Fireside Chat: Buyer and Creator's Perspective of Agentic AI Risks

🎯Agenda

• Decode the production AI stack — Understand what enterprise AI deployments actually look like today, from model serving to the application layer, and how the choice between managed APIs and self-hosted inference shapes your security posture.
• Identify real-world AI threat vectors — Move beyond theoretical risks to learn which AI-specific attacks — prompt injection, data exfiltration, model poisoning, supply chain compromises — are actually being exploited in production environments.
• Architect AI gateways without the performance tax — Explore how to design gateway layers that deliver routing, filtering, and scanning across multiple LLM providers without becoming a latency bottleneck.
• Bridge the gap from AI security policy to runtime enforcement — Discover the patterns that separate organisations successfully operationalising AI security from those stuck in policy documents, including must-have evaluation criteria for AI security tooling.
• Prepare for the agentic AI attack surface — Get ahead of the next wave of risk as autonomous agents that call tools, browse the web, and execute code dramatically expand what enterprise security teams need to defend.