Mathew Graham Chief Security Officer, APAC, Okta
“AI is creating two immediate challenges for enterprises: attackers can move faster and operate at greater scale, while businesses are deploying agents that need access to sensitive data and systems. Both require a stronger approach to identity security. At Okta, with Identity Threat Protection with Okta AI, we use AI and machine learning to continuously assess risk, drawing on signals from our platform and security partners. When risk changes, we can require further verification or end a session, helping contain threats before they spread. A successful sign-in should not mean access goes unchecked.
For AI agents, businesses need to answer four questions: Where are my agents? What can they do? What are they doing? How do we respond when risks emerge? Every agent needs an accountable owner, appropriate permissions and a traceable record of access. We’re a founding member of the Blueprint Alliance, working with leaders including AWS, CrowdStrike, Salesforce, and Zscaler on a shared architecture for securing AI agents.
Our identity security fabric brings together access management, governance, privileged access and threat protection for human, machine and AI identities. Access reviews and audit records provide evidence to support evolving compliance requirements. Enterprises should measure progress through reduced excessive access, faster removal of unnecessary permissions and quicker containment of compromised identities. The goal is to give businesses the confidence to adopt AI while maintaining visibility, control and accountability.”
