Personal AI Is Already Your Customer

Within weeks of launch, Meta Muse became the most common AI agent in the consumer traffic we see. Agents like it now log in, clear OTP, check out, and move money for real people. Most of them look like ordinary Chrome.

In this session we'll show live what these agents can do, sort them into three types, and cover the signals that catch agents that never identify themselves. Then we'll show how leading teams use an authorization matrix to decide, workflow by workflow, what an agent is allowed to do.

Key takeaways

• Agents are already in your flows, and most of them don't tell you.
• There are three kinds of agents: self-disclosing, non-disclosing, and malicious. Detection that depends on agents identifying themselves misses most of them.
• Decide by workflow, not just allow or block. Agents can check balances, but a human should approve moving money.
• Visibility comes first. You can't govern traffic you can't see.

Who it's for

CISOs and Fraud, Risk and Trust & Safety Leaders, Plus the Product, Digital and E-Commerce Teams Who Own the Checkout.

Speaker

Kevin Gosschalk — Founder & CEO, Arkose Labs, and author of After Bots

Can't make it live? Register anyway and we'll send you the recording.