AI Agents Are the New APIs: Insights from Mashery Founder’s Latest Venture
APIs connected websites and apps during the Web 2.0 era. Now, AI agents are emerging as the new way to access and use data from across the internet. While APIs remain relevant, AI agents are quickly becoming the primary interface for data interaction. Managing these agents is shaping up to be the new form of API management.
Oren Michels, who founded Mashery—an API management company acquired by Intel in 2013—is leading this shift with his new company, Barndoor. Barndoor serves as the "control plane for agentic AI," aiming to help enterprises manage and secure AI agents effectively.
From API Management to AI Agent Control
Michels draws a clear parallel between the challenges of API management in the past and today's need to manage AI agents. While Identity and Access Management (IAM) and traditional API management tools handle some aspects, they fall short when it comes to agentic AI. According to Michels, this gap is why AI agents haven’t seen wider adoption in enterprises yet.
He points out that just as network traffic management companies weren't equipped to handle API management back in the day, existing security and management tools aren't fully up to the task for AI agents now. This creates a space for specialized solutions like Barndoor.
Who Needs to Use Barndoor?
The typical users of Barndoor aren't CIOs or security teams. While these roles are essential, they don't directly face the "AI problem." Instead, the people who need solutions are those whose work depends on faster, more efficient selling and operations powered by AI. They need tools to control and guide AI agents without compromising enterprise security or efficiency.
Michels compares this to the early API days when the real API challenges were faced by developers trying to connect mobile apps with backend services, such as hotel booking systems at Marriott. Now, it's about empowering frontline staff and business units to integrate AI agents into their workflows safely and effectively.
Why Managing AI Agents Matters
- Security: Guardrails are needed to prevent misuse or unintended actions by AI agents.
- Control: Enterprises require oversight on how AI agents access and manipulate data.
- Scalability: Managing multiple AI agents across departments without chaos.
Barndoor’s approach focuses on these core needs to enable enterprises to confidently deploy AI agents at scale.
Looking Ahead
As AI agents become more integral to business operations, companies will need specialized platforms to manage them—just like they needed API management platforms before. Those interested in learning more about AI agent management and how to integrate AI into business processes might find value in exploring Complete AI Training, which offers courses and resources on AI tools and automation.
Your membership also unlocks:
