The massive security risks associated with AI agents are rapidly becoming the dominant identity-related concern for customers of the vendor, Okta President and COO Eric Kelleher tells CRN.
While the security risks associated with AI agents are increasingly well-known within the business world, the growing dominance of the topic for many customers of identity security vendor Okta is striking, according to Okta President and COO Eric Kelleher.
In an interview with CRN Tuesday, Kelleher said that the need for management and security of agentic technologies is “consuming our conversations with customers right now.”
[Related: Mistaken Identity? AI Agent Oversight Key To Success]
“They view it as the most important strategic need they have around identity today,” he said.
The rising importance of security for AI agents has become especially apparent during Okta’s most recent quarter, which the company released financial results for Tuesday, surpassing Wall Street expectations.
“Over the past quarter in particular, it really has been made clear to us that this is the No. 1 security threat our customers are concerned about, regarding identity,” Kelleher said. “We believe we’re uniquely positioned to take the leadership role here.”
Key capabilities from the company for securing agentic identities include the vendor’s Okta for AI Agents offering, which was unveiled in September and aims to provide a streamlined integration of agentic identities into an organization’s broader identity system.
“We’ve promoted agentic identity to be a first-class citizen throughout our stack,” Kelleher said. “Our customers can discover and manage an agent very similarly to the way they discover and manage humans.”
This includes capabilities for vaulting credentials related to agents, instituting policies for rotating agent credentials and using identity governance to provision and de-provision agents in a “just-in-time” fashion, he said.
This means that partners and customers can “turn agents on and off when they’re needed and not leave them with standing permissions,” Kelleher told CRN. “They can implement all the security policies on agents that they can on humans. So we’re a trusted provider solving these problems already, and this allows us to extend for the newest version of this problem [related to] agents.”
Agents are still a nascent technology for many organizations, but are spreading rapidly, according to a recent Okta survey. The findings suggest that while nearly all organizations have already deployed agents—90 percent reported doing so—just 10 percent of surveyed organizations said they’re confident that the agents are properly governed and secured.
In other words, with the pressure to adopt new AI technologies and not be left behind, organizations are giving agents access to data and resources “before they’re confident they have a way of securely managing what [the agents] have access to,” Kelleher said.
For the third quarter of Okta’s fiscal year 2026, ended Oct. 31, the company disclosed that revenue climbed to $742 million, up 12 percent year over year. That was well ahead of Wall Street analysts’ consensus estimate of $730.4 million.
Okta’s earnings also beat expectations, with the vendor reporting that non-GAAP diluted net income per share came in at 82 cents for the quarter. That was higher than the 76 cents per share that analysts had been expecting, and above the 67 cents per share the company had reported for the same period a year ago.
