Thanks, everybody, for joining us. This is the first session here on the 2026 RBC TIMT Conference, so thanks, everyone, for joining us. We have a nice chilly room here, which I appreciate because I sit up here all day long and it gets a little toasty. Hopefully you guys keep you guys awake. For the first session here, we have got Josh, I think this is the second time.
Yeah.
This is the second time we've done this?
Yes.
Josh is back.
It is.
Josh Kahn, SVP GM Core Business Workflows. Thank you for being with us, for a repeat customer here. Josh, I think I started out with a similar question last year, but not everybody goes to Knowledge every year and sees you on stage. As group leader, you oversee aspects, HR, finance, legal, just kind of back office functionality.
Yep.
Could you just kind of give us a better understanding of your role and start with your role and kind of what you oversee today, and we'll dig deeper.
Yeah. The role at ServiceNow is what we call a workflow leader. Think of it as a product division leader. We have a common platform and set of capabilities for our core workflows, AI, all of that. Our job, my job, is to take all of that and build products for a specific set of use cases and personas. For example, in my case, as you said, I'm really focused on what you'd think of as the back office. We don't want to call it that because it sounds a little, it sounds like a place you don't want to be. They really are the core of a lot of business functions. My job is to build products that help people optimize the functions in finance, procurement, HR, legal facilities.
Yeah. If I remember right, I was looking back, I think you launched this Core Business Workflow, was it last year or-?
Last year.
Last year at Knowledge.
Yeah.
Since you sort of unified this under sort of your platform now, what are some of the most interesting learnings that you've understood from your customer base through this process?
Yeah. The first is there's a big focus around employee productivity, employee engagement, that is driven largely today by acquisition of Moveworks and what we now call EmployeeWorks, which is the combination of EmployeeWorks and some of our HR capabilities. It's really interesting because we have found times where we go into a customer and they'll have two different initiatives around improving the employee experience, one in IT and one in HR. Because we speak to both constituency and both leaders, we've actually ended up bringing those together. What does that mean? What are they looking for? Day-to-day work as an employee when you need some help from a function could be really hard to get. "Hey, my payroll doesn't seem right. What's the expense policy around taking clients to dinner?" These kinds of questions. "Hey, I moved.
Update my mailing address." Companies want to just make that easier, make it a frictionless experience. That is where the EmployeeWorks product is doing an amazing job of that, giving not just the information people are looking for, but helping them drive action. If you go into a frontier model and you say, "How do I update my address in Workday?" It will give you almost like the docs, the step-by-step instructions, like go to Workday and find this and click there and click that. With our capabilities, because we put the conversational part together with the action part, it says, "Oh, what's your new address?" Click, done. You do not have to go hunt and peck, you do not have to find out where the front door to that other, all the systems are.
That is a big focus for us and one of the big findings. That product in the first three months has already reached $100 million. Customers are up 60% year-over-year. So phenomenal and exciting area. The other thing we are seeing is there is a real desire to optimize the back office with agentic AI, in particular, and agentic workflows. That is where we have something we created called the Core Business Suite.
Yep.
We created Core Business Workflows last year, but we had lots of different product lines that we put together to make that. The reason was we were making it harder on our customers than it should have been because of the way we were organized. Our customers were saying, "Hey, I just want to automate the back office, all these functions. I do not know which one's first, but I am going to get to all of them." We would make them buy 10 different SKUs in order to do that. We said, "This is crazy. Let us just give you one SKU where you can automate the back office." We are going to sum the parts and charge you for the whole thing.
Yeah.
We are going to assume when you start running AI, you will get an ROI and you will need to come back for more consumption. That Core Business Suite is another big part of our growth that, to me, gives a good indication of how people are automating the back office. Core Business Suite was in 15 of our top 20 deals in Q3, with 24 deals over $1 million. That kind of gives you a little bit of the front door and the back.
The back side. I think, one thing we've all understood is, when we think about AI and agentic use cases within IT, we see them every day, and it's very tangible for a lot of people. But it's the back office piece that, I guess I'm wondering, when we talk to customers about that ROI, do you sense a little bit more friction? It's just not as like, "Hey, I'm just a coder and I need to use a frontier model.
Yeah. Clearly, coding is where AI is most mature.
Yeah.
And then I think as you start getting into different use cases, it's less mature, and that's where we're working a lot with them. We have use cases that are in trials with customers, but they're a little bit behind things like our zero-touch service desk for ITSM. So, what I'm seeing in those areas is that, because these are typically thought of as cost centers that provide a business function, there's a set of very clear KPIs that those business leaders are operating towards. They have a cost envelope, and they have three to 10 metrics. So if you're in accounts receivable, you're thinking a lot about how much it costs you to collect $1,000 worth of invoices, how much of your collectibles roll over month-to-month.
And so, what we're working with them on is saying, "Hey, here are the places that agentic can really help you improve those KPIs and bring that cost down.
Yeah.
And so, I think the use cases are clear. We're starting to see those kinds of results, but we're not yet at the scale that you would be in some of the more obvious use cases.
And yeah, so when there is that aha or that unlock moment, the non-coding side of it, what's the trigger for that, where an organization is like, "I get it." Because I know there's and we'll get into some of the questions on some of your FDEs that you're deploying now. But maybe just talk about, what is that spark that an organization sees that triggers that?
Today, it is when they see it working. I think a lot of companies, like early last year, they were in, it was tokenmaxxing. We had leaderboards like, "Use AI, use AI." All of a sudden, everybody went, "Holy crap, we just got the bill. Stop using AI, stop using AI." What they really mean is use AI for things that matter. We build our autonomous workers. For example, in my group, we have an autonomous worker to process invoices. It will take an invoice, it will read it will pull all the data out, it will match it to the purchase orders or the purchase requests, sourcing requests, and it will then look for any issues or exceptions. It has tools to automatically clear those, so if it has an exception, it does not have to go to a person.
It can work on clearing them. Some will fall out to a person, and then it will clear it for payment. That entire autonomous process requires reasoning on behalf of the autonomous worker and the collection of agents that do it. We build that, we test it in our environment. We have some customers who have let us use their data to test it. But really, because there is so much reasoning and so much uniqueness, you want to see it working in five or 10 customers first.
Yeah.
That is where we will take that product out. We will work with a handful of customers. When they see it working and they are like, "Wow, you got 98% straight through processing, and when I do the math on what I am paying you versus what I used to pay to process an invoice, I am going to pay a lot less." They could do that math themselves once they see how this works. Now, the trick from there is turning it into the next 50-.
Yeah.
Which is just marketing the results that people are seeing.
Yeah. We need to kind of put a bow on this. Obviously, you and Bill and Gina were extremely enthusiastic about some of the AI trends taking up the AI ACV target. I think there was a disconnect, though, with how the street interpreted that. From your seat, I think the skepticism on the street is, "Oh, Bill's talking about AI ACV, and yet it's not showing up in the P&L." I know you're not Gina. But can you talk about what these trends that you're seeing that give you confidence that AI is in fact this driver that the team is saying it is?
100%. How people in this room and beyond react to what we do, to you all and to Gina. But what I'll say, what I see in the business and the metrics that I'm excited about, that we're excited about, and I think drives Bill and Gina's enthusiasm, is things like the fact that 50% of our revenue in Q3 was not seat-based. And seat-based revenue is also continuing to grow. So those things are good. Agentic consumption was up 13 times year-over-year. So, we've moved from sort of people using case summarization and knowledge article generation, sort of classic generative AI, to people using our agentic AI at a tremendous clip. We've seen our AI Control Tower, which is a really important part of the strategy, up 6 times year-over-year.
EmployeeWorks, which is a really important part of that overall AI strategy and that front door, with 60% lift in customers year-over-year. I just gave you the data point on Core Business Suite, 15 of the top 20 and 24 deals over $1 million. So there's all these points along the way where we're seeing the AI that we're delivering get used by customers. And to give you a few examples of customer use cases, CVS Health is using EmployeeWorks. So, within 30 days of going live, they saw a 50% reduction in calls to their help desk. That's a tremendous cost savings. Now they're up to 2.5 million conversations, and 75% of people are coming back for a second conversation. So there's some high-level metrics, but then there's customer-specific metrics that back it up. And I think that's the stuff that's getting us excited.
I'll add one more thing, which is, I see the revenue pool for us expanding beyond traditional IT budgets. That's because with autonomous workers, we're starting to tap into things like business process outsourcing revenue streams. Like, why am I paying $3, $4, $5 an invoice to an outsourcer when I can get an AI agent to do it at a lower rate if I'm willing to invest a little in myself and some software? So, that's changing. I think there are places where people will be displaced by technology. That's a different budget pool. Yes, we'll also augment people's work. We'll also do things that people were never doing before.
But all of those things tell me, "Hey, what we're building is getting used, and when it's getting used, we're going to be able to tap into more sources of budget than we ever have before.
I want to double-click. You said something that certainly we picked up on, is the non-seat-based growth that you're seeing. You said that you're not seeing seat compression.
Yeah.
Talk a little bit more about that dynamic, because I know you guys are awfully generous with the consumption analysis credits that you're getting on some of these upgrades. Talk about what are the trends that you're seeing on that.
Yeah.
Because the pushback is like, "Oh, it's just cannibalistic," right? It's just cannibalizing the core.
Yeah.
Cannibalizing the base. Talk about some of these customer buying decisions.
Yeah, I think what I see happening is that people are finding ways to redirect the labor that is being displaced. I will give you an example of that in the space that I am in procurement, is there are some capabilities that we are providing where people used to do them and you do not need people to do them anymore, like in some of the invoice processing work. So, getting a lot more productivity on invoice processing. There is a lot of work that has been going undone before around analyzing spend, analyzing categories, executing on sourcing events. Companies have only been able to actually contract spend for a certain percent of what they spend, because it is time-consuming, it is a big process.
Instead, what they are doing is they are flipping people to start working on more and more sourcing events, actually contract more spend, and negotiate more spend to get a better deal. As we provide more categories of software to support those functions, we can then harness seats in those areas as we need to. That just sort of happens around the horn.
Yeah. I know I asked you this last year. I do not know, my perception listening to Bill and Gina is it feels like we are going much more towards this consumption-based model. Now, I know that a lot of IT buyers still like the idea of buying a seat. There is predictability in that.
Yeah.
Do you imagine a world where we are sitting here in a couple years and, I mean, it seems like that SKU towards non-seat base is just going to continue in the future?
I 100% agree that is the trend line and that is the destination. I do not know how fast we will get there. I do not think it is a revolution where everybody is going to be like, "I just want consumption-based.
Yeah.
Especially coming out of what happened with the Frontier Labs and the bills that people were getting and trying to figure out where to pull together the spend for some of the overages they had. That is a real factor. We are working on pricing models that do the best of both worlds, where you can have what we would call a commit to consume.
Yep.
You can make a commitment around how much you want to buy, and that will be usable across the portfolio. You do not kind of have to worry about, are you spending this much on our customer products and this much on our technology products and this much on our finance products. That can just happen over time, and your level of commitment is going to have an impact on the unit rate. As we see an evolution of consumption-based pricing, it starts to look more like models like that. We have some customers that are expressing interest in things like that. But, to date, it is a sort of a gradual evolution. Yeah.
I guess on that point, because I know Gina talks about some of the uplift to Prime now.
Yeah.
You are seeing ACV uplift as these customers start to think more about flex-based. A lot of people call it flex, but-.
Yeah
That is effectively what it is, right?
Yeah. So, it is a couple things. So, our new pricing model is, number one, it is designed to make sure AI is in everything. Because adding it on, in the beginning, we started with you have your core offerings and then you add AI. And now, with Foundation, Advanced, and Prime, AI is in all of those SKUs. We also put some of the core capabilities in there that help people along the way. So, the SI Control Tower is in there with a limited amount of usage. Our Workflow Data Fabric to help people get all the right data and context into their use cases is in there in a limited amount of usage. And the three tiers are set up to sort of meet customers where they are. So, Foundation is about helping people do work. Advanced is about adding AI agents to autonomously do tasks.
Prime is about autonomous workers.
Yeah.
Fully doing jobs autonomously. Customers can move up that curve. Also along the way, there is more and more of the AI capability.
Yeah.
Essentially, our tokens we call assists in them as well. Look, you could be on the Foundation and use a ton of it and just buy more assists, or you could keep moving up the curve.
Yeah.
Get more capabilities and more assist.
Yeah.
It is a really adaptive model to help customers make sure they are taking advantage of AI in the way they want to take advantage of it. When they get a positive ROI, then they are going to buy more to fuel the scaling of that ROI.
What kind of additional Now Assist credits are you seeing people come back after they initially move to, say, an Advanced or a Prime, and come back and they are reloading their assists at that point?
Yeah. This is where, I know you mentioned FDE.
Yeah.
This is one of the places where we really see a role of FDE in some of our most important customers. We use FDE for a couple things. One is co-innovation, and the other is accelerating consumption.
Yeah.
We saw one customer that we were able to get live on their use case in 12 weeks, and now they're burning an equivalent of $1.5 million a year in assists. If we-
That's incremental. That's beyond-
Yeah.
Yeah.
Yeah. If we left that to sort of the standard process, they probably would've gotten there, but it might've taken a little longer. We're seeing lots of opportunities where, with really important customers who really want to go after a use case, we can deploy FDEs and AI architects to help them get to value faster. We're also doing it to build new use cases. This is a place where you'd say, "Hey, if we didn't do that with them, they wouldn't be using us to solve that problem." One example is a logistics provider who built a like a shipping investigator. When something started to go wrong, it was an agentic way to look into it and figure out what it was. They're saving 100,000 hours a year, and every time they're doing that.
They see 100,000 hours a year, they're like, "Great business case." We see them using our agents to create that business case. That's good monetization for us. Another example in my business is an oil and gas company that was doing supplier onboarding, and we sent our FDEs in to do a use case. There was a very specific part of that supplier onboarding, so they're saving about $100,000 a year on that one use case. It's one use case that was quick to deliver and now opens the door to building the next use case and the next use case.
You guys aren't alone. There's a lot of software companies, enterprise software companies using this FDE model even more so. I don't know, it feels like, are we at this point of the AI adoption curve where it's not as turnkey as maybe coding use cases, where it requires a little bit more handholding? I guess that's question one. Question two, do you envision if we're sitting here a year from now or two years from now, that this is just, there's an agent that helps customers do this and that you don't have to deploy these actual humans to handle that FDE role?
Yeah, I think there are use cases that are extremely mature where you don't need an FDE to go do it. I would say our, what we call zero-touch service desk for IT, HR, we have them in procurement, accounts payable. That's basically an autonomous worker that takes incoming cases and just fully resolves them. Figures out what's the person asking for, what are the paths I have to resolve it, tries one, didn't work, tries the other one. Those kinds of use cases are pretty common today. I mean, things that we don't provide coding agents to the market, but we're using coding agents ourselves. In my group, we were able to bring one of our products in by six months with fewer people. We didn't need consultants to do that.
Yeah.
We didn't need FDEs from anyone. We just did it. I think there are real use cases where it's like, man, people are ready to go. I think where the more emerging use cases, a lot of them in the back office, are places where it helps to have the FDEs. Once we do five of those and we have it in a product and we know the configuration dynamics that you need across multiple different customers, that becomes more like out-of-the-box software.
Yeah.
I think, in some ways, it's similar to the evolution of software over time is people started by building custom things, then someone's like, "Hey, I'm going to package that into an offering." If you can buy it off the shelf, you're always going to be better off. It's going to faster time to value, higher quality because more customers are pounding on the same code and have found any problems you'd hit, and you're going to ride the innovation curve of the vendor. They're going to release improvements all the time. If you're doing it yourself, you have to build those improvements.
I expect to see, and we are seeing, we've created over 300 improvements to the product already through our FDEs, so I expect to see that make its way into software, and we'll be delivering more and more out of the box, where today it's our first and earliest and most common use cases.
Bill's stance on this is that this is sort of the accelerating this adoption curve. That's when Bill talks about and Gina talks about the AI ACV doing what it's doing, there's an optimistic view that these FDEs are going to accelerate that deployment and that value creation.
Yeah. They're accelerating it right now in select customers, and it will accelerate it at scale.
Yeah.
As those things become part of the core product.
Yeah.
Right. The zero-touch service desk is in ITSM Prime. So, that capability now makes its way to the market at scale, and it doesn't need an FDE to be deployed every time.
Control Tower, you can't talk about ServiceNow without talking about that opportunity. From your perspective, you had a lot of time to absorb and understand how customers see ServiceNow's role in the industry. Talk about ServiceNow's right to win that orchestration for AI agents. There's just a lot of people going after that, but uniquely, why are customers turning to you for that?
Yeah. First, we were pretty early on in realizing how important visibility and governance was going to be in AI, and the realization that you're going to need not just to be able to see it, but to act on it, the kill switch concept, which I think a couple months ago everybody became acutely aware of the need for. It's not just that on AI. It's the fact that we have the underpinnings that make that all possible. For years, we've been shipping and are the industry leader in providing a configuration management database, CMDB, as a part of our TX offering. That CMDB tracks every asset in the enterprise, every identity in the enterprise.
All of that information is what we've used for years to power things like our Security Vulnerability Response, which, when a new vulnerability comes out, we know every server and device in the enterprise, and we can automatically then patch them with the workflow capabilities. That combination of CMDB and workflow that we've had for 20 years and have an industry-leading position on is the core of what now the AI Control Tower, the SI Control Tower is really driving for customers. That's one big part of it is like, hey, 20 years of foundation that went into making this possible. The second is we're treating it as a heterogeneous play. I think there's other places where you look and what they're really doing is managing only their agents, whereas we can manage our agents, we can manage third-party agents.
We have one customer that's already managing 2,000 AI use cases with the SI Control Tower. Accenture, with 800,000 employees, is using the Control Tower to manage all their AI use cases from ServiceNow.
Yeah. Talk about then, I guess, one of the other buzzwords in enterprise software is headless.
Yeah.
And sort of abstracting the UI and just interacting with the database from an API. Talk about ServiceNow's stance on headless as it relates to the rest of the enterprise.
Yep. Headless is an important part of our strategy. It's an important part. It's not the cornerstone. We have EmployeeWorks, which is our front door, and we want every one of our customers using EmployeeWorks. Not every one of our customers is going to use EmployeeWorks. Some of them will use EmployeeWorks and other things. What headless does is it lets us build best-in-class experience when they're using the ServiceNow stack and still create a great experience if they're not. I'll give you an example of this is in procurement. One of the big challenges companies have in indirect procurement is when employees need something, they have no idea where to go. The form asks them a bunch of questions they don't know how to fill out. It's 400 questions long. It's a pain.
We created, together with EmployeeWorks and built into our new experience, Luxee, a way that employees just go in and have a conversation about what they're looking for. "Hey, we're opening a new office. I need 20 new desk chairs and 20 desks and 20 docking stations and 20 monitors." It'll say, "Oh, okay. Here's how we're going to handle that." It'll guide them through the process. It's visual. It's not just a form. It's pictures. It's guidance. It's automation. All of that is within the experience. We do that with EmployeeWorks. If you were using Claude, we could do the same thing. Not quite the same experience because we have this ability to generate graphics in EmployeeWorks, and still working on some of that over there.
But the point is, if you come to us and you say, "Hey, I am going to use Claude," we can still sell you our Source-to-Pay Operations product, and you are going to get a tremendous amount of front door AI value out of it. And so, in that way, we believe the best way for us to get the maximum surface area and the maximum amount of consumption, because every one of those conversations is consumption for Source-to-Pay, is to support multiple front doors. Now that said, as I mentioned, it is not our core strategy is to be headless, but it is an important part that we see opening the door to more revenue and more consumption.
You talk to customers all day long, and if you sit back and you listen to what they are saying, if you were to say, "I think the industry is getting this wrong about AI," or, "I think you guys are all wrong about this," and maybe where you have the highest conviction, especially if we are sitting here three years from now.
Yeah. So I think what people are getting wrong is that AI is going to create more I think it is going to create more jobs and more economic value. And they are going to be different, but there is going to be more. So when I look at where we deploy AI in the back office and Core Business Workflows, I look for three different use cases. One is work that is beneath the human economics today. One example is a ton of invoices go unpaid, but it is not worth paying a person to chase them down. If you give it to a collections agency, you get pennies on the dollar back. AI agents can actually go do that work that you were not doing at all before. No negative consequences of that in terms of how your workforce feels about it.
The second is, yes, there are going to be some jobs that need to change, because an AI agent will be able to do that job. And then the third is places that AI agents can make a human more powerful, better, more productive. And so if you look across that spectrum, there is really only one of those that impacts today's workforce, and there is a whole bunch of stuff that has to happen to make those work in order to create the value out of them. So I think people are getting wrong that AI is just about replacing humans, and I think they are missing the part where there is tons of stuff that is not happening today. And there is tons of stuff that humans do that they could be so much better if it was just served to them.
Yeah.
If you look back in a few years, I hope what people will see is Core Business Workflows, we are making a tremendous amount of progress across all of those. Companies are getting better ROI, performing better on their KPIs, spending less money because they are using our software to power the back office.
Spending less money overall, but more with you guys.
Exactly. Absolutely. 100%.
Yeah.
Like I said, there are multiple pools of dollars here that we can tap into because some of that, take the collections use case. That might come, sorry, at the expense of some of the collections agencies.
Yeah.
Accounts payable invoice processing, it might be at the expense of some of the BPO providers. Everybody is going to have to adapt to this new market, and we're not restricted to software TAM and IT TAM. We can get into a whole lot of other categories now.
Your conviction on the long-term targets versus a year ago, higher, with what you've seen from an enterprise?
100%. Gina gave you forward-looking guidance. I think the fact that at the beginning of the year, we said we would do $1 billion in AI ACV, and now we're coming in out of Q3 saying we're going to do $1.7 billion, tells you we're seeing a lot of really good trends in that direction. And I think we can continue to do the things we're doing today.
Great. Well, we are out of time. Josh, from all of us at RBC, thanks, and we'll look forward to next year.
Yeah. Thank you. Appreciate it.