As companies give AI agents access to internal systems, a new cybersecurity problem is emerging: how do you secure software workers that can act on their own?

Startup Cymphony has raised $30 million, including a $25 million Series A co-led by Sequoia Capital and SMBC Fin Atlas Beyond Fund. The company is developing technology designed to help enterprises manage the security of their growing AI workforce. (TechCrunch)

AI agents are becoming corporate identities

Traditional enterprise security was built around human employees.

An employee receives an account, is assigned permissions and uses specific applications.

AI agents complicate that model.

A single agent can potentially interact with multiple applications, databases and external services, sometimes at machine speed.

Cymphony’s premise is that companies need ways to monitor and control these digital workers as they become more deeply integrated into business operations. (TechCrunch)

A new security layer

The startup’s funding comes as businesses increasingly experiment with autonomous AI systems.

An agent might handle customer support, analyze documents, create reports or interact with enterprise software.

Each capability creates another potential access path into corporate information.

That means security teams need to understand not just who has access to a system, but which AI agents have access, what they can do and how those permissions are being used.

The rise of non-human workers

The term “AI workforce” is becoming increasingly common because agents are beginning to perform activities traditionally associated with employees.

But software agents do not behave exactly like people.

They can operate continuously, process huge amounts of information and potentially execute many actions in a short period.

That makes conventional employee-security controls difficult to apply directly.

A new startup category

Cymphony is part of a growing group of startups building security infrastructure specifically for AI agents.

The category overlaps with identity management, cybersecurity, governance and observability.

The underlying question is increasingly straightforward:

If an AI agent can act like an employee, should it have its own identity, permissions and security controls?

As businesses deploy more agents, startups attempting to answer that question could become an important part of the enterprise technology ecosystem.