Your AI Agents Have Credentials. Now What?
A 30-minute conversation on the 3 things AI breaks about identity, and the key questions to take back to your team
AI agents already hold credentials inside most enterprises. They request access, read data, and take action — often inside a browser session that belongs to a human. In 30 minutes, Michael Leland, Field CTO at Island, and Andrej Safundzic, CEO at Lumos, walk through what agent identity breaks, where it breaks, and what to do about it.
You’ll come away understanding:
- Why identity programs built for employees do not stretch to cover AI agents
- The assumptions about ownership, timing, and oversight that agents break
- Why granting access and governing what an agent actually does are two different problems
- What good looks like when agents hold privilege, from ownership through expiration
- How to decide how much of your governance to automate, and what to require before you do
Speakers
Who should attend
This session is for security leaders and the identity practitioners who work with them, including anyone who owns IGA, access reviews, or non-human identity. It will be most useful if your developers are standing up AI agents faster than your access request queue can absorb, or if you have agents in your environment today and cannot say who owns them. No prior knowledge of agentic AI assumed.
.svg)
.svg)
