Infosys is helping clients across industries achieve manpower savings of 5% to as high as 35% with the adoption of its poly-AI, or hybrid artificial intelligence, architecture to integrate AI technologies, a senior executive said.
“It (manpower savings) varies depending upon the different processes and different industries and how much information you get,” said Balakrishna DR, executive VP and global services head for AI and industry verticals at Infosys. “But (overall) anywhere from 5% to 30% or even 35% could be achieved depending upon a business process,” he told ET.
The Bengaluru-based IT giant has created a flexible architecture ecosystem that helps businesses easily migrate AI solutions across models. This has helped its clients save on manpower, by carrying out the same work with fewer workers than before.
“That is what we call poly-AI or hybrid AI architecture…We believe there will be newer and newer models and techniques emerging,” he said. “So, you are better off as an enterprise, not tying to one particular tool or a model.”
The company has moved from initially talking about assistance or copilot — for helping people do their job better by providing additional inputs — to autonomous agents capable of independent reasoning, planning, execution and actions and completing end-to-end business operations, similar to humans, he said.
“It is a digital twin for what people were doing. Of course, you need to have some human in the loop, reviewing the actions,” he said.
Reusable agents
Amid slowing discretionary technology spending, tech clients are demanding more outcome for a lesser price. This has muted revenue growth as well as dented margins for software service providers like Infosys and larger rival Tata Consultancy Services.
The companies are now focusing on building generative AI-backed solutions beyond call centres, personal assistance, advisory, marketing campaigns and content generation.
Infosys, India’s second largest tech outsourcing firm, is working on creating an agentic AI foundry and reusable agents, with 200-plus such agents focusing on enterprise-wide solutions.
Meanwhile, in Infosys’ recent survey with 1,500 senior executives, 95% reported negative consequences of enterprise AI usage in their company in the past two years. As much as 78% viewed ‘responsible AI’ as a critical enabler for enterprise AI success.
According to Balakrishna, while data and talent continue to be a challenge, the wait for the software services industry to meet the current skill and talent mismatch will not be too long. He believes companies like Infosys are focusing on learnability and AI skilling and building AI builders, AI masters and AI experts of existing as well as new employees.
“Currently, I think close to about 90% of employees have been skilled on AI,” Balakrishna said, adding that the company is building next grade AI builders who are able to write, integrate solutions and build AI use cases.
“Then we have the AI experts where employees have deep capabilities, even including creating a foundational model from scratch, model tuning, model optimisation, etc. We even have AI masters with core, deep expertise,” he said.
Currently, sectors leading in agentic AI adoption are financial services, telecommunications, manufacturing, consumer packaged goods, logistics and services industries.
However, Balakrishna said, “The best benefit that you would see is in areas where you have processes that are labour-intensive, which run across long durations.”
For a manufacturing client, Balakrishna said, Infosys used its AI solutions on direct and indirect procurement functions to look across sourcing, supplier management and contract lifecycle management. “We are talking about savings of tens and hundreds of millions of dollars annually,” he said.
Similarly, Infosys is helping improve the audit function of a professional services company where it has seen a 40% higher turnaround time and cost efficiencies of about 20%. For a telecommunications firm, Infosys solutions have helped achieve 22% revenue growth for a particular product line and also enabled cross-sell and up-sell opportunities for its call centre agents, he said.
Independence Day 2025
Op Sindoor to water cutoff: PM Modi slams Pak in I-Day speech
GST reforms by Diwali to cut daily-use taxes: PM Modi
Terrorism, tech, more: PM's I-Day speech highlights
The Bengaluru-based IT giant has created a flexible architecture ecosystem that helps businesses easily migrate AI solutions across models. This has helped its clients save on manpower, by carrying out the same work with fewer workers than before.
“That is what we call poly-AI or hybrid AI architecture…We believe there will be newer and newer models and techniques emerging,” he said. “So, you are better off as an enterprise, not tying to one particular tool or a model.”
The company has moved from initially talking about assistance or copilot — for helping people do their job better by providing additional inputs — to autonomous agents capable of independent reasoning, planning, execution and actions and completing end-to-end business operations, similar to humans, he said.
“It is a digital twin for what people were doing. Of course, you need to have some human in the loop, reviewing the actions,” he said.
Reusable agents
Amid slowing discretionary technology spending, tech clients are demanding more outcome for a lesser price. This has muted revenue growth as well as dented margins for software service providers like Infosys and larger rival Tata Consultancy Services.
The companies are now focusing on building generative AI-backed solutions beyond call centres, personal assistance, advisory, marketing campaigns and content generation.
Infosys, India’s second largest tech outsourcing firm, is working on creating an agentic AI foundry and reusable agents, with 200-plus such agents focusing on enterprise-wide solutions.
Meanwhile, in Infosys’ recent survey with 1,500 senior executives, 95% reported negative consequences of enterprise AI usage in their company in the past two years. As much as 78% viewed ‘responsible AI’ as a critical enabler for enterprise AI success.
According to Balakrishna, while data and talent continue to be a challenge, the wait for the software services industry to meet the current skill and talent mismatch will not be too long. He believes companies like Infosys are focusing on learnability and AI skilling and building AI builders, AI masters and AI experts of existing as well as new employees.
“Currently, I think close to about 90% of employees have been skilled on AI,” Balakrishna said, adding that the company is building next grade AI builders who are able to write, integrate solutions and build AI use cases.
“Then we have the AI experts where employees have deep capabilities, even including creating a foundational model from scratch, model tuning, model optimisation, etc. We even have AI masters with core, deep expertise,” he said.
Currently, sectors leading in agentic AI adoption are financial services, telecommunications, manufacturing, consumer packaged goods, logistics and services industries.
However, Balakrishna said, “The best benefit that you would see is in areas where you have processes that are labour-intensive, which run across long durations.”
For a manufacturing client, Balakrishna said, Infosys used its AI solutions on direct and indirect procurement functions to look across sourcing, supplier management and contract lifecycle management. “We are talking about savings of tens and hundreds of millions of dollars annually,” he said.
Similarly, Infosys is helping improve the audit function of a professional services company where it has seen a 40% higher turnaround time and cost efficiencies of about 20%. For a telecommunications firm, Infosys solutions have helped achieve 22% revenue growth for a particular product line and also enabled cross-sell and up-sell opportunities for its call centre agents, he said.
