In a detailed interaction with ET Now, Kumar Rakesh, Analyst, BNP Paribas laid out a nuanced view of how AI could reshape traditional IT services — warning of near-term pressure but arguing that the longer-term picture may be more balanced than current market reactions suggest.
Two Phases of AI Impact
According to Rakesh, AI implementation in enterprises is unfolding in two distinct phases.The first is automation of business processes — a continuation, but at a much faster pace, of what robotic process automation (RPA) had already begun.
“So, we see AI getting implemented in enterprises in potentially two phases or two parts of the enterprise. The first is the business processes. Wherever things can get automated, they will get automated. This part of the work was already being incrementally automated. We used to see the work being done by RPA, where 10% to 15% productivity benefit was already being achieved. What now we are seeing is all of those kinds of productivity benefits are being accelerated through the use of agentic AI models.”
These are structured, repeatable tasks governed by predefined workflows — ideal candidates for AI-led automation. The implication, he noted, is reduced manpower intensity and cost savings for enterprises.
“And that starts getting automated using AI tools. So, that is the first part, which is the business processes getting automated and, as you can imagine, this would be more deflationary in nature because this is largely for cost saving, reducing the work intensity which was earlier, and this is disruptive in a negative way.”
The second phase, however, is potentially more transformative — and positive.
“The second part to this is the business transformation. Now, the enterprise data is sitting and trained in all the large language models and using that some of the new applications can be built, maybe hyper customisation for their customers and that is something which we believe will start creating new use cases, more businesses for IT services companies, and that will be inflationary in nature and disruptive in a positive way.”
In essence, while process automation reduces revenue pools in the short term, AI-led transformation could create entirely new revenue streams over time.
Front-Loaded Disruption
The challenge for the sector lies in timing.
“Now unfortunately, the business process disruption is more front-loaded. We are already seeing the impact of that over the last one, one-and-a-half years. Some of it has already started showing up. But the business transformation inflationary new work will take longer because these models have to evolve a little more. Enterprises have to get more comfortable to use them and integrate them into their end businesses and that takes a little longer journey and potentially at some stage we would start seeing the benefit of the AI implementation as well for IT services companies.”
Rakesh believes the overall impact may eventually be neutral — but far from linear.
“Overall, we would believe that it would be neutral, but the impact will not be linear. First, there will be a deflationary pressure followed by inflationary pressure. Again, not every IT services company would be impacted equally. It depends on what they are more focused on, which part of the service line they are more focused on because the impact will not be linear and symmetric across all the parts of their services businesses and that is how we would expect this to pan out in the coming years.”
Are Markets Overreacting?
The recent selloff in Indian IT stocks has reignited concerns about long-term terminal growth. While industry body NASSCOM has argued that fears around AI disrupting Indian IT services are misplaced — suggesting instead that AI will augment capabilities — investor sentiment remains fragile.
Rakesh believes the sharp stock price reactions reflect deeper uncertainty.
“So, we will have to first understand what exactly investors are looking at or sending a message in terms of these sharp stock price reactions. What they essentially are looking at and when we are speaking with most of the foreign institutional investors, they are incrementally getting concerned about what would be the terminal growth of these companies, what would be the long-term sustainable growth for these business models.”
He acknowledged that investors themselves lack clarity.
“Now, the challenge which they are also facing, investors, is that we do not have a clear answer for this as yet that how actually this disruption plays out and who gets impacted what. And with that high level of uncertainty, at a time when the technology is evolving at a very fast pace, they are choosing to sit on the sidelines. They are trimming the position in this sector to sit on the sidelines to first better understand how this exactly plays out and then once they have a better understanding maybe come back and then start building position again in specific names who are the net beneficiary of this trend.”
Fear vs Greed Cycle
Drawing parallels to last year’s exuberance around data centre capex announcements, Rakesh described the current mood as part of a familiar market cycle.
“What we have seen especially over the last week, 10 days, couple of weeks is I would call something very similar we saw last year. If you recall last year any company which used to come out and announce their significant capital expenditure for data centre building, the stock used to react very positively to that and then came a time when stocks stopped reacting to it and now we are in a phase when any new expansion in capex is seen negatively and stocks have started reacting quite sharply on the negative side to it.”
His conclusion: sentiment has swung decisively toward fear.
“What I am trying to say is that there is a cycle of fear and greed which plays out and right now we are at an elevated level of fear. Any new frontier model getting launched or new version coming out is seen significantly as a negative to the sector, not just services but software as well. And we think that this is a cycle which will also peak at some stage. We would get to a point where any new model launches by these frontier models will not have any stock price reaction. We think that is a good trigger to watch out for and that will potentially start indicating this fear among investors has largely peaked and that could be a good entry point in the sector.”
Stock Picks in a Volatile Phase
With valuations for some IT majors approaching multi-year lows, Rakesh sees selective opportunity emerging.
“So, our top pick in the sector among the largecaps is Infosys and Tech Mahindra. Infosys, we see as one of the key beneficiaries of AI implementation because of the capabilities that they are building. We also see Tech Mahindra significantly transforming their business both in terms of revenue growth and margin and that should imply that they should grow faster than the overall industry. So, these two are our preferred pick among the largecaps. We like Persistent Systems among the midcaps. We think Persistent Systems is the key beneficiary of AI implementation. It is already showing up in the revenue performance which is starting to decouple from employee growth. So, we are already seeing some initial signs of that performance.”
His preferred large-cap bets include Infosys and Tech Mahindra, while among midcaps he highlighted Persistent Systems.