AI-driven automation and autonomous agents are increasingly embedded in exchange operations, market-making, custody, and on‑chain infrastructure. That introduces novel risk vectors—agentic decision loops, API abuse, model drift, data exfiltration, vendor and model-supply‑chain weaknesses, and regulatory gaps around AML/KYC and AI-specific rules. As of 2026-02-10, heightened regulatory scrutiny and the operational scale of multi-agent systems make comprehensive AI security and risk tooling a priority for crypto firms. This topic covers three linked categories: AI Security Governance (real‑time monitoring, behavior visibility, and incident response), Regulatory Compliance Tools (audit trails, policy enforcement, and vendor risk), and AI Governance Tools (model validation, provenance, and lifecycle controls). Representative tools include Xilos, which focuses on 100% visibility into connected services and agentic AI activity; StackAI and Relevance AI, enterprise no‑code/low‑code platforms for building, deploying and governing autonomous agents and workflows; Monitaur, an insurance‑grade governance platform centralizing policy, monitoring, validation, and vendor governance for regulated firms; Mistral AI, which delivers open, efficiency‑oriented foundation models and production tooling with privacy and governance features; and Vertex AI, Google Cloud’s managed end‑to‑end platform for model training, deployment and monitoring. Together these tool classes help exchanges and crypto infrastructure teams instrument agent behavior, enforce policy across model lifecycles, centralize vendor and model inventories, and produce audit-ready evidence for compliance regimes. Choosing the right mix depends on an organization’s exposure to market risk, regulatory footprint, and preference for managed versus open model stacks—balancing visibility, control, and operational agility while reducing systemic AI risk.