Roughly 80% of AI projects fail to deliver what organisations expect, according to Julen Mohanty, VP of IT Automation and CIO at Gallagher's India operations, in an interview published by Express Computer. Mohanty, who oversees technology for the insurance broking firm's nearly 19,000 employees across nine Indian cities, attributes the problem not to the technology itself but to what companies expect it to accomplish. His central argument is that enterprises too often begin with the technology rather than the business problem it's meant to solve.
Mohanty's test for any AI initiative is straightforward: does it increase revenue, reduce cost, or manage risk? At Gallagher, automation has transformed processes like proposal generation, which previously required pulling information from more than a dozen systems, though the output still passes through human review before reaching customers. Time savings from automation vary significantly across different processes, ranging from under 30% to more than 70% depending on the workflow. Gallagher's India operation, which primarily supports the company's US and UK businesses while expanding domestically following its acquisition of Edelweiss General Insurance, began around 2006-07 as a contact centre before growing into broader technology support and business operations.
"If you have a hammer in your hand, you start searching for nails everywhere," Mohanty says, describing how the rapid availability of generative AI and increasingly capable agents has created a temptation to insert technology rather than first ask whether a process actually needs it. He considers headcount reduction the wrong starting point, arguing that automation may eventually change the amount of human effort required but that should be a consequence of improving the process rather than the objective. Mohanty coined the term "Humbotai" nearly a decade ago, combining human, bot, and AI, based on the observation that while a machine may execute an action, responsibility still belongs to a person. "You cannot have accountability to a bot," he notes. "You cannot fine a bot, you cannot penalise a bot."
The core issue, according to Mohanty, is that companies look for places where technology can be inserted rather than identifying where it can produce a measurable business outcome. This approach runs counter to how many enterprise AI programmes are being conceived amid the growing capabilities of AI systems. His philosophy extends to data: "Data is not the new oil, data is a new cash machine," he says, emphasizing that data becomes valuable only when an organisation knows what it contains, connects it to a business problem, and can act on it. At Gallagher, this means measuring process strength before automation, identifying which steps can be automated, establishing baseline timing, and then measuring what changes afterward. The variation in results challenges the assumption that AI produces a standard efficiency gain that can simply be applied across an enterprise.
Mohanty's outlook extends beyond individual projects to India's global capability centre evolution. He questions whether the term accurately describes what many such operations actually do, estimating that fewer than 5% of GCCs in India undertake meaningful product innovation or file patents. While India has established itself as a formidable technology delivery base, he argues the next step is creating more intellectual property rather than primarily executing work defined elsewhere. On security, as AI systems become more autonomous, he compares the challenge to checking a drum for leaks: you inspect it first, then put it under water and look for bubbles. "You cannot stop the rain," Mohanty says. "You have to carry an umbrella." For a CIO navigating the current AI cycle, it's also a useful definition of the job—technology will keep changing, but the harder task is deciding where it belongs, proving that it works, and ensuring someone remains accountable when it doesn't. The insurance sector's emphasis on human oversight in consequential decisions suggests that businesses betting on full automation may be solving for the wrong variable. Organisations chasing efficiency gains without first establishing what success looks like risk building systems that satisfy no stakeholder beyond the technology team itself.