Efforts led by the Agentic Authentication Technical Working Group and the Payments Technical Working Group will focus on three core areas. Verifiable User Instructions will enable users to authorize AI agents through “clear, phishing-resistant mechanisms” that do not expose user credentials.