23 - 09 - 2026

India’s Enterprise AI Opportunity Lies Beyond Pilots, SAP Labs India MD Says

Enterprise artificial intelligence must move beyond isolated pilots and task-level automation if it is to fundamentally change business processes, Sindhu Gangadharan, Managing Director of SAP Labs India and Head of Customer Industry Solutions at SAP, said at the India Advantage Summit 2026.

Speaking during a fireside discussion titled “Leadership in the AI Era,” Gangadharan said the next phase of enterprise AI would depend less on experimentation with increasingly capable models and more on understanding the business problems organisations are trying to solve.

“AI is not just a technology transformation. It is actually a leadership transformation,” Gangadharan said. “What is the transformation that you’re trying to drive? What are the outcomes that you’re trying to solve for?”

The session was moderated by Jose Jacob K, Chairman of the TIA Summit and Future ICT Forum.

Gangadharan argued that business context would increasingly determine the effectiveness of enterprise AI. SAP, she noted, has systems touching 87% of global business commerce, while 99 of the world’s 100 largest companies use SAP, according to figures cited during the discussion. The company also draws on business knowledge across more than 26 industries.

That positioning, she said, allows AI to be embedded directly into business processes such as procurement, strategic sourcing, lead-to-cash and workforce management rather than being treated as an additional layer on top of existing software.

SAP’s knowledge graph connects business data, applications and relationships between business objects, providing AI systems with greater visibility into how an organisation operates. Its business AI platform combines AI skills, agents, assistants and industry knowledge, while its Joule copilot enables employees to interact with enterprise systems conversationally, including through voice.

The distinction is important as companies move from experimenting with AI tools to redesigning business processes around specific outcomes.

“It’s not just about taking AI and applying it on an existing process,” Gangadharan said. “That’s not going to take you far at all.”

India’s opportunity in AI at scale

The discussion also focused on India’s potential to develop AI applications that are affordable, multilingual and capable of serving large populations.

Gangadharan linked that opportunity to India’s digital public infrastructure, which has enabled digital services to operate at a scale of more than a billion people. Similar foundations, she said, could support AI applications in sectors such as agriculture, healthcare, education and financial services.

India’s talent pool is another factor. Citing the Stanford AI Index 2026, Gangadharan said India ranked second globally in AI talent, with more than 2.75 million AI professionals and annual growth of 55%.

She also pointed to the country’s linguistic diversity and the development of systems designed to operate across more than 22 Indian languages. Language-specific optimisation could improve both accessibility and the cost of deploying AI, she said, noting that standard models can take three times as long to process Hindi compared with optimised approaches.

Initiatives such as Bhashini, combined with India’s startup ecosystem, could help demonstrate how advanced AI systems can be delivered at mass-market cost.

“India doesn’t necessarily have to compete only on frontier technology,” Gangadharan said. “We can also demonstrate how to make AI accessible to a billion people at an affordable cost.”

SAP Labs India, which Gangadharan described as SAP’s second-largest research and development hub outside Germany, is also working with startups in areas where specialised technologies can be combined with SAP’s enterprise reach. Examples discussed at the summit included supply chains, procurement, logistics, carbon neutrality and AI-enabled battery operations.

Jacob illustrated the potential with a hypothetical financial-services use case for small businesses. An AI agent could explain a merchant’s transactions, seek permission to review UPI history, generate an immediate credit assessment and potentially offer a loan linked to future sales.

Responsible AI cannot be separated from speed

The discussion also addressed the tension between rapid AI adoption and responsible deployment. Gangadharan said the two should not be treated as opposing objectives, particularly for enterprises handling sensitive operational and financial information.

“We can’t see this as separate things,” she said. “Speed combined with safety and security is an absolute no-brainer.”

She linked the issue to the trust customers place in enterprise software. Around half a million customers rely on SAP applications, according to figures cited during the discussion. “Trust is earned over the years, and we know it can be lost in a second,” Gangadharan said.

While leadership has to establish the boundaries for AI deployment, she said those guardrails should not prevent experimentation. Employees and partners should be able to test ideas, learn from failures and refine solutions within clearly defined limits.

SAP’s approach includes developing focused models for specific business domains while also integrating external models into its platform under controlled conditions. The objective is to give organisations access to different AI capabilities without removing them from an enterprise governance environment.

The discussion also extended responsible AI to the future of work. Gangadharan said businesses should look beyond efficiency and potential job displacement towards upskilling, relearning and helping employees apply their capabilities more effectively.

In high-consequence areas, she stressed that human oversight remains important. Medical diagnosis, for instance, could become significantly more efficient with AI assistance, while the clinician would continue to evaluate the output and determine treatment.

AI works in production?

For enterprise AI, the more difficult challenge may not be building a model but identifying the right problem to solve and integrating the resulting system into real operations.

Gangadharan cited a billing agent developed for a global professional-services company with more than 800,000 professionals. The system helps determine employee deployment across projects while accounting for regulatory and legal requirements, pricing structures, costing options and professional seniority. According to Gangadharan, the solution has replaced work previously performed by more than 1,000 executives.

In another case, SAP worked with one of the world’s largest vaccine manufacturers to use multiple AI agents to rethink manufacturing processes. Vaccine production involves thousands of components, each subject to considerations including supply constraints, regulatory requirements, availability, cost and delivery schedules.

SAP has also developed three agents for manufacturing quality management to support the 8D problem-solving process, a methodology originally developed by Ford in the late 1970s. Gangadharan said the system is being deployed across hundreds of plants.

She attributed these deployments to their focus on specific operational problems rather than technology demonstrations. SAP’s forward-deployed engineering teams work alongside customers, combining technical expertise with an understanding of business processes and the employees who carry them out.

“Building is much faster these days because we have the technology,” Gangadharan said. “But if you don’t know what to build, that’s where the problem comes in.”

Data quality and architecture remain equally important. AI systems need access to appropriate datasets and a clear understanding of how those datasets should be used to train or guide models. In precision-sensitive environments such as vaccine manufacturing, probabilistic outputs also require strong controls and validation.

The discussion at the India Advantage Summit ultimately placed the enterprise AI challenge less on the availability of increasingly powerful models and more on the ability of organisations to define meaningful problems, establish trust, bring domain experts and technologists together, and redesign processes around measurable outcomes.

For India, the opportunity extends beyond adopting AI developed elsewhere. Its combination of digital infrastructure, technical talent, linguistic diversity and a large domestic market could also provide a testing ground for AI systems designed to operate at scale and at significantly lower cost.

By Subramanya Joshi