ServiceNow CEO Bill McDermott: Agentic AI puts IT back in control

LAS VEGAS — Over the last fifteen years of enterprise software, a sizable-number-but-maybe-not-a-thousand flowers bloomed as the growth of software-as-a-service allowed startups to build focused applications for just about every business function under the sun, which emboldened small teams to pick the tools they thought were best for their jobs. If you believe ServiceNow CEO Bill McDermott, agentic AI will bring an abrupt end to that era.

"Right now we're in an agentic AI moment where the industrial software complex of the 20th century is being integrated, and in some cases, drastically reduced in numbers of applications and moved to the ServiceNow platform," McDermott said in an interview with Runtime this week at Knowledge 2025, ServiceNow's annual conference for customers and partners. Around 25,000 people attended the conference this week, up 25% from last year.

As worries about the economy accelerate, CIOs are regaining control over the sprawling application footprints they tolerated for several years in hopes of keeping employees happy. Every enterprise software vendor with a pulse is telling them that AI agents will allow them to cut costs while improving productivity, and ServiceNow is determined to be part of that conversation.

"What's happening is technology is the only way out. It's not kind of, sort of; it's the only way out," McDermott said. Over the course of the interview, he discussed ServiceNow's decision to take on Salesforce, why he thinks agents are the remedy for economic worries, and the speed at which he promised companies can get up and running with agentic AI on ServiceNow's platform.

This interview has been edited and condensed for clarity.

The headline this week seems to have been the CRM announcements, and the not-so-thinly veiled strategy of taking on Salesforce, which has been in this market for forever and practically invented the cloud-based version. What gives you the confidence that you can take share there and dislodge them from that position?

Agentic AI has changed everything. The unique attribute of ServiceNow is the AI platform for business transformation. And to transform businesses, you have to serve every function of a company. So if you think about a customer relationship, just look at it mathematically: The net present value of a very satisfied customer is any company's greatest asset. So if you're in a situation where you're going to satisfy a customer, you need engineering involved. You've got to make a great product. You need pre-sales involved. You need sales. You need post-sales. You need finance, you need legal, you need HR. If you have a dissatisfied situation you have to correct, every function of a company has to get involved in that.

The fastest-growing business in ServiceNow is CRM, that's why we declared and put a flag in the ground that we decided to take it all.

If you think about autonomous agentic AI, the control tower for how that runs in a corporation will come out of IT; the technical community, the Chief Digital Information Officer. Alternatively, you would say, okay, well, why doesn't the sales director do it? Because the sales director doesn't know enough about technology to run the agentic AI strategy for a corporation.

We're in a unique situation here where IT is the foundation of ServiceNow and always has been. We've extended the perimeter to include the employee experience, the customer experience, the engineering experience, and all the core business functions of a company now on one platform.

We have the whole middle office covered. We've had the middle office covered now for many years. And then on the service management side, the fastest-growing business in ServiceNow is CRM, that's why we declared and put a flag in the ground that we decided to take it all.

You talk about agentic AI and autonomous agentic AI like it's here. A lot of the companies I talk to don't quite feel that way. They're still struggling to get those apps into production. What do you think is the missing link that gets them over that hump?

We introduced a reasoning model [Tuesday] that is a game changer. You're no longer searching for data, you're no longer going through spreadsheets and PDFs and emails and internet links and all the other social networking techniques. These reasoning models are actually going through very complex decision-making procedures and actually coming up with an autonomous agentic AI decision. 

We've done this now for IT. We've done this now for security. We've done this now for risk, we've done this now for governance, and we're rolling this out across corporations. So we're in a whole new world. And the thing that companies don't know yet, but once they find out they won't know how they ever live without it, is that ServiceNow has it ready to roll today. And that's the difference between the pretenders and the doers. We thought of this six years ago and started engineering it with hundreds of Nvidia and hundreds of ServiceNow engineers in the same room, shoulder to shoulder building this stuff.

For the last 15 years, companies built their enterprise software stacks around the best of breed idea, where people just pick whatever works for your department and go with that. You're talking about IT regaining control, this feels much more of a centralized approach to how teams use different applications within their workflows. What changed there?

Right now we're in an agentic AI moment where the industrial software complex of the 20th century is being integrated — and in some cases, drastically reduced in numbers of applications — and moved to the ServiceNow platform. Over the last almost six decades now, organizations were structured functionally, and they bought the functional system, whether it was finance or HR or sales, etc. On top of that, these global companies have many different locations, [and] in some cases, each location or each regional theater did their own thing. So some are in the cloud. Some are on premises. Some are in hyperscaler clouds. Some are on different release levels. Some never integrated in the first place, where they still communicate with spreadsheets and stuff.

So what's happened now, — and you see this in the public sector, the public sector is saying, I want efficiency, I want effectiveness, I want OpEx out so I can serve the American people better — if you look at companies, they're behaving very much the same way. They're saying, why am I paying for something that isn't delivering me a positive ROI? And if I can collapse the software industrial complex of the 20th century and simply use those functional apps — pick me the best one, get rid of the rest, and I'll use it as a database — and that data then can feed into ServiceNow, and ServiceNow can go end-to-end across every department using autonomous agentic AI to move these companies a lot faster, to get positive ROI and get results.

Lower OpEx, higher ROI, happier customers, happier employees. It's as simple as that.

You're in a unique position to kind of evaluate some of the things that are going on with the economy right now. And in the first quarter, it seems that tech companies, by and large, did pretty well, but from what I understand looking at some of the supply chain issues and some of the things that are expected to hit later this month, folks are starting to get a little more nervous. What are you seeing when you talk to customers about their plans in the next, say, six to nine months?

I see that they are uncertain, as you rightfully point out. But in the midst of the uncertainty, the reason I can tell you for sure that autonomous agentic AI is still number one, two and three on their top priority list is because they have to invest in their technology estate to take cost out.

They have to go someplace to say, "we have to thin out the spend on the things that don't matter so we can make room for the things that do matter." They have to be able to take the operating expense to its leanest form possible, because they have to make revenue assumptions that take into account that there could be less revenue, which means you have to have less expenses with less revenue to hold your margin profile at the same time.

This is the biggest learning at the same time: The CEO is saying, "I can't also take my eye off the ball when it comes to business model innovation and the dream of our future," because if you don't invest in the short run, you're obviously going to be very, very jeopardized in the long run. What's happening is technology is the only way out. It's not kind of, sort of; it's the only way out.

Think about it this way: You have massive shortages of IT professionals in the world economy. There's millions of open jobs in the global economy for IT professionals, it's even worse for AI professionals, only one in four AI jobs are actually going to be filled because there's a lack of skills. There's a huge gap. So they're going to be looking for companies that have autonomous agentic AI capabilities to make up for the shortage of the labor that in different times they would turn to to solve problems. So I think that from our perspective, whether it's an uncertain time or it's a growth time, we're going to do just great, which is why I actually upped the guidance of the company.

Even industries that are hit by the tariffs, for example, which is one of the most uncertain things right now; if you take an automobile, they've got on average 30,000 parts in a car. And if you put the tariffs that are currently structured on the table, that would be on average $10,000 price increase per car.

In the old days, you would have a system of record. You would have business logic within that system of record, and you would begin augmenting the rules in this business logic to account for tier two, tier three suppliers from the most friendly tariff environments to ultimately compose the vehicle. With autonomous agentic AI, I can do it in sub-second speed on the fly, and so the economic value to an auto manufacturer as an example, is in the billions.

We are the fastest time-to-implement-and-execute-adoption of any enterprise software company in the world.

I think that is what people are looking for. What is the big idea in every industry where I can apply the power of autonomous agentic AI and a platform like ServiceNow's?

But people are on that journey right now, right? And they're confronting these changing economic conditions in the short term, which appears to put much more pressure on them to see a return from their investments in the technology than maybe they would if they had a more stable economy.

The faster they go live, the faster the adoption cycle takes place, the more money they derive from the decision, either on a cost takeout or revenue growth. Uniquely, and this is a big thing, we are the fastest time-to-implement-and-execute-adoption of any enterprise software company in the world. So instead of, like, months, we're talking days and weeks.

Do you have a benchmark for that? 

We have plenty of heuristics. But if you take a good-sized company, they should be live on ServiceNow in 60 days. So this CRM solution should be installed, up and running and adopted in 60 days. Some of them can be faster, depending on the complexity of the environment. And what most of them do is they sandbox the data into our [configuration management database], and then they start to retire the various layers of technology that they're overspending on. So what's really cool is they derive the business value from the original case, and then they start exiting applications.

I had a bank very recently retire 256 applications onto the ServiceNow platform, and they saved so much money, they actually presented it at their investor day. I didn't even know it until an investor took a picture of the slide and sent it to me and said, "holy cow."