviewReports and schedules stay useful when someone changes role or leaves.
Demonstrate ownership changes, permissions, run history and schedule authority.
Continuity reviewThe pilot has a clear operating model before customer data enters Novem.
Agree the service boundary, monitoring, incident route, backups, recovery and customer responsibilities.
Deployment planIf the boundary is unclear, stop.
Name the active context and stop. Do not search another workspace.
Explain which access is missing without exposing the protected resource.
Stop before publishing a new output and record why the run could not complete.
Show what was requested and stop until the required access is available.
Bring your questions before the pilot starts.
We agree who can act, where data goes, how runs fail, how long information is kept and who responds to an incident.