ServiceNow is ‘Not’ a Boring Company

There is no doubt that everyone wants to work with ServiceNow. The company sits somewhere between OpenAI and TCS—a unique position. It neither builds models nor delivers them to the last mile with a lakh-strong workforce. That is exactly why it has stood the test of time without dominating headlines every week.

Every few weeks, another benchmark drops. OpenAI launches a new reasoning model, Anthropic improves Claude Code, Google catches up with Gemini. Yet, Sumeet Mathur, Senior Vice President & Managing Director of ServiceNow India Technology and Business Centre, believes everyone is looking at the wrong scoreboard, even though ServiceNow works with all of these companies. 

“Intelligence has outpaced execution,” he tells AIM. “The models can think. They can’t act.”

That single line explains why ServiceNow’s AI strategy looks very different from much of Silicon Valley’s. Instead of competing to build the next frontier model, the company is betting that enterprise AI will ultimately be won by those who can make intelligence useful inside organisations.

That may sound similar to what Accenture, Infosys, TCS, or Wipro do—but it’s not. ServiceNow works with those companies as well.

“The future value in technology is probably going to be much less on the intelligence and model part,” Mathur says. “It is going to be much more on the orchestration part. The technology is moving from intelligence to orchestration of intelligence.”

The AI Control Tower

That thesis couldn’t have come at a more interesting time.

Last week, ServiceNow reported another strong quarter, beating analyst expectations on revenue and earnings while raising its full-year guidance. Investors responded positively, sending the stock higher after the results.

CEO Bill McDermott used the opportunity to argue that AI isn’t reducing the need for enterprise software. It is increasing it.

“If you look at the terms of our contracts, they’ve actually gotten longer,” McDermott said after the results. “There’s going to be more AI. There’s going to be more incidents, and all these things drive more and more volume to ServiceNow.”

His comments came just a day after OpenAI disclosed that one of its advanced AI agents had escaped a controlled cybersecurity evaluation and compromised the infrastructure of AI startup Hugging Face before being detected.

McDermott immediately contrasted ServiceNow’s approach.

“We have a kill switch that stops AI agents that go rogue,” he said, referring to the company’s AI Control Tower, which gives enterprises a central platform to monitor, govern and secure AI agents.

For ServiceNow, the next enterprise AI battle is not about who has the smartest model. It’s about who can keep thousands of AI agents under control. And that, the company believes, is not something even the Indian IT firms are prepared to do. 

‘People Think We’re a Boring Company’

ServiceNow doesn’t launch consumer chatbots or text-to-video models. It doesn’t dominate social media conversations every time a new model arrives. Yet, behind the scenes, the company says nearly 90% of the Fortune 500 use ServiceNow in some capacity, running close to 100 billion workflows every year.

Mathur is well aware of the perception problem.

During the conversation, AIM joked that ServiceNow often comes across as a “boring company” when compared to AI startups that dominate headlines with every new release.

Mathur disagreed almost instantly.

“Oh, hell no,” he laughed.

The perception, he believes, comes from the fact that the company rarely builds products consumers directly interact with. “The buzz is mostly around models right now, not outcomes.”

That distinction matters.

OpenAI launches a new reasoning model and the internet talks about it for weeks. Enterprises, however, have a different problem.

Customer data sits in one system. Finance operates on another. HR uses something else. IT has its own stack, and procurement has another.

Making all of those systems work together is a far less glamorous challenge than releasing a chatbot. But it is also the problem enterprises pay billions to solve.

“We believe we are the operating system for enterprises,” Mathur says.

Founded more than two decades ago, ServiceNow originally became known for IT service management. According to Mathur, that became both the company’s biggest success and its biggest misconception.

“We became the victim of our own success.”

IT service management was simply the first workflow the platform solved. Today, the company connects HR, customer service, finance, procurement, security and dozens of enterprise workflows on the same platform.

That, Mathur argues, is what makes AI actually useful.

AI Can Think. It Still Can’t Work

Inside ServiceNow, the reaction to AI was much calmer than elsewhere. It was almost as if the company had been preparing for it.

“When ChatGPT came about, it wasn’t a revolutionary moment for us,” Mathur says. “It was the next logical step.”

The company had already spent years embedding AI into enterprise software after acquiring Element AI and integrating machine learning across multiple products.

Then suddenly, every board wanted AI. Getting those pilots into production, however, has been a very different story.

“It’s very easy to build proof-of-concepts,” Mathur said. “It’s much harder to get to production.” The reason, he argues, has little to do with model intelligence.

“You start with data. You contextualise that data into a business process. Then you feed it into a reasoning engine. Once you get the intelligence, what do you do with it?”

That question, according to him, is where most enterprise AI projects fail.

Where is Intelligence Headed?

Mathur points to a recent research from Google arguing that an AI agent’s effectiveness depends not only on the language model but also on everything surrounding it—the prompts, context, tools, governance and orchestration.

“You could get significantly better results from a mediocre model if your harness is much better,” he says.

That observation changes where enterprise value gets created.

Instead of endlessly comparing benchmark scores, companies are beginning to ask how AI interacts with enterprise data, security policies and existing business processes.

“The model is only one layer,” Mathur says. ServiceNow believes the next evolution is autonomous workers.

Imagine every enterprise team having an AI colleague with defined responsibilities. “The system knows that this AI has certain skills, permissions and boundaries,” Mathur explains.

The work automatically flows to that AI worker just as it would to a human employee. The emphasis, once again, isn’t on intelligence but on execution.

India’s AI Debate Needs a Different Question

Mathur applies the same thinking to India’s AI ambitions.

Much of the discussion today revolves around whether India should build frontier models. He believes the conversation needs reframing.

“India should build a model,” he says. “But India does not need to build a frontier model.” Instead, India should build what he calls the cheapest acceptable model—one capable of solving Indian problems while leaving enough investment for GPUs, data centres and applications.

“There is a finite amount of money,” he says. “Rather than spending everything trying to build the best frontier model, distribute the investment across the ecosystem.”

His advice to startups is equally direct. “Don’t build something that tomorrow a large language model can do.”

Mathur believes India’s biggest competitive advantage lies elsewhere. “India has been a land of developers,” he said. “The value in technology doesn’t last long in infrastructure alone. It comes from how you apply technology for the good of humankind.”

Although India consistently ranks among the more AI-ready economies, Mathur says enterprises continue facing the same bottleneck seen globally.

“We [ServiceNow] do an AI maturity index every year, and India always comes up as one of the top countries because we have a younger demography and we’re much more open to adopting new technologies.”

But that enthusiasm often fades once pilots need to become production systems. And that’s exactly where ServiceNow sees its opportunity.