In case you’ve somehow missed the splashy TV ads starring Idris Elba, ServiceNow would like you to know that “The world works with ServiceNow.” The tagline is catchy, as is all the talk about connecting “every corner of your business.” But could that just be marketing? Not as far as I can tell. After attending the company’s annual Knowledge conference in Las Vegas last week — and after years of studying the company — I believe it’s not just hype, because the details released at the show back up its AI ambitions.

ServiceNow is building on its background in IT service management and more generally in workflow automation to attempt something potentially huge in the world of enterprise software. I’ll get to some important caveats later, but if CEO Bill McDermott and the rest of the ServiceNow crew pull it off, it could mark an important shift in the world of tech.

I’ve attended hundreds of conference keynotes, and I thought that McDermott and his lieutenants did a spot-on job of describing — and demoing — what I’m calling “enterprise AI nirvana.” This is an AI platform that supports any AI model, any agent, any type of data, any workflow, any cloud and any system for any industry. The platform already has connectors to every enterprise application and data source you count on, and it’s ready to give you a “single pane of glass” view into your operations so you can manage all your workflows, very much including agentic AI workflows, regardless of which parts of your business they cut across.

This is consistent with what McDermott has been talking about for years: getting rid of the “swivel-chairing” that has your hard-working enterprise employees consulting something like 17 different business apps per day to get their jobs done. For me, the difference between this year’s Knowledge conference and last year’s event was that the presenters increased what I would call the “reality bits.” They put more meat on the bone and gave real-world examples with real customers. Even better, they took the radical step, in the second slide of the entire keynote, of explaining how this will save you money.

I joke that this is “radical” because so many times at industry events in the past couple of years, executives from whichever company have talked breathlessly about saving the world with AI, transforming everything with AI and so on. Sometimes the hyperbole is enough to set off a gag reflex. What CEOs really want to hear from an enterprise SaaS vendor is how the offering saves them money or makes them money. At some level, it really is that simple.

Mind you, McDermott is every bit a visionary — plus maybe the most energetic CEO I’ve ever seen onstage — who does not hesitate to embrace the spotlight to say things like this:

He’s also not shy about citing predictions that AI will generate $22 trillion in economic impact by 2030, including $4 trillion in opex reduction. He jovially despises the fragmentation of data and the siloing of functions that hamper productivity and create what he — rightly — calls “a $10 trillion tax” on the U.S. business economy.

To give McDermott his due, he’s earned the right to talk like this. Not many enterprise SaaS vendors are achieving double-digit growth across the whole company. Some of them hit that growth mark in certain pockets, for example Oracle ERP or SAP’s cloud-based ERP. But even important and very well-managed companies like these or Salesforce or IBM have single-digit growth as a whole. Yet ServiceNow has been racking up annual revenue growth north of 20% for years on end. From all available evidence, the company has cultivated a communal thirst among its employees to win by making life super easy for customers — and it has a track record of execution to back it up. So I would not bet against ServiceNow.

So what exactly is ServiceNow doing with its AI Platform? The picture above shows the enterprise AI nirvana I described earlier. In short, the platform is built to use AI — especially by orchestrating AI agents — to speed up every part of a business by managing workflows much better. In fact, speed was a major emphasis throughout, including during an analyst Q&A session when I asked McDermott about initial reactions to the platform from big enterprise customers. The CEO said that the customers he talked to came back again and again to the level of knowledge the platform conveys about company operations and the speed at which it operates.

The keynote presentations offered plenty of real-world examples of this. While I’m impressed that the new platform has answered so many questions about data ingestion, agentic orchestration and so on, in business terms it’s more compelling to know, for instance, that it has already saved Siemens 1 million work hours.

The big pharmaceutical company AstraZeneca served as another key example. ServiceNow helps the company speed up several parts of its business, such as reducing the time for a supply request from 20 or 30 minutes to a matter of seconds. That sounds even better when you realize that AstraZeneca has 60,000 of those supply requests annually. Combined with other time savings, ServiceNow says it returns 90,000 work hours to AstraZeneca every year.

During the keynote, there was a demo of the ServiceNow platform solving a problem for AstraZeneca. In the simulation, an AI-driven monitoring device in a lab detects an error. Based on a photo of the device producing the error, the AI agent determines it’s broken, confirms the warranty and purchases a replacement. By the way, other companies should study how ServiceNow did its demos during this conference — genuinely outstanding.

Getting to the nuts and bolts of the technology announcements, three big advances stand out to me:

All the components shown in the diagram above enable the kind of autonomous operation demonstrated in the AstraZenica example. This could improve and accelerate the handling of myriad different issues in potentially every area of physical and digital operations, as well as in marketing, sales, customer support and more.

CRM is a particular target, and the fact that ServiceNow is aiming at that target is another indicator of the breadth of its ambitions. During his presentation, COO and chief product officer Amit Zavery said, “Traditional CRM is broken. It is a patchwork of solutions held together with duct tape and chewing gum.”

These are not the words of a company that has come to make friends with the incumbent players in CRM like Salesforce.

My colleague Melody Brue, who has covered the CRM market for years, has written a detailed analysis of how ServiceNow is approaching this market and what it hopes to achieve. In sum, the company wants to build on the years it has spent developing its customer service management platform, uplevel it with agentic AI and win business with a workflow-centered approach. I wouldn’t count out Salesforce, the market leader by leaps and bounds, and the other smart players in CRM, but in terms of the potential for market disruption? Pass me the popcorn.

It’s a running joke that no B2B tech conference these days is complete without an appearance by Nvidia CEO Jensen Huang. In this case, however, I’ve watched Huang and McDermott play off of each other for years, including well away from the keynote stage, and it seems clear that they have a genuine friendship and a special business relationship. It makes sense: they’re both razor-sharp leaders who are endlessly eloquent in putting forward new visions for their industries that challenge the status quo.

At the conference, Huang voiced his opinion that “ServiceNow is destined to be the best platform — the operating system of AI agents.” (He has a knack for choosing particularly apt superlatives for each of Nvidia’s big partners.)

The big announcement for the two CEOs was the launch of the Apriel Nemotron 15B LLM, which was created using Nvidia tools and ServiceNow domain data. This reasoning model is designed to power AI agents that can respond faster and handle complex workflows — right in line with main themes from the conference. While the ServiceNow AI Platform is built to handle any LLM you throw at it, I have to think that this model will make a particular difference with ServiceNow’s agents.

As he was laying out ServiceNow’s grand strategy during his keynote, McDermott asserted that “Only we can do this,” but I’m sure that Microsoft, Google and AWS would take issue with him about that. There’s an important distinction to be drawn here between east-west and north-south workflows; “east-west” refers to agentic workflows that cross different systems (HRM, HCM, ERP, CRM), while “north-south” refers to workflows that live in just one of these systems. All of the big CSPs are going after east-west workflows, and Microsoft already has a stake in north-south via Dynamics 365 to go along with its east-west capabilities. Many CIOs I talk with are DIY-ing east-west in the spirit of reducing cloud costs, keeping control and spanning data on-prem and in the cloud.

Even beyond the big cloud service providers and their rapidly growing AI businesses, the incumbent players in CRM (and ERP, HCM, etc.) are not just sitting on their hands. The big question is, who are enterprises going to trust with their east-west workflows: Companies like CSPs that do this already? Or a north-south ERP-SCM-HCM vendor that wants to expand into east-west? If ServiceNow is able to achieve even part of its ambitions, some of those incumbent vendors are in trouble. Because for ServiceNow’s customers to afford this technology, those enterprises are going to have to dial down some of the other software expenditures they’re making. McDermott has said a few times that his customers want their SaaS vendors to fail so they can pay ServiceNow more.

It’s clear that ServiceNow is going for the brass ring. There’s no doubt in my mind that they have the funding, the growth and the record of execution to do it. But make no mistake, getting there is going to take a lot of hard work. Sure, if the platform can do everything that McDermott and Zavery say, you can go workflow by workflow across an enterprise and potentially improve everything sooner or later. Here’s the thing, though: when you connect that many workflows, it can easily turn into spaghetti, so to speak — hard to keep in line. Plus, you’re probably going to have to change the way you do data integration. In very many cases, this means you’re going to need a global system integrator to help you get there if you want to boil the ocean or set up a new data lake in RaptorDB. At that point — even if ServiceNow delivers beautifully on every promise — there will be an aspect of this that’s like other IT transformation projects. It might turn out to be better than the norm, but it’s still not going to be magic.

What the company needs now to convince prospects that this is possible is a steady supply of customer stories, whether we’re talking about the examples mentioned onstage at Knowledge, one-page summaries, 10-page case studies, video testimonials, TCO calculators or whatever else works to get the word out. The C-suite should already love this because of the promised simplification and cost savings. But as the go-to-market motions reach into the operating levels of enterprise IT, not everyone will be so easily sold. There are entrenched ways of implementing IT — including with preferred incumbent vendors — and there are entrenched ways of doing business, all of which ServiceNow would be happy to revolutionize. But plenty of people in the north-south trenches will need to be convinced, and the way to do that is with real-world examples of ServiceNow’s wares working as advertised. If Idris Elba’s gruff, witty assertions in the TV ads translate into a thousand real-world stories like AstraZeneca’s, the sky’s the limit.