After saying its doing incredible work with generative AI in the previous quarter, Infosys has yet again emphasised on its dedication to generative AI at its latest Q2 FY25 results, but shied away from revealing the revenue from generative AI.
On the brighter side, Infosys has finally revealed that its working on small language models for its clients for various applications. “It’s an incredible approach that leverages various open source components along with a narrow set of industry data and Infosys’ proprietary dataset,” said Salil Parekh, CEO and MD.
Some of the popular open source models include Llama, Mistral, Phi, and Gemma, alongside India’s very own small language model, Sarvam 2B.
Meanwhile, in March, Yann LeCun, the chief of Meta AI revealed that he met with an Infosys co-founder who was funding a project based on Llama 2. “He’s funding a project to fine-tune Llama 2, the open source model produced by Meta so that it speaks all 22 official languages in India. It’s very important for people in India,” said LeCun. He did not reveal which co-founder he met.
Parekh added that Infosys believes small language models will provide clients with a powerful tool, allowing them to build business logic on top of it and unlock new potential. “This combination will form the foundation of the small language model, which is being tailored for different industry applications,” he added.
“It’s a differentiated strategy, and although we haven’t shared much yet, we are already having promising discussions with clients. The work has begun, and we’re integrating generative AI deeply across key areas,” he added.
Multi-Agent Framework
Parekh also shared a notable case involving building a multi-agent framework for a client, where the agents handle specific business processes almost entirely on their own. “We are building enterprise generative AI platforms and multi-agent frameworks for clients,” he added, saying that Infosys leadership in this area is making its clients happy.
“This shifts the way the client can approach their operations, allowing them to scale up more efficiently by automating significant parts of their processes,” he added.
According to Parekh, generative AI use cases are not limited to any particular sector; it’s becoming integrated into nearly everything. “Whether it’s a large transformation project, cost efficiency initiative, or productivity improvement, generative AI is playing a role,” he said, while adding that generative AI is driving productivity gains and is part of almost every project that Infosys undertakes.
Another example is in the telecom sector, where Parekh said that Infosys implemented an enterprise generative AI platform. “Currently, 70,000 employees are using this platform to develop their own use cases, whether in knowledge management, customer service, or coding,” he added, saying that these aren’t just proof-of-concept projects but actual, completed projects where clients are already experiencing tangible benefits.
Infosys is probably the only Indian IT company which is talking about building AI agents right now.
Revenue, Freshers, and GCCs
Infosys reported a 2.2 percent quarter-on-quarter rise in net profit for Q2 FY25, amounting to Rs 6,506 crore, up from Rs 6,368 crore, falling short of market expectations. The company’s revenue for the July-September period grew by 4.2 percent compared to the previous quarter, reaching Rs 40,986 crore.
The company raised its full-year revenue growth guidance to 3.75-4.5 percent, building on the unexpected hike to 3-4 percent announced in July. It reported an operating margin of 21.1 percent for the fiscal second quarter, with large deals totaling a TCV (total contract value) of $2.4 billion.
In constant currency terms, revenue grew by 3.3 percent year-over-year. Parekh highlighted that the $2.4 billion in large deals during Q2 underscores its strong market position.
“The growth was broad based with good momentum in financial services. This stems from our strength in industry expertise, market leading capabilities in cloud with Cobalt and generative AI with Topaz, resulting in growing client preference to partner with us”, said Parekh, highlighting the importance of generative AI projects.
Infosys is on track to onboard 15,000 to 20,000 freshers at the group level in FY25, though it did not break down the numbers between current and previous years. All freshers will be onboarded within that range.
Infosys is also working with GCCs in India in industries like telco and life sciences. When clients are setting up their GCCs, Infosys collaborates with them, particularly in the Build-Operate-Transfer (BOT) model.
“We’re actively involved in the building and operating phases, and when they transfer, we continue to support them. Our efforts extend to GCCs in India, where we assist with scaling their operations, including recruitment efforts,” added Parekh.
The post Infosys is Finally Building Small Language Models and Multi-Agent Framework appeared first on AIM.
