Infosys is looking to transition anywhere between a quarter to a third of its workforce into specialist positions and also create a team of about 6,000 frontier engineers, according to people who attended a recent company event where the plans were discussed.Infosys has more than 320,000 employees, but the growing impact of AI on billing models based on the number of people deployed on projects is putting pressure on the traditional headcount-driven pricing structure.An Infosys executive’s presentation showed that 270,000 employees had received AI training, while 80,000 were engaged in working with coding tools.
Infosys hiring and specialist workforce plans
The Bengaluru-headquartered IT services company is recruiting talent from leading Indian institutes as well as US campuses, with the company often competing with AI labs for the same pool of candidates.The management shared these details with partners, advisers and analysts during a three-day annual Americas Confluence held in Washington, DC, last week, people familiar with the discussions told ET.Three analysts separately told the financial daily that the country’s second-largest IT services company acknowledged this AI-led challenge.Also Read | IITs to stop campus hiring by 22 companies for 2 years after 150+ job offers to 2026 graduates went unhonoured; Oracle reportedly withdrew around 40 offers“Infosys said it hasn’t restructured, which is a noticeably different approach from what many companies are doing right now. It says a lot about how it’s thinking about talent and change management, and about where it expects to need people in future,” said Dana Daher, executive research leader at HFS Research, who attended the confluence.“It’s also hiring from top Indian institutes and US campuses … In some deals it’s taking on clients’ surplus staff and redeploying them,” she said.CEO-designate Ashiss Kumar Dash and incumbent CEO Salil Parekh shared the stage to discuss ways of generating value from AI.The company, which has revenue of $20 billion, is navigating the challenges created by one of the most significant technology shifts faced by the IT industry.The central theme of the flagship regional thought leadership event was helping enterprises “unlock AI value” while moving from experimentation with AI to deployment at scale across organisations. The same broad message has featured across Infosys’s 2026 Confluence events in Asia Pacific, Europe, the Middle East and Africa, and the Americas.Senior management discussed the company’s strategy alongside broader conversations on AI, its implications for pricing, the approach to enterprise AI and plans for building AI capabilities at scale.Apart from Parekh and Dash, the speakers and panellists included Anant Adya, head of Americas delivery and leader of Infosys Cobalt, the company’s AI-powered cloud platform, as well as Anand Swaminathan, executive vice president and global industry leader.Speakers at the event also featured Verizon CEO Dan Schulman, Wharton School professor Ethan Mollick and former NASA astronaut Peggy Whitson.As AI-driven productivity continues to reshape delivery models, Infosys is placing greater emphasis on outcome-based pricing and maintaining a model-agnostic strategy. The company is working with open-weight models for enterprise customers as demand grows for greater control over their data.In a presentation, CEO Salil Parekh reiterated the company’s first-quarter disclosure that AI-related revenue accounted for 8% of total revenue, amounting to $1.6 billion, and was growing at a double-digit rate on a quarter-on-quarter basis.Parekh also identified six emerging areas where AI could create new services opportunities: engineering, data for AI, process AI, agentic legacy modernisation, physical AI and AI trust.According to Dana Daher of HFS Research, Infosys highlighted its cloud-first, AI-first composable stack, Topaz Fabric, which combines AI agents, services and models. The platform is now processing around 100 billion tokens, with 35 implementations completed and more than 100 underway.“It routes work across roughly 50 models, and Infosys has tuned some open-weight models for clients who want more control over their data. It’s also now packaged in a way that’s easier to buy and deploy,” Daher said.She said Infosys was approaching outcome-based pricing cautiously and described the future business model in terms of “human capital, token capital and IP”.“Infosys is moving toward that deliberately. The risk for the whole industry is a race to promise ever-bigger savings, with nothing left to reinvest in talent and innovation,” Daher said.Daher also said Infosys had characterised itself as a “custodian” of client context. An energy-sector client, meanwhile, said partners contribute to scaling that context but do not own it.“That tension, over who holds the knowledge, is going to shape a lot of these relationships.”Holger Muller, principal analyst at Silicon Valley-based Constellation Research, also posted about the confluence on microblogging platform X. He said Topaz Fabric had onboarded 35 clients, while 118 implementations were in progress, spanning more than 22 business and IT focus areas.