Artificial Intelligence | News | Insights | AiThority

Barndoor unveils AgentProfile, ending borrowed identities and blind costs for every agent

Barndoor AI | Secure Access and Governance for AI and MCP

Barndoor ties each agent’s profile, credentials, and spend to a single governed record, the first step toward governing the entire agent lifecycle

Barndoor AI, the governance platform for AI agents and MCP-connected tools, introduced AgentProfile, a new capability in its AI agent governance platform that assigns each agent a discrete profile of its own: a unique record, a human manager, and credential sets for the tools, models and data that agent’s job requires, independent of human credentials.

Most organizations running agents today have no single record that ties agent activities together. MCP access is usually managed in one place, model keys in another, with cost data and agent ROI generally hard to track or simply non-existent.

Legacy identity systems anthropomorphize agents

Legacy identity access management and governance solutions were built for humans who log in, stay authenticated, and log out or offboard when they’re done. Agents behave differently. They make tool calls at machine speed, and take action at runtime.

“Agents require their own governance, not something borrowed from humans,” said Oren Michels, CEO and co-founder of Barndoor. “We call these models enthusiastic interns for a reason; they don’t know right from wrong, they just want to do the job well, and they’ll use everything they’re allowed to use to get there. A human employee will pause and ask. Agents don’t have judgement built in, which is why they need a governance approach of their own, not a workaround built for people.”

Also Read: AiThority Interview with Matej Bukovinski, Chief Technology Officer at Nutrient

Barndoor raises the profile of every agent

Identity is no longer a one-time authentication requirement. Legacy IAM grants credentials at authentication and stops there. Barndoor governs every agent operation, the tools and models agents use, how they’re used and accessed. With AgentProfile, Barndoor enables organizations to fully govern agent:

Related Posts
1 of 43,584

CONTEXT

  • Establish every agent as its own profile, ensuring every action is mapped to an agent, not a shared human or machine identity
  • See everything in an agent’s profile: activities, tool usage, model calls, spend, and more

COST

  • See what every agent is spending. Because every model call and tool call logs back to the agent’s profile, organizations can see token usage and cost per agent, not pooled across a team or shared account

ACTION

  • Continuous enforcement: every agent action is governed continuously, not granted just at sign in. Legacy IAM authorizes access once and stops watching
  • Set least privilege agent permissions, not inherited from the human who deployed it
  • Scope and create credentials that are attached to the agent profile, not a user session
  • See everything in an agent’s profile: all agent activities, tool usage, model calls, spend, and more are tracked and monitored
  • Deprovision centrally. Remove an agent without impacting business continuity – cleanly remove access to all connected systems
  • Change policies without impacting agents. Since access policies are managed through the Barndoor platform, not in the agent’s code, switching all agents from one model provider to another is a configuration change, not a code change

“Barndoor is building a system that addresses how agents really operate, not retrofitted from human identity and access. It’s time to sunset identity-only governance and look to the future. AgentProfile is the next piece that manages agent profiles from one record, not fragmented across different places,” said Alexander Richard, Cyber Security Specialist, Syndio, the leader in Decision Intelligence for Pay.

Barndoor AI agent governance platform

Barndoor is the AI Agent Governance Platform, governing each agent’s profile, tool access, model access, cost, and data protection across the full agent lifecycle, from credentialing through runtime policy enforcement to deprovisioning.

Also Read: ​​AI systems – Interoperable AI systems: Connecting models across platforms

[To share your insights with us, please write to psen@itechseries.com ]

Comments are closed.