AI Dev 26 X SF | Idan Raman: The Identity Crisis of Browser Agents

Why It Matters

As enterprises move to agent-driven workflows, legacy authentication models and ad-hoc proxying create major security, reliability, and compliance risks—pushing firms to adopt specialized identity and network infrastructure. Companies that fail to invest in agent-native identity controls risk credential leakage, access blocks, and operational debt.

Summary

Idan Raman, CEO of Ankor, laid out what he calls an "identity crisis" for browser-based AI agents, arguing that retrofitting 20 years of human authentication for autonomous agents uncovers deep, unexpected complexity. He described common but insecure customer approaches—embedding or obfuscating credentials in prompts—and explained why those fail at scale due to credential exposure, bot-detection barriers, and unreliable third-party proxy IPs. To solve these issues, Ankor has built integrations with bot-detection vendors, its own enterprise-grade VPN, and secure agent infrastructure to present trusted, consistent identities to websites. Raman illustrated the wider pattern with a cautionary billing-system anecdote: seemingly simple features often balloon into costly, brittle in-house systems.

Original Description

As computer-use models become smarter, the bottleneck for their adoption is becoming clear: 20 years of web identity tech must be securely adapted for AI agents. In this talk, Idan Raman from Anchor Browser explored how enterprises and builders are solving this challenge.