Most global software companies treat their India operations as the engine room: essential, enormous and rarely consulted on where the ship is headed. ServiceNow has spent the last few years dismantling that assumption, and the numbers are now hard to argue with. More than 45 percent of the company’s global product engineering is carried out in India, which accounts for over 20 percent of a workforce of roughly 28,000, according to figures the company shared earlier this year. Around 6,000 of those employees sit in Hyderabad and Bengaluru, and Hyderabad is ServiceNow’s largest engineering site anywhere in the world, as chief technology officer Pat Casey has put it.
For Sumeet Mathur, Senior Vice President and Managing Director of ServiceNow’s India Technology Centre, the logic is almost arithmetic. “India is a very critical market, a critical geography for ServiceNow,” he says. “In every company, there are two major functions: the build and the sell. If you look at the build part of it, 50 percent of all ServiceNow’s product development happens in the country.”
Why should a CIO care where a vendor’s engineers sit? Because the answer increasingly determines what lands in your roadmap, and because Indian enterprises are in the most interesting, and most precarious, stage of enterprise AI: spending faster than almost anyone, and governing slower than they should.
A build centre with a vote
Mathur is careful to separate headcount from influence. The India centre, he argues, is not a delivery shop executing instructions from California.
“We have a lot of leaders who are global engineering leaders for their product line,” he says. “Whether it is a hundred-million-dollar product line or a billion-dollar one, it doesn’t matter. So it is not only the quantity, but also the decision-making power. They are shaping the product portfolio and the new tech that is coming out.”
The ownership is concrete. ServiceNow’s platform line, which the company does not monetise directly but on which everything else is built, has deep India engineering roots. Mathur says the company is known as a workflow orchestration company, and that the workflow designer sits at the heart of that identity.
Strategic portfolio management, he notes, is run entirely by an India-based leader, and the software and hardware asset management lines are “completely run out of India.” He also points to India’s involvement across IT service management, HR, customer service, and security and risk.
The pipeline is still widening. The India team tripled in size in the two years to mid-2025, and engineers make up about 85 percent of it. The company has said it plans to keep hiring in Hyderabad this year, with a premium on AI engineers, machine learning specialists, research scientists and forward-deployed engineers.
On the sell side, Mathur describes India as one of the four or five prioritised geographies for ServiceNow globally, a view that senior leadership has echoed, with President and Chief Product Officer Amit Zavery recently describing India as one of the top five countries in the world in terms of value to the company.
But Mathur says the go-to-market story in India plays out along three distinct dimensions at once.
The first is the familiar one: Indian enterprises buying for themselves. The second is the global customer whose technology ownership has migrated into its India centres. “We have a lot of global customers here working from India who are basically rising up the value chain in their own organisation, and hence the decision-making and influence is coming,” he says. The third is the partner ecosystem, which spans the global integrators and India-based systems integrators alike.
The second cohort, the global capability centres, is where Mathur sees the most strategic leverage. “Because ServiceNow is present in almost 90 percent of all Fortune 500, almost every GCC of sizable scale is our customer already,” he says. The GCCs, though, are no longer just users. “They have a large technology base here. So it gives us an opportunity, and we are doing it in multiple ways, co-designing the solutions with them, which also helps us build better products and also makes sure that as they use those products, we have stickiness.”
The company’s Hyderabad innovation centre, opened in 2023 with nine purpose-built rooms, is where much of that co-design happens: customers arrive, whiteboards come out, and engineers who will write the code sit across the table from the people who will live with it.
India needs AI, but AI needs India
There is an old line in Indian technology circles that India needs AI. Mathur flips it. “India needs AI, but AI needs India,” he says. “For AI to work better, you need more and more data to fine-tune it. That is why while India needs AI, AI needs India.”
His argument is about scale. Indian conglomerates run more employees, more systems and more legal entities than most of their European counterparts, and building products against that complexity is a forcing function. “It helps shape our product strategy and roadmap much better,” he says, describing three benefits: performance at scale, better product direction from demanding technical customers, and the leverage of a large partner ecosystem on the doorstep.
Where the moat actually is
Mathur is blunt about the commoditisation of models. Frontier and open-weight models are available to everyone, which raises an uncomfortable question for any enterprise hoping to differentiate with AI. “If every enterprise has access to the same frontier models or open-weight models, where is your competitive moat?” he asks. “The competitive moat for enterprises is in the context of their business. The context is the data that flows through them. The context is the intelligence they have built over the years, whether it’s in their humans or in their systems.”
His case for ServiceNow follows from that. Much of that context lives where work is orchestrated, and the ability to ingest data from anywhere, maintain zero-copy relationships with partners and apply end-to-end workflows on one integrated platform is, in his telling, the advantage. He dismisses the alternative bluntly: “You can’t do an autonomous workforce if you are building on top of silos.”
He is equally unthreatened by the model makers themselves. “Whether it’s Microsoft or Google or Anthropic or OpenAI, you can go and find that we are partners with everybody,” he says, adding that the same is true of the Accentures and Deloittes of the world, and the Databricks and Snowflakes. “Everybody runs us.”
The user experience he envisions is deliberately boring. “As an employee, you don’t actually care whether your problem is with an IT system or a payroll system or HR. You just want to say, I have a problem with this.” The orchestration, identity and routing across HR, ERP and CRM systems should happen out of sight.
The governance gap
That vision runs straight into the data on Indian readiness. ServiceNow’s Enterprise AI Maturity Index 2026, which surveyed 4,500 senior leaders including 350 in India, found that Indian enterprise AI investment grew 119 percent in a year, ahead of the 110 percent global figure. AI already consumes 16.6 percent of the average Indian IT budget in the survey, projected to reach 21.3 percent by 2027.
The maturity numbers are less flattering. While 54 percent of Indian organisations are deploying AI agents, only 11 percent have moved to autonomous workflows. Just 22 percent have AI testing, auditing and risk-assessment processes in place, and only 18 percent have replaced fragmented legacy systems with an integrated platform. AI-enabled workflows scored 41 out of 100. Casey, for his part, has said that most new automation projects with major Indian customers in banking, manufacturing, technology services and GCCs now include agentic elements from the outset. The ambition is clearly there. The plumbing is not yet.
Mathur’s prescription is to stop bolting AI onto the old process. “You can bolt AI on top of your existing things, but it doesn’t impact your top line or bottom line,” he says. “But if you take a step back and reimagine your business process, then you spot opportunities to do much better.”
He also sees the industry’s vocabulary shifting. “There will be a lot more autonomous workforce in the enterprises, end-to-end and outcome-driven,” he predicts. “People will start talking more about the outcomes and less about the number of agents I have in the system.” Most organisations, he observes, have pointed AI at productivity and efficiency, “the bottom-line plane,” while “very few have focused on reimagining their business processes to drive more revenue,” the top-line plane.
Security, and the harness problem
Autonomy raises the stakes on security, and Mathur’s view of where risk sits is instructive for any CIO writing an agent policy. “The models never go rogue,” he says. “The model is immutable. Once the model is released, you and I interacting with it doesn’t change it. It is the harness on top of it that orchestrates it: what permissions, what access, what tools.”
That is the thinking behind ServiceNow’s push into security, which includes the Veza identity acquisition completed in March and the $7.75 billion Armis deal that closed in April. By correlating the identity graph Veza brought in with the company’s CMDB asset graph, Mathur says, a customer can see permissions across the infrastructure, identity and asset layers at once, “so I can remediate it faster.” It is also why he describes ServiceNow as now the eighth-largest enterprise security company, and the fastest growing. For Indian manufacturers and hospitals running ageing industrial and medical systems, he adds, the industry has largely “woken up to the reality” that visibility into those estates is no longer optional, though maturity varies widely by sector, with financial services further ahead.
What comes next
The global financials suggest the bet is paying off. ServiceNow reported second-quarter subscription revenue of $3.877 billion, up 24.5 percent, and said its AI business crossed $1 billion in annual contract value. But Mathur is less interested in the scoreboard than in the shift in the questions CIOs now ask him.
“We are getting fascinated by which is the leading model of the week, rather than trying to solve the problem for the enterprise,” he says. “But it is heartening that when you talk to the CXOs, they are all now coming back to: what problem does it solve? We have moved from ‘should we use AI?’ to ‘I am using AI, but I am not driving outcomes.’ The conversation has pivoted.”
That pivot is the real story of Indian enterprise technology in 2026. The country has proved it can build the software, staff the centres and write the cheques. What remains is the harder, less glamorous work: untangling legacy estates, governing the agents already deployed, and rewiring processes so that AI shows up on a CFO’s income statement and not just on a vendor’s slide.
If ServiceNow is right, the winners will not be the enterprises with the most agents or the newest model. They will be the ones that understood, early, that the intelligence was never the scarce resource. The context was. And in India, there is no shortage of context, only the question of who will have the discipline to put it to work.