generative artificial intelligence: GenAI production deployments on the rise in India: EY India
Additionally, 15-20% of domestic enterprises and 30-40% of global capability centres (GCC) of multinationals have rolled out gen AI PoCs, EY’s analysis of the last two quarters showed.
“Clients are doing a lot of experimentation but scaled production deployments of use cases are still few and far apart,” Mahesh Makhija, technology consulting leader at EY India, told ET.
At the same time, AI, gen AI, and businesses having to get their data in order so as to leverage these technologies, is likely to be one of the largest drivers of revenue in fiscal year 2025, leading to sizable gen AI deals for Indian IT, he added.
“FY25 will be a year where people will get a little more organised on the kinds of experiments and PoCs they’re doing, so there will be a large number of these projects which will happen in a slightly more coordinated fashion. But I expect to see pretty large deals being run by these IT services companies,” Makhija said.
This fiscal year will also likely see deployments of some of the first AI agent bots in areas like customer service, sales, underwriting and collection. Though the technology for this is as yet a work-in-progress, Makhija said the return on investment can be “dramatic”. He added that some AI agents could potentially lower business’ costs by 60-80%, particularly if they can be leveraged for more than a single use case.
Discover the stories of your interest
EY found that the gen AI PoCs that are succeeding in going into production are in areas like document intelligence — for instance, bots that can generate answers based on underwriting documents for financial services or knowledge for sales agents. Gen AI for productivity assistance, such as to transcribe patient-doctor interactions in medical settings, and bots for marketing automation are also going into production.Companies are witnessing value uplift in delivery-related areas like application development and support, infrastructure and operations, cloud migration and workload modernisation, EY has found. Financial services, healthcare and IT services are the fast-moving sectors for AI adoption.
There are challenges, however, such as regulatory uncertainty and unwillingness of customers to sign up for long-term projects because of concerns around data accuracy, misinformation, bias, security and privacy.
“As you start getting more and more into the companies which are aggressively trying to set up their own custom models, etc., that’s where there is some level of hesitation because the law is not yet clear,” Makhija said.
Since many businesses are still experimenting, they also hesitate to pay upfront the large costs needed to access the advanced large language models in the market.
