Indian IT majors flag trust, data readiness as enterprises scale AI

Representative image

• Indian IT firms: Scaling AI requires trust, data readiness, and strong governance.
• Enterprises struggle with data quality, explainability, and organizational readiness.
• Governance and trust must be built into AI systems from the start

Did our AI summary help?

Indian information technology (IT) services companies are warning enterprises that scaling artificial intelligence (AI) beyond pilots will depend on trust, data preparedness, and governance, as AI moves from experimentation to core business.

Executives from two of India’s top IT companies, HCLTech and Infosys, said enterprises are past the awareness stage of AI adoption, but face tougher questions around data quality, explainability, and organisational readiness as deployments expand across workflows.

Gaurav Dhakar, Global Head of Industry AI Solutions at HCLTech, said most organisations have moved beyond basic AI awareness and are now navigating a multi-stage maturity curve.

“So most of the organisations today have passed by the awareness milestone,” Dhakar said at the 20th India Digital Summit in Bengaluru on January 29, describing an S-curve that runs from awareness and enablement to integration and autonomy.

“Do you have data for it?” he asked, highlighting that many enterprises are working backwards to structure datasets so AI systems can deliver meaningful outcomes.

Trust, explainability take centre stage

Ashok Panda, Vice President and Global Head of AI and Automation Services at Infosys, said trust becomes non-negotiable when AI systems move into regulated environments such as banking, healthcare, and life sciences.

“AI, by nature, is more of a probabilistic model,” Panda said, adding that enterprises accustomed to deterministic systems struggle with this shift. “If it is a life science client or a financial service client, the data has to be accurate. You can’t have 99 percent accuracy, it doesn’t work.”

“Every time I ask the same question, it should give the same answer,” he said, adding that privacy controls must ensure models do not access data users are not authorised to see.

He warned that risks multiply at scale as AI systems interact with siloed enterprise data.

Scaling depends on governance

Both executives agreed that governance cannot be bolted on after deployment.

Dhakar said leaders are increasingly concerned about how success is measured as AI adoption changes cost structures and productivity metrics. “Your measurement criteria are changing,” he said.

Panda said governance and trust must be designed into systems from the start. “Trust cannot be an afterthought,” he said.

As enterprises push AI deeper into operations, Indian IT majors say the next phase will be less about hype and more about fundamentals, clean data, clear accountability, and systems that can explain their decisions at scale.