For nearly 30 years, Virtusa operated as a product and platform engineering services company. But today, it is aggressively repositioning itself for the age of generative and agentic AI.
The company’s new offering, Helio, avoids the heavy, one-size-fits-all platform model. Instead, it delivers modular, configurable AI solutions that plug directly into whatever tech stack an enterprise already uses.
In a conversation with AIM, Nitesh Banga, CEO of Virtusa, explained how the company sees AI adoption evolving across industries, why organisations struggle to build these capabilities in-house, and what makes ROI so complex in this space.
This shift in enterprise AI maturity comes as global investors are split on the trajectory of the AI boom. While organisations are experimenting aggressively, market commentators are questioning whether the sector’s rapid acceleration is built on solid fundamentals.
Responding to broader industry debates, including recent commentary from figures like Michael Burry about the AI bubble, Banga believes that the AI cycle resembles early cloud adoption, not the blockchain hype curve. “AI is probably the next cloud, not the next blockchain,” he said.
Banga stated that fluctuations are inevitable and expressed his strong belief that AI and agentic AI are here to stay. He highlighted that stability will only be achieved once enterprises modernise their data infrastructure and complete the foundational work that underpins AI.
Besides, Burry, Mark Cuban, investor and entrepreneur, recently warned that many AI investments are overspending, drawing parallels to the 1990s search-engine/tech race and saying that the current AI frenzy could end like a bubble.
On the contrary, IBM CEO Arvind Krishna said in a recent podcast that there is no AI bubble. He noted that while some of this capital will inevitably be wasted, the broader economics of AI still justify the aggressive investment, particularly in the consumer AI space.
Krishna said that a company building a highly attractive AI model with gains of half a billion users can generate significant value. “If you build a slightly better model by spending another $50 billion and that can attract another 200 million users, it seems to make economic sense,” he explained.
Meanwhile, Virtusa, supported by Helio, aims to become an AI-first services company by 2030.
“Helio will be at the tip of the spear, and all our other service lines will attach to it,” Banga said. The company reported that about 17 % of its current revenue is already AI-first and that this portion is growing at a higher profitability than traditional digital services.
What About ROI
One of the biggest misconceptions in enterprise AI is how to measure ROI. According to Banga, a distinction must be made between gross ROI and net ROI.
Productivity improvements may reduce headcount requirements, but once the costs of compute and models are added, the real benefit often shrinks or disappears. He believes the industry is obsessing over the wrong metric.
“There is a gross ROI, and there is a net ROI. You may reduce the work from 100 people to 40, but you must add the cost of compute…to know the real ROI,” Banga said.
He said that the world of automation has undergone a dramatic shift over the past decade. Enterprises first embraced deterministic automation through RPA, then moved into predictive analytics and later into cognitive automation. But the arrival of generative AI has completely changed the equation.
“The flip has happened from automation to creation,” he said, describing how AI systems are now capable of generating artefacts, decisions and experiences that never existed before.
Within the product development lifecycle, Virtusa sees an enormous opportunity beyond simple coding assistance. While the industry talks endlessly about AI coding tools, Banga noted that coding accounts for only a fraction of the lifecycle. “Coding efficiency is only 30% of the PDLC. The real impact lies in the rest of the lifecycle,” he said.
The IT services company is investing heavily in agentic testing, architecture generation, reverse engineering for forward engineering and AI-based CI/CD. Their testing suite, for instance, can reduce testing time by up to 60% by automatically generating test cases, test data, and execution flows.
Why Enterprises Can’t Build AI Platforms Alone
Despite the excitement around AI, most organisations face a serious reality check when they try building their own AI platforms. The challenge begins with data readiness.
As Banga put it, many organisations still have fragmented data, no single source of truth, and incomplete cloud journeys. Even before modelling begins, enterprises must realign infrastructure, prepare data, set up compute resources and modernise legacy systems.
While enthusiasm is high everywhere, as the CEO, he believes no sector can yet be called truly mature. He shared that tech ISVs (Independent Software Vendor) such as Google and AWS are long-time Virtusa customers.
Banga mentioned that among traditional enterprises, some sectors are progressing faster than others. Communications firms are moving ahead as network optimisation and software-defined networks become increasingly AI-driven.
Life sciences companies, especially in early-stage drug discovery and clinical research, are rapidly incorporating AI into lab and R&D workflows. Financial services players are using AI to streamline intensive operations like onboarding and KYC. Banga added that even manufacturing is exploring computer vision for warehouse optimisation and automation.
Still, he believes that the journey is just beginning: “I wouldn’t say any sector is mature, to be very honest. The industry is still early in the journey.”
The post ‘AI is Probably the Next Cloud, Not the Next Blockchain’ appeared first on Analytics India Magazine.
