Artificial intelligence is generating a new cybersecurity dilemma. Companies increasingly need to govern not only who can access their systems, but also which AI agents may access them and what those agents can do once they’re inside.
That might be a big opportunity for Okta (OKTA), according to Bank of America Securities.
Okta’s Oktane 2026 conference led BofA analyst Tal Liani to lift his price target on the stock to $220 from $200, according to a Sept. 24 research note emailed to TheStreet.
The analyst kept the Neutral rating but pointed out the disconnect between the promise of Okta’s AI approach and how little it’s really doing for the firm now.
Okta used its annual conference to cast itself as an identity-security control layer for AI bots. Among the developments BofA called out were free Agent SSO, Shadow AI Agent Discovery, Agent Gateway, and Agent-to-Agent Connections.
Those technologies might help Okta go beyond standard authentication and into governance and security rules for what autonomous artificial intelligence systems can access and do, BofA said.
The opportunity that may arise later could be substantial.
The management message at the conference indicates that what Okta calls “Agentic Identity” might represent a larger total addressable market than identity security currently is, BofA said.
But investors may have to wait for that chance to make a significant difference in Okta’s financial outcomes.
Bank of America sees a bigger AI opportunity for Okta
BofA raises Okta’s price target, citing improved confidence in its AI potential, not a material shift in the near-term profits forecast.
The $220 price target is based on a 12 times expected calendar 2027 enterprise value-to-sales multiple for Okta, up from 11 times before, Liani said.
The higher multiple is a better reflection of the improvements in Okta’s narrative post-Oktane, the analyst added.
But BofA framed the conference as building on early momentum, not as anything that alters its investment thesis. The business thus kept its rating at Neutral, but lifted its price target.
At the heart of that possibility is a rather straightforward dilemma caused by more autonomous A.I. systems.
AI agents must have identity.
Those identities need permits. Companies need to know who agents may access, detect agents that are running without authority, terminate access when appropriate, and prevent hacked or deactivated agents from continuing to utilize credentials.
Okta wants its platform to be between the agents and the apps they utilize.
Agent SSO is a major component of that plan. BofA said it’s offered at no extra cost inside Okta’s basic single sign-on solution.
Agent SSO may offer agents a unique identity and short-lived tokens instead of using long-lived application programming interface keys.
The decision to open out the authentication component generally isn’t about making money from it immediately. It’s an issue of distribution.
“Free Agent SSO seeds the upsell,” BofA said.
BofA thinks free Agent SSO may lead organizations and software developers to embrace Okta’s agent identity standards and create a funnel for premium goods across discovery, governance, and runtime security.
The economics might get a lot more fascinating in those compensated skills.
Related: Okta’s AI boom just created a new security problem
BofA said Okta for AI Agents has features including Shadow AI Agent Discovery, Agent-to-Agent Connections, lifecycle governance, and Agent Gateway.
The paper notes planned features such as visual mapping of permissions between agents and resources and a “kill switch” that would deny requests made with tokens still in an agent’s possession after the agent is deactivated.
That difference is significant for BofA, since it takes Okta beyond just certifying an AI agent.
It may give the company a hand in restricting what the agent can really perform.
Okta could reach customers that use rival identity platforms
One of the most crucial portions of Okta’s AI approach might include clients that don’t use Okta as their main workforce identity supplier.
BofA cited Okta’s commitment to embrace third-party identity suppliers as a particularly noteworthy step.
That implies companies might use Okta’s AI-agent security products, even if another vendor is managing their core employee identification infrastructure.
That approach is extremely essential to Microsoft.
Microsoft accounts for nearly 40% of the market, according to the BofA analysis. By supporting third-party identity systems, Okta has the option to target Microsoft-centric clients without forcing them to tear out their current identity infrastructure first.
That alters the possible sales pitch quite a bit. It could offer adjacent AI-agent solutions such as governance, privileged access, and runtime detection, rather than needing to persuade a corporation to migrate its primary identity system to Okta.
Ultimately, BofA believes Okta must be a trusted identity-intelligence layer spanning both Okta clients and enterprises with various identity systems.
The AI-agent possibilities may also go beyond workers. BofA cited Okta’s Auth0 business as another area where it may develop as companies continue to roll out customer-facing AI agents.
Okta’s AI Identity for Commerce is designed to help companies manage transactions originating from proprietary and third-party AI assistants. Simultaneously, the Auth for Universal Commerce Protocol is meant to allow agents to identify items, generate shopping carts, and make transactions on behalf of users, the note said.
That might open up prospective consumers for Okta inside a company beyond the conventional IT and cybersecurity departments to product, engineering, digital, and customer-experience teams.
Related: Jim Cramer has a strong message for AI investors
BofA warned that the customer-facing standards and implementation are still early. Still, customer-facing AI agents remain a possible key future driver of corporate AI adoption, the analyst said.
The big picture is that every task you add to autonomous software is another identity that organizations may someday have to manage and safeguard.
If it occurs at scale, identity security might become a key infrastructure layer underpinning the agentic AI economy.
Okta’s AI economics show promise, but it’s still early
The main reason BofA isn’t becoming more optimistic on Okta right now is simple: The potential is far easier to perceive than the financial effect.
Okta for AI agents is charged per user today, since most installations are still tied to employee processes, BofA said. As agents grow more autonomous, the analyst anticipates that the price structure will change.
Early deals do provide some glimpse into what AI agents may potentially imply for Okta clients, however. In a few of large agreements addressed at Oktane, BofA noted, expenditure on AI-agent products totaled about $1 to $1.50 for every roughly $4 spent on the larger Okta platform.
That kind of incremental expenditure might be meaningful, but BofA noted specifically that the sample is too small to derive a normalized attach rate.
So the bank’s broader takeaway from Oktane is somewhat positive but cautious.
Free Agent SSO may assist in driving adoption. The proposed monetization layer with lifecycle governance and runtime enforcement is Okta for AI Agents. Okta’s other solutions would then be able to capture the associated demand for governance, privileged access, and threat response.
That framework supported BofA raising its price target to $220 from $200, although it wasn’t enough to prompt it to elevate Okta from Neutral.
The moves do not have a major impact on BofA’s near-term financial forecast, it added. The AI-agent potential will be obvious in fiscal 2028 rather than now, the analyst said.
That’s what distinguishes Okta’s AI narrative from just adding another generative-AI capability to an existing software product.
The company’s bigger bet is that autonomous agents will be yet another type of digital identity that companies need to identify, validate, regulate, and maybe even shut down. This could mean a much larger number of identities that enterprises will have to safeguard.
BofA sees enough upside in that scenario to value Okta more highly. For now, however, it still needs more proof before it can be positive on the stock.
Related: Wall Street panicked over AI. Then came an 8-figure cybersecurity twist
