ACTTRIDENT GOVERN™

Govern the actor. Bound the authority.

Bind people, AI agents and digital workers to identity, purpose, delegated authority, supervision and runtime policy.

“Who or what is acting, under whose authority and within which limits?”

Agent Identity

Every agent and digital worker is bound to human sponsorship, purpose and delegated authority — with approval and supervision built in.

DEOS™ Agent Continuity

Memory and resource governance that preserves digital-worker continuity across sessions, models and handovers.

MarCosTrident™ Runtime Compliance

Turns regulatory obligations into runtime policy — enforced at the moment of action, with evidence to match.

ActTrident ASSURE

AI Security, Supervision, Usage governance, Resilience and Evidence — one assurance model for human-led AI, shadow AI and autonomous workflows.

ACTTRIDENT GOVERN / ASSURE

Govern AI use. Control AI actions. Evidence the outcome.

Govern AI use. Control AI actions. Evidence the outcome.

Govern AI use. Control AI actions. Evidence the outcome.

ActTrident ASSURE brings AI Security, Supervision, Usage governance, Resilience and Evidence together within ActTrident Govern, helping organisations govern human-led AI, shadow AI and autonomous workflows through one assurance model.

ActTrident ASSURE brings AI Security, Supervision, Usage governance, Resilience and Evidence together within ActTrident Govern, helping organisations govern human-led AI, shadow AI and autonomous workflows through one assurance model.

One assurance model for human-led AI, shadow AI and autonomous workflows.

A

A

AI

Human-led generative AI, AI-enabled applications and autonomous workflows.

ASSURE provides a common governance model across the different ways AI enters enterprise work.

S

S

Security

Protection against injection, information leakage, tool misuse and unauthorised actions.

Security controls should follow the action and its context, not rely only on whether the AI service itself has been approved.

S

S

Supervision

Accountable oversight and action-specific human approval where appropriate.

Not every AI action needs human approval, but consequential activity should have clear supervision and escalation boundaries.

U

U

Usage governance

Policies governing which AI services, data and activities are permitted, including shadow AI.

Govern AI use based on the action, data, service and organisational policy rather than relying only on blanket access controls.

R

R

Resilience

Controlled behaviour through outages, change, revocation and recovery.

Governance should continue when models, services, credentials, runtimes or delegated authority change.

E

E

Evidence

Inspectable and verifiable records of decisions, observed outcomes and coverage gaps.

Assurance should provide evidence of what happened, why a decision was made and where control coverage may still be incomplete.

Approve AI and agents in context.

Bound authority and tool use.

Preserve evidence. Expose gaps.

HOW ASSURE FITS THE ACTION

Govern autonomous action, not just autonomous identity.

Govern autonomous action, not just autonomous identity.

An AI agent can authenticate correctly and still attempt an action outside its delegated authority. Govern connects identity, sponsorship, purpose, runtime policy, supervision and evidence around the action itself — the continuity and delegation model DEOS™ provides, extended with ASSURE’s assurance dimensions.

Bring shadow AI into the governance model.

Bring shadow AI into the governance model.

Employees can adopt AI services faster than traditional governance can classify them. ASSURE helps organisations define permitted AI use, data boundaries and action policies while preserving evidence of governed activity and coverage gaps.

REQUEST

AI or agent requests an action

ASSURE EVALUATES

AI usage policy

Authority

Security context

Supervision requirement

Resilience state

ACTTRIDENT CAN

Allow · Advise · Confirm · Step up · Escalate · Block

EVIDENCE

Evidence preserved through Proof Fabric

Conceptual sequence. Enforcement responses vary by integration and policy.

Part of ActTrident Govern.

Enforcement and evidence connect with Protect and Prove; the platform remains Protect, Prove, Govern and Optimize.

Autonomous does not mean unaccountable.

AI may execute the action. Humans remain accountable for the outcome.