Skip to content

Plans & entitlements

A workspace’s plan determines which features it can use. Entitlements are how those features are gated and enforced across the product. Gating is a security boundary: a capability a plan does not include is enforced server-side, not merely hidden.

PlanFrameworksNotable inclusions
StandardCore set (GTSAF, EU AI Act, NIST AI RMF)Systems, intake, assessment, risks, policies, model cards, AI analysis, CSV export.
AdvancedCore plus MAESTRO, ACRS, NAGFThe full product except the agentic stack: control testing, workpapers, findings and evidence registers, reports, AI chat, policy generation, audit trail.
EnterpriseAll, including ATFEverything in Advanced plus the agentic stack: Agentic CISO, Gateway, Claw, and Discovery.
TrialAll, including ATFThe full product, time-limited, to showcase Enterprise capabilities.
PlatformAllGamut operator tier.
No AccessNoneLogin only.

The autonomous agent stack, Agentic CISO, Gateway, Claw, and Discovery, carries the highest operational and security risk, so it is reserved for Enterprise (and Trial). ATF is grouped with that stack: an agentic framework is only useful where there are governed agents to assess. ISO/IEC 42001 and 42005 assessment modules are never plan entitlements; they are enabled only after a valid licence confirmation is recorded.

Effective access is the intersection of two limits:

  • The plan tier the workspace is on.
  • The role product ceiling of the user. A Standard-role user is capped at Standard capabilities even on an Enterprise plan, and an Advanced-role user cannot reach the agentic stack.

A feature is available only when both the plan and the role allow it. This is why a capability can be present for one colleague and absent for another in the same workspace.

Beyond features, plans set numeric quotas, enforced server-side, that scale with tier:

  • AI analysis: daily and monthly call limits.
  • Policy generation: separate daily and monthly limits, zero unless the policy_generation entitlement is on.
  • Assessments: a maximum number per workspace.
  • Model level: which Claude models the plan may use, with a default model per tier.

Quotas keep AI usage predictable and are part of how Gamut keeps model usage governed.

Every gated capability binds an entitlement feature to one or more RBAC permissions, so a user needs both the role permission and the plan entitlement. The check runs server-side on every request, so the interface and the API agree and gating cannot be bypassed from the client.

Administrators can see the workspace’s current entitlements in Administration. To change plan or discuss enterprise capabilities, contact the Gamut team.