A shared vocabulary for cognitive runtimes, domain governance, provenance, human oversight, replay, recovery, and responsible intelligence.
Cognitive intelligence infrastructure
A governed platform boundary for systems that understand context, reason across constraints, decide accountably, act within policy, and learn from reviewed outcomes.
cognitive-runtimedomain-intelligence-layer
Cognitive runtime
The coordinated execution boundary connecting context, memory, knowledge, reasoning, planning, constraints, provenance, action, observation, and recovery.
decision-provenancedeployment-control
Domain intelligence layer
The layer that adapts shared cognitive capabilities to domain concepts, data sensitivity, policies, workflows, users, and oversight.
cognitive-runtimegovernance-profile
Decision provenance
Approved references and status that help a reviewer understand the context, evidence, constraints, proposal, approval, execution, and observation associated with a decision.
auditabilityreplay
Governance profile
A domain or application-specific set of policy, role, approval, escalation, data, security, and audit expectations.
human-in-the-loopconstraint-evaluation
Constraint evaluation
The review of policy, role, data, environment, safety, and domain conditions before a workflow may propose or execute an action.
governance-profilehuman-in-the-loop
Human-in-the-loop
A workflow pattern in which a person reviews or approves a proposal before consequential execution.
human-on-the-loopescalation
Human-on-the-loop
A bounded workflow pattern in which a qualified operator supervises execution and can intervene, pause, or recover under policy.
human-in-the-looprecovery
Auditability
The ability to review approved workflow events, decisions, actions, observations, changes, and corrections at the appropriate data and retention boundary.
decision-provenancereplay
Replay
A controlled reconstruction or simulation of workflow sequence and state for review; replay is not automatically production execution.
auditabilityrecovery
Recovery
Domain-owned pause, rollback, compensation, correction, or escalation behavior used when a workflow cannot continue safely or within policy.
replayhuman-on-the-loop
Assistive intelligence
Intelligence that supports a person or system workflow while preserving review, accountability, provenance, and domain-specific constraints.
human-in-the-loopresponsible-intelligence
Responsible intelligence
The practice of making assistive system behavior bounded, reviewable, explainable at the approved level, correctable, and accountable to people and policy.
assistive-intelligencegovernance-profile
Deployment control
The environment, release, identity, observability, rollback, and ownership boundary reviewed before a workflow is made available.
governance-profilerecovery
Escalation
A controlled transfer to a qualified person, domain owner, or approved service boundary when uncertainty, consequence, or policy requires additional review.
