Account and permission inheritance
ZenovaOS AI works through existing monitoring platform accounts, so plant and device visibility follows your existing account scope.
Security
Security, data governance, and verifiable output
Account boundaries, controlled data access, quality review, report validation, and transparent unavailable-data statements.
Controlled data-access boundaries
AI capabilities use controlled data-access paths for plant, device, alarm, trend, report, and knowledge workflows instead of direct ad hoc access.
Transparent unavailable-data reasons
If the monitoring platform or device collection layer does not provide a field, the answer explains that boundary instead of inventing a value.
Pre-launch review
Scenario checks review evidence, charts, layout, and unsafe fallback patterns before work is treated as production-ready.
Data governance and verifiable output
Every boundary exists so the output can be verified.
Account boundary: AI visibility inherits your existing accounts and permissions — no side doors.
Controlled access boundary: agents access platform data through controlled services and never hold user tokens directly.
Pre-launch review: scenario output must pass an acceptance check before it reaches production.
Report review: deep reports pass a quality-validation role before they are shared with management.
Transparent unavailable-data statements: missing fields come with verified reasons, not invented values.
Authorized data boundary
AI only works inside the authorized data boundary.
This diagram shows that ZenovaOS AI does not bypass monitoring platform permissions, does not hold user tokens directly, and keeps outputs traceable to authorized data, controlled retrieval, and pre-launch review.
01 Existing account permissions
Your existing monitoring account permissions determine visible plants, devices, reports, and history range.
02 Controlled internal access
AI capabilities retrieve data through controlled services without exposing passwords or tokens to the model.
03 Evidence-based reasoning
AI generates answers, charts, and reports from real data, business definitions, and knowledge-base evidence.
04 Quality-reviewed output
Scenario gates and report validation check numbers, charts, missing-data explanations, and deliverable format.
Data policy
Traceable AI starts with narrow data boundaries.
The pilot security review should confirm which accounts, plants, reports, device metrics, and mobile roles are in scope before a workflow is evaluated.
Keep generated answers grounded in monitoring platform returns, standard metrics, or documented business rules.
Avoid exposing passwords, tokens, refresh tokens, or infrastructure secrets in AI-visible outputs.
Use authorized historical evidence retrieval, not for bypassing account scope.
Review report outputs through task-level quality checks before sharing management reports.
Next step
Once the security boundary is clear, move into a controlled pilot.
Validate with one account scope, one workflow, and one set of acceptance metrics before expanding to more plants and roles.