The enterprise adoption of AI, each in India and globally, has shifted away from employee-led initiatives to leadership-driven programmes, with senior executives and center managers now steering top-down deployment of the expertise.
As an example, over 80 per cent of Indian enterprise leaders mentioned that they’re already conversant in AI brokers compared to 66 per cent of workers, in line with country-specific findings reported by Microsoft India in its annual Work Development Index (WTI) 2025 report final month.
Globally, 79 per cent of leaders imagine that AI will speed up their careers versus 67 per cent of workers. It’s in stark distinction to final 12 months’s WTI report by Microsoft which confirmed that workers have been the important thing drivers of AI adoption.
This shift in momentum is what’s propelling generative AI pilots into manufacturing, in line with Himani Agrawal, Chief Working Officer (COO), Microsoft India and South Asia. “When it begins with the leaders, it’s severe adoption, it’s severe ROI and issues which imply enterprise returns. That’s the way it will get perceived, conceptualised, and adopted,” Agrawal informed The Indian Specific on the sidelines of the Microsoft WTI occasion in Noida not too long ago.
Her remarks come in opposition to the backdrop of an ongoing debate over who’s accountable, and who needs to be accountable, for the adoption of AI instruments inside an organisation. Final month, Coinbase CEO Brian Armstrong mentioned that engineers who refused to enroll to make use of AI coding instruments have been instantly fired. The crypto change’s mandate drew criticism for disregarding particular person selection.
The Indian Specific sat down with Agrawal in addition to Manpreet Singh Ahuja, Chief Digital Officer, PwC India; and Rajesh Kumar R, Government Vice President and Chief Info Officer, LTIMindtree, to debate leadership-driven AI transformation, the failings in generative AI pilots, AI fluency as a core talent, and extra.
Management-led AI adoption
When requested why this strategy was higher than employee-led initiatives, Singh pointed to a market pattern from 2023-24 the place most individuals have been skeptical of AI. “The dialog was, ‘please come do what you are able to do, so long as you ship the ROI again throughout the six-month or one-year interval’. The marketing consultant was sceptical and the client was sceptical,” he mentioned.
Story continues under this advert
“From there, the mindset shifted as a result of the management is getting extra engaged, they’re utilizing it at a private degree. The ecosystem is starting to make use of it. They’re much more certain and assured,” Singh added.
A latest survey of three,000 professionals by hiring platform Certainly discovered that center managers are main AI upskilling with 49 per cent of the respondents aged between 35 and 54 stating that they’re actively searching for extra AI coaching, in comparison with 41 per cent of respondents aged between 18 and 24.
In the meantime, Kumar mentioned that IT companies firm LTIMindtree needs to be “seen as an AI firm fixing our prospects’ issues with AI, and we should be adopters internally.”
Scaling generative AI pilots
A latest examine performed by Massachusetts Institute of Expertise (MIT) grabbed headlines after reporting that 95 per cent of 300 US-based companies that had invested someplace between $35 billion to $40 billion in generative AI, noticed little to no returns largely as a consequence of flawed enterprise integration of AI.
Story continues under this advert
Nevertheless, Singh struck a optimistic observe, evaluating AI adoption to enterprise capital investing: “They put cash throughout 20 startups, 14 of them could not work out however the six that do will create disproportionate worth. Now, in case you apply the identical mindset to an organization seeking to undertake AI, the place you let these 20 concepts come to life and supply seed capital, you should have the 5 superheroes come all the way down to the desk and unlock worth which is starting to occur.”
When requested about scaling AI pilots inside LTIMindtree, Kumar mentioned that the corporate has fostered a tradition of experimentation which inspires the idea of failing quick.
“You fail fast sufficient so that you simply don’t spend an excessive amount of cash. If you happen to begin counting each failure, it’d add as much as a share however total it will likely be a profitable journey,” he mentioned, citing the instance of LTIMindtree’s venture to offer each worker a digital companion. Kumar additionally mentioned that too many constraints can derail AI adoption and restrict innovation.
We allotted adequate assets to create a buffer zone for workers to experiment freely with AI and construct options with out worrying about consumption limits, he additional mentioned. “After we get to a degree of maturity, then the optimisation is available in, then our vast rollout is available in. That’s the tactic.”
Story continues under this advert
Knowledge readiness as a bottleneck
Efficiently scaling AI adoption additionally requires firms to organise datasets, as per PwC’s Manpreet Singh.
“Whenever you attempt to scale, you need the underlying datasets to be robust for you to have the ability to get foolproof responses whenever you’re coping with the client or worker. Whereas AI brokers can perform as retrieval instruments no matter whether or not you’ve the info or not, if you wish to hyperpersonalise that dialog again to each buyer, you want buyer information to know what they like and don’t like,” he mentioned.
“As you resolve that information layer, I believe AI will change into an increasing number of actual,” Singh added.
AI brokers and ROI
Over 59 per cent of Indian enterprise leaders are already deploying AI brokers to automate workflows throughout complete groups, in line with Microsoft’s WTI 2025 report, with 93 per cent of leaders anticipated to deploy AI brokers to increase workforce capabilities within the subsequent 12-18 months.
Story continues under this advert
Extensively touted as the following AI frontier, agentic AI options are additionally compute-intensive and costly to run. When requested how the AI brokers wager is affecting Microsoft’s bottomline, Agrawal mentioned, “Any new expertise, product, or service comes with excessive prices whenever you introduce it available on the market. However as you scale adoption, the fee all the time retains coming down.”
Costs per token have fallen steadily for the reason that launch of ChatGPT and Azure OpenAI because of the consumption curve and ongoing expertise innovation, she added.