AI-powered authentication framework launched by Descope

Descope’s new Agentic Identity Hub addresses a critical infrastructure gap facing enterprises as they adopt AI agents into business workflows. The platform provides secure authentication and authorization between autonomous AI systems and enterprise applications, tackling an emerging challenge many organizations only discover after implementation begins. As machine identities already outnumber human identities by up to 45-to-1 in enterprise environments with projected growth of 150% in the coming year, this infrastructure solution becomes increasingly vital for maintaining security while enabling new AI agent capabilities.

The big picture: Descope has launched a specialized identity platform that solves authentication challenges between enterprise applications and the growing ecosystem of AI agents operating on behalf of users.

• The Silicon Valley startup’s Agentic Identity Hub enables secure, standards-based connectivity that traditional authentication methods weren’t designed to handle.
• The solution arrives as organizations increasingly deploy AI agents to automate tasks and improve productivity, creating new security vulnerabilities at machine scale.

Key components: The platform architecture consists of three integrated elements that work together to establish a secure identity infrastructure for AI agent interactions.

• Inbound Apps capabilities make applications and APIs agent-compatible through OAuth-based authentication with user consent screens, creating secure pathways while keeping users informed about what data and actions agents can access.
• Outbound Apps functionality streamlines connections to external tools and enterprise systems with pre-built templates for over 50 third-party applications, eliminating complex OAuth implementation burdens.
• MCP Auth SDKs and APIs secure remote Model Context Protocol servers, addressing authentication needs for this rapidly growing protocol developed by Anthropic that standardizes how large language models connect with external tools.

Why this matters: The scale of machine identity proliferation significantly magnifies security risks for organizations adopting AI agents.

• Research indicates 76% of security leaders anticipate machine identity growth of up to 150% over the next year, creating urgent authentication challenges.
• The platform’s fine-grained authorization approach combines both roles and scopes, representing a shift from traditional coarse-grained API authorization based solely on roles.

Technical innovation: Descope’s solution accommodates complex authentication realities that traditional systems weren’t designed to handle.

• AI agents often request API access on behalf of individual users and may interact with multiple APIs within a single function call, requiring more sophisticated authorization models.
• The system stores tokens with different permission scopes at both user and tenant levels, enabling both individual and organization-wide AI agent operations.