Building Enterprise AI on Azure: Security, Governance

AI pilots are easy. Enterprise AI is harder.
This session will explore how organisations can move from isolated AI experiments to secure, governed and reusable AI platform patterns on Azure. We’ll cover what an AI application actually looks like, the threat landscape for agentic systems, and how a centralised AI Gateway can help enforce controls like least privilege, model guardrails, rate limiting, content safety, observability and cost management.
we’ll also explore out how these patterns can support thinking around PSPF and ISM alignment, including classification, authorised access, residual risk, agent permissions, prompt injection controls, sensitive data exposure, and monitoring of AI model usage.

Agenda
5:00pm - Mingling
5:30pm - Presentation and open discussion
6:30pm onwards - Food / networking

The session is designed to be practical and conversational, with audience participation encouraged.