1. Keep data and configuration tenant scoped
Workspace access, templates, jobs and plan entitlements are designed around tenant boundaries. Users should only process documents inside the authorised workspace for that customer or business unit.
2. Use plan-based retention as the default policy
Current plan definitions include document-history periods that are enforced by the retention workflow for eligible terminal jobs. Retention can vary by plan and contractual requirement.
The current self-service defaults are defined in the approved plan catalog. Contracted deployments may require a different approved policy.
3. Use controlled account-data and deletion requests
Customer account-data export and deletion requests are reviewed through controlled account workflows. Destructive actions are not performed silently from a user interface request.
4. Do not use customer data for global training without approval
Customer corrections and operational documents should not be used to train a general-purpose MicroExtract model without explicit authorisation and a lawful basis. Tenant-specific learning, where enabled, should remain approval gated.
5. Do not promise a self-service processing region selector
Processing location depends on the deployed environment and contracted architecture. The standard self-service product does not currently expose a customer-controlled region selector.
Data residency or processing-region requirements should be confirmed during solution scoping before contractual commitments are made.
6. Keep human review where risk requires judgement
Confidence, validation and automation do not remove accountability. High-impact, ambiguous or exception cases should remain reviewable by an authorised person, with the decision kept traceable.
Need help with a real document workflow?
Use representative samples, define the required output and validation rules, then test before production use.