Good morning, everybody. Welcome to Investor Summit at Octane 26. So we've got about an hour. We're going to take questions in the room, but we're going to be joined by Todd McKinnon.
I mean, we're pretty financially aligned in R&D. So when we groom the portfolio, we're looking at like what are the major drivers in terms of current revenue, what are the bigger bets down the line, and we have no fear of moving resources even if we have to cross the lines between the different platforms. So we're here to win, and that's the only way to win.
Yeah. One thing I want to be clear on, everyone, so we combine OP Zero and Okta R&D, so it all works for Rick now. So he can move pieces around, make sure the investment across the entire portfolio is aligned on the biggest impact things. I think it's been a great thing. It's really helped Okta for AI agents in the gateway and support.
Thanks, Sia Park. Thanks so much for hosting the session. John, maybe for you to change it up a little bit. I'd love to talk about some of the observations that you have from specializing Salesforce, right? It seems like it's been a great success, basic opportunity which we all clearly excited about. Do you still feel like there's low-hanging fruit for better execution just in the core workforce and OP Zero businesses? Does that make sense?
Indeed. I mean, I think the momentum, the durability of our whole business is really strong. It's a really big part of the success we've had this half. A large part of that is the improvements and the accelerated benefits of sales productivity. For 18 months, we've had a very stable operating model in go-to-market. We've benefited from specialization, like I said, in several forms. We specialized around our public sector opportunity. We specialized in commercial and down segment around like specific customer need, around new logo acquisition and install base and account management, and like I said, in platform specialization up above that for enterprise and strategic. It's super clear that when we think about specialization, and obviously many companies at our pace and our stage with this, you know, broad range of product capability and demand that solve, you know, specific problems for customers and our stakeholders that we manage. There are various choices around how you specialize, and we take our north star from what is it that the customer needs? What is it that the customer needs to be successful in the thought leadership, relationship, capital, technical expertise and partner ecosystems so that they can invest in the Okta and OP Zero platforms and think about really as generational strategic bets over the long term, and that's how we think about specialization. And so if you look at all of the incredible innovation that we've built in around the Okta and OP Zero platforms over years, like they're very broad platforms. They have a lot of use cases covered. They require a lot of specificity, a lot of understanding about the competitive landscape, the buying processes of stakeholders when they think about those platforms, different identity use cases change. So it does require learning, and our tenure in sales is very high. Like I said, productivity is very strong and our attrition is low, and so we're getting the benefit from that. And so when we think about the future, we feel about, you know, shifts in those areas carefully and prudently, and we really think right now we're on the right path in the structure and the team, the capacity we've got and how it's yielding in pipeline and sales productivity.
So, Todd, on the keynote, you pretty much highlighted why you made the Okta assets so available to everyone, and clearly, as we just highlighted, it's the frictionless on-ramp for customers to adopt more. I just wanted to connect the dots here a little bit since we already have several thousand customers who are already using Okta for AI agents, and you mentioned in the keynote that OP Zero is a connection point, and then if you need more complex use cases, requirements for governance and runtime enforcement, just in terms of the monetization journey for the customers on Okta for AI agents, how is that group progressing?
Yeah, so I was meeting with a large, well-known insurance company, and they're an Okta customer. They use Okta for customer identity and workforce identity, and the conversation was all around cross-app access, and they wanted to, they're very excited about cloud supporting it. They're very excited about a bunch of their applications supporting it, and so they want to understand that. So they were super excited about just being able to do that in their environment. So they have agent SSO, which means that they can start using the Okta directory of agents. They can start doing cross-app access, and they're very excited about that. And it was like two meetings in one week when they realized that they needed full lifecycle. They needed support for connections that weren't cross-app access. And you know, it's a full upgrade to Okta for AI agents. So that was kind of like, I was really excited about that because that's our strategy. It's like kind of seed the opportunity with cross-app access, deliver value with that, but also deliver value with and monetize the value with Okta for AI agents. Which is why when you put it all together too, you kind of see why third-party IDP support is important. You can, you know, that even if it's a company that doesn't use, like this large auto manufacturer I was talking about, they don't use Okta for their IDP, but we could run the same play there, cross-app access, upgrade to Okta for AI agents. So yeah, it's a, we're like I said, we're in a good position, still early, we have a lot of work to do, but the strategy and the tactics, we're seeing early evidence they're playing out quite nicely.
Great. Steve Paney with McQuarrie. Thanks for doing the Q&A. Yes, sir. I'm wondering, question probably for you or anyone else that wants to comment. Given kind of the early state of agentic identity and agentic governance, and you're one of the few vendors with a GA product out there, how is this impacting customers' efforts to adopt agentic AI more broadly? And where are you seeing, like if you were to think about the use cases that are driving customers to consider Okta for AI agents, what sorts of agentic use cases are priority for them, and where are they in some cases holding back on adopting agentic technology because of this early state of the control technology?
Yeah, I think it's a really good question. I don't think we've, any of our customers for Okta for AI agents, I don't think we've introduced them in the customs of agents. I think these are companies that have been kind of on the path. You look at Ramp or Yahoo or Dell, or internally our own example, it's, you know, these companies see the potential and are on the journey. Either they're heavily invested in platforms that are offering the agentic capabilities or they're building their own, and so the need kind of smacks them in the face, and we're there to service that need. I think when you know, if you just think about it like as an adoption curve, these would be called early adopters, I think, but you know, and you know, we're trying to make the pie bigger by codifying and clarifying how things fit together with the blueprint and trying to really hasten the big part of this bell curve of adoption. Because I mean, it's coming, right? I mean, every AI company that's going to be successful is, you know, it's almost like loses meaning because saying they're going to be agentic is like, it's the same thing as saying they're going to adopt the future of technology. So it's going to happen. It's just a matter of how long, and you know, the reality is it's probably going to be longer than we all think. But it's probably going to be bigger than that. Yeah. So it's super motivating.
Thanks for doing this. Mike Cos with NEM. Two questions. First on the AI monetization. Can we talk about where we are in terms of price discovery and what is the market willing to pay in terms of valuing an agent?
Yeah, I think so. This is a really important thing we think about a lot. Right now our pricing for Okta for AI agents is per user, which sounds weird, like it's for agents, why would it be per user? And that's, you know, if you got a chance to check out the keynote today, we're a pretty pragmatic company. We want to meet customers where they are. And customers, and this is like kind of like the previous answer where these companies are adopting AI agent solutions, but they're really helping a user and making a user more productive, and the flows all kind of start for a user, whether it's the Dell example where the AI system is like in the interaction of a customer service or support person is like figuring out based on the AI model the next best action. A lot of the use cases are tied to a user, so it kind of makes sense for someone like Dell or Ramp or Yahoo or many other customers where they say, hey, I want to pay for this per user, and it's like basically an uplift of the per user price. And we're seeing like, you know, if the whole using round numbers obviously, but I can think of deals, sizable deals that have been, if the overall platform or Okta is, you know, is like, you know, four, the AI agent part is one or one and a half, so it's significant part of uplift for the whole suite. And that's exciting, but I think this is going to evolve. This pricing model is going to evolve, and it's going to be, it's not going to be per user. There's going to be something else. It's going to be per agent. That's problematic now because there's not a standard way on how you count agents. By the way, it's something we can add to the blueprint, a systematic way to count them. It sounds simple, but it's going to really benefit customers because right now the problem with counting is that if you have an agent like our internal agent called Dex, is that one agent or is that 7,000 agents because that's how many people use it? And you know, of course, if you're paying by the agent, you want that to be one. If you're, you know, charging by the agent, you want that to be 7,000. So the industry will figure it out. I think a big impetus will be the longer running agentic workflows get, the more autonomous workflows get, the less coupled they will be to a user session, and that'll really force us all to figure this out. But we'll figure it out. And I have no, the value we're adding and the value blueprint's adding. Like I said, I'm not worried about monetizing this. I think this opportunity could be bigger than all of identity in the fullness of time. We'll get there.
Maybe for a quick follow-up, I'd just love to ask John, like go-to-market on this point, like it feels like each touch point by default has to be more bespoke to help each customer on their journey. And I'd just be interested, like what are the additional resources you're putting behind this to help customers?
There's two or three things I'd say on that. We prepared for this year really waiting for the moment, and it gave us an opportunity to reconsider our sales methodology and go-to-market. So the structures that we've got in place to create a framework for how we go into the market with thought leadership around AI, like we completely rebooted our sales methodology at the beginning of this year specifically for that purpose. Secondary to that is we want go-to-market folks to come to Okta and it be the best experience for them, not only in terms of making them identity experts, but also like fusing how they go-to-market with AI tools. And so we spent a lot of time and effort and investment giving them amazing tools when they're showing up with customers that they have the right information that's specific to the use case. It's specific to the industry. It's highly relevant to really the innovation that's coming very quickly out of R&D to really give them the tools to excel, differentiate themselves and meet customers where they are. The third piece on this is you're right, by each customer use case is bespoke. It's highly specific to where they are on their maturity curve of agents, and an incredible part of that is preparing ourselves with thought leaders around those customers who work with them for years on their own technology strategy, and for that we have to rely on our partner ecosystem. So we've spent a lot of time and effort with our partners educating them on this massive opportunity, but really preparing them to bring their specific leadership about the company in question to bear as they engage with us with our expertise around AI security, and that's really playing out. We had an incredible partner summit yesterday. Some of the performance from our consulting partners, the advisory firms and companies are working within enterprise and strategic, it paid massive dividend. We look at when we look at the performance of our big deal cohort, really big part of our growth story in H1. I look at all those deals. I mean, our top 20 deals in Q2, they were like sold side by side with partners, co-selling with us. Our biggest deal in Q2 was partner-sourced. So a massive part of that is being there with our partners providing advice, technical expertise that is relevant to their bespoke agentic use cases.
Hi, it's Greg Mos from Mizuho. So I have a bit of a follow-up on agent SSO. It's been GA for a month, and the demo you showed this morning was pretty slick. What I'm wondering though is how much of an on-ramp for Okta for AI agents do you think that this can be? And from a go-to-market standpoint, are there any strategies or mechanisms that you might be able to put in place to get a healthy percentage of customers to sort of matriculate from agent SSO to Okta for AI agents going forward?
Yeah, I mean, you can almost think of it as a very clear analytical dashboard that we can use for pipeline, and so we know all the customers that are onboarding onto it that can be fed to John and his team can engage. And I mean, honestly, you don't really have to wait for that. It becomes pretty obvious in terms of all the features and functions that you need to manage the agent lifecycle once you're onboarded to it. But it does give them the first litmus test of understanding of like what is the experience as a result of not having to do all these like consent-based prompts through, you know, cloud console or what have you.
It was interesting, and it was interesting internal debate about should we make the cross-app access capability add it to every product of Okta and make it widespread. And it was the interesting debate. The debate against doing that was, oh, we don't want to give it away, we can charge for it, we can monetize it. The argument for doing it was like, you know, obviously the pushing the cross-app access standard, giving an answer to these ISVs, and they ask how many people could do this. The answer is, you know, all of them. That's compelling. But the other point is like, listen, we're pouring all this energy and innovation into Okta for AI agents, and you can't build enough value around the blueprint and all the problems there are to solve that you can't get them an upsell. It's like, really, come on, you got to work harder. And the result was like clear that we do have compelling value in Okta for AI agents way above the cross-app access and the stuff that in the directory entries and the capabilities there. And I think we're going to be pretty happy with that decision.
Yeah, I mean, it's going to take a lot of effort and focus and momentum to get cross-app access adopted broadly in the market and the ecosystem, and a major part of how we're going to be tracking customer success moving forward. I mean, we obviously have a set of customer health metrics that we measure in every kind of standard like CSN motion, and how we think about agent SSO is going to form part of the consistent evaluation of the value that we're adding across all forms of how they're being successful with their stack.
Jim Kim, Loop Capital. Just following up on Mike's question around go-to-market on Okta for AI agents, how heavily are the partners involved in the initial go-to-market? Because obviously we hear a lot of good things about the pipeline from the earnings calls. How are those generated on that side? And also just walking the floor for a few minutes before I came in here, did pick up signs that maybe there's a greater services component related to Okta for AI agents, maybe almost two to three times versus your single sign-on? Does that change the way that you do go-to-market and you got to tie into services to make sure that you're engaging the prospects properly?
