Skip to main content
Version: Latest

Use Cases

Every organization faces the same fundamental problem: authorization is fragmented across applications, and adding new partners, jurisdictions, or AI agents makes it worse. PBAC solves this with one policy plane that adapts to any domain — the same authorization model that governs a corporate workspace also governs cross-jurisdictional health data sharing or regulated financial services. What changes is the data, not the infrastructure.

Scenarios

ScenarioThe problemHow PBAC helps
EnterprisePost-M&A identity chaos, per-app authorization, no unified auditOne policy plane across business units and acquired companies
HealthcareCross-jurisdictional data sharing with different privacy rules per jurisdictionFederated authorization with consent-as-obligation and full audit
FintechRegulatory mandates for delegated, consent-based access across institutionsStandards-based authorization with dynamic partner onboarding
AI AgentsAutonomous agents accessing internal systems with no auditable authorization modelTrust-tiered agent tokens with single-use issuance and per-call policy

AI agents and MCP

The AI agents use case shows how PBAC governs autonomous software clients with the same policy model used for humans:

  • Software statement claims carry agent identity (agent_provider, agent_model)
  • OPA enforces trust tiers at token issuance — no code changes to add or demote a provider
  • MCP servers are protected via real-time token introspection on every tool call
  • JIT single-use tokens and reduced TTLs constrain agent access windows
  • Agent-to-agent delegation via RFC 8693 token exchange

Next steps