Human oversight is not an afterthought in Digital Humans — it is the structural foundation that makes the platform viable for enterprise deployment. Every digital employee operates within a defined approval framework that specifies which actions they can take autonomously, which require human review, and which are prohibited entirely.
The approval framework is organized into four tiers:
Tier 1 — Autonomous Actions. Routine, low-risk, high-volume tasks that digital employees execute without human intervention. Examples: a Customer Support Digital Employee responding to a password-reset request with a verified email link; a Finance Digital Employee categorizing a recurring software subscription expense under $500; a Research Digital Employee compiling a daily industry news digest. These actions are logged and auditable but require no human approval.
Tier 2 — Notification Actions. Tasks where the digital employee executes but notifies a human supervisor. Examples: a Sales Digital Employee sending a follow-up email to a prospect after a demo (notification to the account executive); a Marketing Digital Employee scheduling a social media post (notification to the marketing manager); an Operations Digital Employee updating a project status to "at risk" (notification to the project owner). The action proceeds immediately but the human has a window to intervene if needed.
Tier 3 — Approval-Required Actions. Tasks that require explicit human approval before execution. Examples: a Finance Digital Employee initiating a payment above $5,000; an HR Digital Employee scheduling a performance review meeting; a Procurement Digital Employee selecting a vendor for a contract above $50,000. The digital employee prepares the action, presents it with supporting context (why this vendor, what are the alternatives, what is the budget impact), and waits for approval.
Tier 4 — Prohibited Actions. Tasks that digital employees are never permitted to execute, regardless of approval. Examples: terminating an employee, signing a legal contract, making a strategic acquisition decision, overriding a security policy. These actions are hard-coded into the agent's permission set and cannot be overridden through the approval workflow.
Escalation rules define what happens when actions stall: if a Tier 3 approval request isn't acted on within 4 business hours, it escalates to the approver's manager. If still unaddressed after 8 hours, it reaches the department head. Emergency actions (suspected fraud, security breach, system failure) bypass the normal escalation chain and route directly to a designated incident response team.
The blueprint includes a full permission matrix template covering all eight department role templates, with configurable thresholds, escalation chains, and override protocols. Organizations can customize every threshold and approval path to match their risk tolerance and operational tempo.