AI Infrastructure

We have implemented an AI steering committee and looking for ways to have visibility into data that leaves our company destined for AI SaaS platforms. We have an administrative policy and have mostly identified which AI platforms are being used.

* Identify platforms that help secure AI apps, copilots, chatbots, and agents across internal and external use cases. * Our main focus is to reduce the practical security risk of AI adoption as usage grows across 60+ AI & ML projects in 22+ entities. * This initiative is not about broad cloud posture or general GRC workflows. It is about securing AI driven applications at build time, test time, and runtime. * We want solutions that can help us address data leakage, prompt injection, unsafe outputs, policy violations, excessive permissions, and weak agent behavior before these risks scale further. * There is real urgency here. We expect this area to move in Q3 2027 as more AI use cases move closer to production and external exposure increases. Target: - Select 1 or more vendors for a focused PoC or PoV in Q2. - Validate whether the solution can improve security for AI applications and agents without slowing down delivery too much. - Prioritize vendors that can show fast value in areas like runtime protection, red teaming, policy enforcement, and visibility into how AI components behave. - Build a credible path from pilot to broader use across entities with different levels of AI maturity.

As our organisation increases the use of AI coding agents and autonomous workflows, we need infrastructure that allows these agents to operate safely without relying on unmanaged user devices, shared credentials or unrestricted access to production systems. This initiative is focused on secure execution environments, controlled infrastructure access, credential isolation, developer tooling and the operational foundations required to scale agents securely.

As AI agents are increasingly deployed across our organisation, we need consistent visibility and control over how they behave, what tools and data they access, and what actions they perform. This initiative is focused on establishing a dedicated security and governance layer for AI agents, including runtime monitoring, policy enforcement, MCP and tool governance, identity-aware controls, and auditable agent activity.