Shares of Okta rose more than 3% after the company detailed its expanding footprint in AI-agent security at the Goldman Sachs Communacopia + Technology Conference on Wednesday, presenting a product lineup and organizational changes aimed at capitalizing on enterprise demand for securing autonomous software agents.
Okta’s president and chief operating officer, Eric Kelleher, described “agentic identity” as the third phase in the company’s evolution — following its initial move into cloud computing and its subsequent focus on identity-driven security, which he noted accounts for roughly 80% of successful cyberattacks. The company launched “Okta for AI Agents” in general availability on April 30 and has since closed dozens of transactions, including a multimillion-dollar deal with a Fortune 50 healthcare company whose deployed agent count surged from 50 to 1,500 within weeks of implementation.
Okta published a document titled “Blueprint for the Secure Agentic Enterprise,” built around three security questions: where agents live, what they can connect to, and what actions they are permitted to perform. The company’s universal directory now treats agentic identities as a first-class category alongside human workforce identities, customer identities, and service accounts. Agent single sign-on is available at no additional charge to all workforce-platform customers, and identity-security capabilities such as governance, provisioning and auditability have been extended to agents.
Revenue breakdown for the second quarter showed workforce identity accounting for approximately 59% of total revenue, with customer identity making up the remaining 41%. New products contributed 30% of bookings, and Okta reported pipeline levels at record highs. The company said it now serves more than 20,000 customers with over 8,000 integrations.
Kelleher characterized the primary competitive threat as confusion in the market rather than any single rival. He pointed to the rapid pace of change among frontier AI model providers as a driver for Okta’s neutrality strategy. “Our buyers are seeing this need to be agile,” he said, noting that major models are leapfrogging each other every 30 to 45 days. Okta also proposed Cross-App Access as an open standard extending the Model Context Protocol, and its Auth0 for AI Agents offering supports that standard out of the box.
On the operational side, Okta completed the acquisition of Permiso on the same day the company announced its earnings, a move aimed at accelerating identity-threat protection. The company also secured IL5 authorization for its FedRAMP business, the highest non-classified level for federal government operations. Ric Smith, who joined about a year ago as head of technology from Oracle, Medallia and SentinelOne, oversees the company’s R&D direction.
In a sales overhaul implemented roughly 18 months ago, Okta reorganized its seller groups around buyer personas rather than product lines. One group specializes in the Okta platform for CIO and CISO buyers, while another focuses on the Auth0 platform for developer audiences. Both are supported by a dedicated AI conversation team. Most initial agreements for agentic identity are structured as one-year contracts, though some multi-year deals have been completed. Pricing remains an uplift on existing per-user pricing, a model Kelleher acknowledged is imperfect for agent architecture.
Okta shares closed at $172.74, up 3.07%, with after-hours trading at $172.51.
