Babak Hodjat, Chief AI Officer, Cognizant
Enterprises are coming to the realisation that AI is no magical “pixie dust” and customisation with business processes is key to reap benefits of Agents, Cognizant’s Chief AI Officer Babak Hodjat, said.
Speaking to businessline on the sidelines of the AI Summit here, Hodjat said that IT services companies “own the last mile” and understand the client’s domain inside out, and that context is the missing piece in the ‘agentification’ journey.
“Agents need to be designed and defined to model the processes and organisations of our clients, and needs to be done in a trustworthy way,” he said. “There’s customisation, tailoring, engineering, safeguarding that has to happen,” he adds. Cognizant, he says, is not just ready for this transition but also “ahead of everyone else,” noting their investments in this regard including in AI labs.
Cognizant recently opened its India AI Lab in Bengaluru. The India AI Lab extends Cognizant’s AI Lab in San Francisco, which was recently granted its 61st US patent.
On partnerships
Speaking about the company’s recent partnerships with various AI native startups including Anthropic, Hodjat called it a win-win tie-up. “While we are empowering our associates to use the models and bringing them to our clients, Anthropic also wants to be a serious, reliable player in the enterprise world, and we can help them with that,” he said.
The AI chief at the Nasdaq-listed major says they are not just using AI internally for employees and project delivery but also measuring the impact of it on net new logos [clients]. “We’re measuring how much we’re actually making use of AI and delivering to our clients and what kind of productivity we are bringing to the table,” he said.
Laying down the ways in which the company, which has majority employees in India, is preparing its workforce for the transition, Hodjat said they are running boot camps, hackathons, and also specialised context engineering courses. Another way is by actually hiring more, he points out. “We’re looking at freshers and bringing a whole bunch of them into the company because they understand AI and they use AI much more naturally,” he said.
While higher up in the pyramid, there’s scepticism and there’s reluctance and conservatism, lower down you have ‘naivety,’ he says. However, this can be tackled by the right training and we need that energy and innovation, he added.
AI Labs ecosystem
Speaking about the AI Labs ecosystem, Hodjat calls it a place where they do core AI research. One major breakthrough they have recently seen is in fine-tuning large language models, where the Lab invented a new way to fine-tune that uses Evolutionary Strategies instead of Reinforcement Learning,” he said. This method, essentially, is less compute-hungry. “In India, in particular, we’re researching multi-objective reasoning systems where you’re trying to reason for more than one outcome at the same time, and we are also looking at joint research with IITs and others,” he said.
In India, the AI Labs core team is around 30 people, mainly AI PhDs and researchers, and some AI engineers. “We want the research to be cutting edge, and want it to be grounded in what our clients are looking for,” he says.
And what stands in the way of completing the transition from IT services to AI services? “We just need that expectation to be tapered down — that AI is not magic but just an engineered principle,” he said.
Published on February 20, 2026
