Mid-Size Indian IT Might Be ‘Vibing’ a Little Too Much with …

The confidence gap between larger firms and their mid-sized peers stands out in the Indian IT industry when it comes to generative and agentic AI. 

On the one hand are firms like TCS, Infosys, and Wipro, that are building hundreds of AI agents for clients but are still pessimistic about the upcoming quarters and do not expect much growth. On the other, smaller firms are vibe coding and reporting AI revenue, such as the startups reporting $100 million ARR with a single month of spiral growth.

Cognizant just made it into the Guinness world record for world’s largest online generative AI hackathon with 30,601 working prototypes. The vibe coding event had more than 53,000 Cognizant associates across 40 countries. 

CEO Ravi S Kumar is extremely confident with generative AI. He aims to generate 50% of the firm’s code using AI within a year and states it has already achieved 20% of that target. “We aim to be a pioneer in this area,” he said, which is in line with Cognizant’s announcement in 2023 of investing $1 billion in generative AI over three years.

What’s in it for Smaller Firms?

Other examples of AI confident smaller firms are Sonata Software and Happiest Minds, which are not only quantifying it, but projecting bold numbers. Coforge, Persistent Systems, Mphasis, LTIMindtree, and Tech Mahindra are building libraries of AI agents, rolling out new coding practices, and showcasing AI-driven platforms at a pace in comparison to which, the big four seem cautious.

The question is whether these smaller firms are genuinely ahead or whether they’re guilty of what the industry is beginning to call “agentic washing”.

In a recent interaction with AIM, Apoorv Iyer, global head of GenAI practice at HCLTech, one of the larger firms which is showing confidence with AI, said that there is the possibility of agentic AI washing in the IT industry, as reporting AI-related revenue is vague since it is now part of all services.

Read: HCLTech GenAI Head Calls Out Agentic Washing in Indian IT

Sonata Software is the most aggressive in its projections. CEO Samir Dhir has said the company expects AI-enabled services to make up 20% of its revenue over the next three years. Based on its current run rate, that could mean more than $300 million coming directly from AI by FY28. 

Also, Happiest Minds has already started breaking out generative AI revenue, with $1.48 million, or 2.3% of its revenue, coming from this segment in the last quarter alone.

These disclosures echo what Accenture and ServiceNow have been doing globally, but they are unusual for India. The largest IT firms have avoided separating AI revenue so far. Only TCS had reported its generative AI pipeline of $1.5 billion in Q1 FY25, but since then, it stopped doing so.

Infosys and HCLTech highlight the number of agents they’ve built, not what those agents are earning.

“There’s a lot of superficial claiming of agency, with people describing simple tasks and labelling them as agents. It is not really about the quantity of agents, it is about the complexity, quality and adoption of AI within your customer base and enterprise,” Iyer added.

In other words, everyone is chasing the AI label, but what that means in rupee terms is unclear. It is possible that mid-size firms might also be doing the same. 

To take another example, similar to Cognizant, Coforge is also experimenting with vibe coding. It has tested tools like Cursor and Windsurf, building a secure setup for its internal TechCon conference where engineers could describe applications and have them generated by on-premise LLMs. 

As for the revenue from all of this, Vikrant Karnik, EVP for digital, data, cloud, and AI at Coforge, told AIM that reporting AI revenue is not exact. “It’s a little bit like asking a hotel to list electricity as a revenue line,” he said. “We’re here to deliver shareholder value for our customers. If AI helps us do that, we’ll use it.”

Read: Coforge’s AI Reality Check for Indian IT to ‘Strip Away the Fat’

Mid-Caps Pulling Ahead with Agentic Washing?

Even with the ambiguity, the numbers tell their own story. Persistent Systems reported $389.7 million in revenue last quarter, up 18.8% year-on-year. Profits jumped nearly 39%. The company has now delivered 21 straight quarters of double-digit growth. 

Coforge grew even faster, up 56.5% year-on-year, with profits more than doubling. It also became the seventh-largest IT firm in India, surpassing Mphasis.

Both companies are attributing their success to AI. 

Persistent credits a platform-driven approach similar to ServiceNow, embedding AI into client projects instead of keeping it as a side offering, for its growth. Coforge has built AgentSphere, a library of 100-plus AI agents for sectors like travel and healthcare, and is betting on scaling them.

Mphasis booked its highest TCV in eight quarters, with 68% of those wins being AI-led. LTIMindtree launched BlueVerse, a unit offering 300 specialised AI agents across industries. Tech Mahindra deployed more than 200 AI agents while managing a 34% year-on-year jump in profit.

Compare this to the Big Four. Wipro doubled its large-deal bookings but still saw topline revenue decline. TCS managed just 1.3% growth, with constant currency revenue falling 3.1%, and Infosys delivered modest growth but still lagged behind mid-cap peers in pace and clarity.

Enthusiasm among smaller firms though, raises a red flag. Declaring 15–20% revenue through AI when the category is in its infancy might be overreach since even their larger counterparts choose not to do so. 

Many of these deals are small contracts. For mid-sized companies, a $5–10 million AI engagement can look transformational against a smaller revenue base. For giants like Infosys or TCS, the same deal is barely a rounding error.

Also read: AI ROI is Better for Coforge, Persistent, Mphasis, L&T than the Big Four Indian IT

The post Mid-Size Indian IT Might Be ‘Vibing’ a Little Too Much with AI appeared first on Analytics India Magazine.

Generated by RSStT. The copyright belongs to the original author.

Source