At this year’s Oktane 2025 conference, the spotlight was firmly on AI security—and for good reason. With AI agents now embedded in enterprise workflows, Okta positioned their security as a top-tier priority, placing them on the same footing as human employees when it comes to identity and access management. The message was clear: in the era of artificial intelligence, AI agent security is enterprise security.

AI Agents as “First-Class” Identities

One of the most significant announcements was the introduction of Okta for AI Agents. This new solution elevates AI agents to “first-class” identities within the Okta ecosystem. Much like employees, these agents will now be subject to:

• Governance frameworks

• Granular access controls

• Continuous auditing

This shift ensures that AI agents no longer operate as shadow processes but as managed entities within the enterprise identity fabric.

A New Standard for Interoperability

Okta also announced the Cross App Access initiative, an open protocol developed in partnership with industry standards bodies. This extension of OAuth authorization controls simplifies how AI agents communicate across applications while maintaining robust security. In essence, it sets a foundation for consistent and cohesive security standards across the entire tech landscape.

Tackling “Agentic Sprawl”

As organizations deploy more AI agents, they risk falling into agentic sprawl—a state where forgotten or over-privileged agents create hidden vulnerabilities. Okta’s platform updates directly address this challenge by offering:

• Discovery tools to surface rogue or unmanaged agents

• Governance features to enforce appropriate access levels

• Visibility dashboards to prevent sprawl from escalating into breaches

Security as an Enabler of Innovation

Okta’s CEO, Todd McKinnon, emphasized that securing AI is not about slowing progress but about enabling responsible innovation. By embedding AI security into the core of identity management, Okta seeks to prevent enterprises from undoing a decade’s worth of progress in cybersecurity.

Collaboration Across the Industry

While Okta is taking the lead, the company stressed that no single vendor can secure AI alone. The open standards unveiled at Oktane are designed to foster collaboration across the identity and security ecosystem, making AI safety a shared responsibility.

Why It Matters

The key lesson from Oktane 2025 is straightforward:
As AI agents evolve into an integral part of the workforce, treating their identity and access management as optional is no longer viable. Organizations that adopt AI without securing it risk exposure to the same vulnerabilities once faced in the early days of cloud adoption. The future of enterprise security now hinges on recognizing AI agents as part of the identity fabric—and managing them with equal rigor.

At LeLaboDigital, we see this shift as a pivotal moment. Just as cloud security redefined enterprise architecture a decade ago, AI identity security will define the next era of digital transformation.

Rami Zebian
CEO, LeLaboDigital