From homegrown AI applications and code assistants to autonomous AI agents — is your security keeping up?

Organisations are developing homegrown applications and adopting agentic AI to automate workflows, interact with systems and complete complex tasks. While these technologies can accelerate innovation, they also create an expanding attack surface that traditional application security controls often struggle to protect.

Join Brian Martin from Integrity360 and Ben Bloomfield from SentinelOne as they explore how attackers target custom-built applications through insecure code, exposed APIs, vulnerable dependencies, weak identity controls and cloud misconfigurations. They will also examine the distinct risks introduced by agentic AI, including prompt injection, excessive permissions, data leakage, tool misuse and the manipulation of agents into performing unauthorised actions.

Attendees will learn how to build security into the application development lifecycle, assess the permissions and connections given to AI agents, apply effective guardrails and maintain continuous visibility over how applications and autonomous agents behave. 

This session will provide guidance into safeguarding your organisation's software development while facilitating the secure integration of agentic AI.