Infosys CEO Salil Parekh laid out the most detailed picture of how Infosys is scaling AI inside the company and across its client base, even as the company continues to avoid bifurcating AI revenue. He was speaking during the company’s Q3 earnings call.
At the centre of this push is Infosys Topaz and a new agent services layer called Topaz Fabric, which now also includes Cognition’s AI software engineer Devin. Further, Parekh said the platform is designed to help enterprises deploy, manage, and govern AI agents across their clients, turning what were once experiments into production systems.
Parekh said that the Cognition partnership is not filling any gaps but is helping Infosys deliver AI agents into one of its clients’ environments.
“The advantage is we have a detailed understanding of the technology landscape, and we have a good understanding of the industry constraints or opportunities. And that, combined with the software agent with Cognition, becomes a very powerful combination in the merry clan,” he said.
The scale of AI agent deployment is now significant. Parekh said Infosys is working with 90% of its top 200 clients on AI programmes, a level of penetration that shows AI has moved far beyond pilots.
Across these engagements, Infosys teams are running 4,600 active AI projects. It claims to have generated more than 28 million lines of code using AI tools and built over 500 AI agents. The company is also expanding its forward-deployed engineering teams to embed these capabilities more deeply in customer environments.
Its large deal wins jumped to $4.8 billion in Q3, up from $3.1 billion in Q2. Parekh said it was led by contracts with AI-led modernisation, automation, and agent deployment as part of the core scope of work.
Parekh outlined how Infosys sees the AI market evolving in the months to come. He identified six emerging AI-led value pools that he believes will drive the next wave of enterprise spending. These include AI engineering services, AI data platforms, business operations agents, AI-driven software development and legacy modernisation, AI deployed on physical devices, and AI trust and risk services.
“These six AI-led value pools could unlock a large incremental opportunity,” Parekh said, adding that Infosys believes it is uniquely positioned to capture share across all of them.
At the same time, he acknowledged that AI is compressing parts of the traditional services business. Productivity gains from automation and agents are shrinking some legacy revenue streams even as new ones open up.
In context, its Q3 revenue rose 8.9% year-on-year to ₹45,479 crore, but net profit fell 2.2% to ₹6,654 crore as margins came under pressure.
The shift is also showing up in hiring. Infosys added 5,043 employees in Q3, raising its headcount even as peers like TCS are cutting staff while aggressively reskilling for AI. Infosys continues to hire, but it is also retooling its workforce around AI, data, and agent-driven delivery rather than relying solely on manual effort.
Infosys raised its FY26 revenue growth guidance to 3.0-3.5% in constant currency, compared with the 2-3% range it had given in Q2. “We believe we are uniquely positioned to capture market share across these value pools and emerge as the leading AI value creator for global enterprises,” Parekh said.
The post Infosys Runs 4,600 AI Projects, 500 Agents Across Its Enterprise Clients appeared first on Analytics India Magazine.
