Aegis is the autonomous governance agent engine — running continuous monitoring, escalation, and assurance workflows across the Citadel platform.
Manual risk classification doesn't scale beyond a handful of AI systems. Obligation mapping is error-prone and inconsistent. Evidence gaps go undetected until audit — when it's too late to fix them without a scramble.
Aegis agents analyse AI systems against EU AI Act risk tiers, sector-specific requirements, and your organisation's risk appetite. Classification proposals — not decisions — presented for human review.
Each AI system is mapped to its applicable obligations across EU AI Act, PRA SS1/23, FCA Consumer Duty, ISO 42001, and GDPR. Complete, consistent, auditable.
Aegis continuously scans for gaps — obligations without controls, controls without evidence, systems without risk assessments. Alerts surface before auditors find them.
Every Aegis output is a proposal. Risk classifications, obligation mappings, and gap alerts all require human confirmation. An acceleration layer, not a black box.
Aegis is the intelligence layer — it reads context from Atlas and pushes actions to Sentinel.
Aegis agents read the governance knowledge graph to understand system relationships, existing obligations, and control coverage before making any classification or mapping proposals.
Confirmed risk classifications and obligation mappings create tasks in Sentinel — review cycles, approval workflows, attestation requests, and monitoring checkpoints.
Aegis is deployed in Phase 2 alongside Sentinel. Once your AI inventory and governance graph are established (Phase 1), Aegis agents begin autonomous classification, mapping, and gap detection — always with human-in-the-loop confirmation.
Book a demo to see how governance agents accelerate risk classification, obligation mapping, and evidence gap detection.
Free practitioner analysis of AI governance developments. Not press-release summaries.
Join compliance leaders from regulated industries. Unsubscribe anytime.