Top business process management (BPM) companies such as Genpact, EXL and WNS are making strategic shifts in the way they operate amid artificial intelligence-led automation. These companies are building and deploying AI agents that enable end-to-end automation of workflows for greater efficiency.
While Genpact is moving all its existing work into the agentic architecture, EXL feels that 60-70% of jobs in traditional areas like customer service and customer transaction processing may go away. WNS said there is a move away from the labour-, time- and material-based commercial models towards value-based pricing, as customers focus on profitability, efficiency and outcomes.
The agentic AI use cases range across processes in customer service, order management in retail and ecommerce, accounts payable, insurance claims processing and underwriting, supply chain, back-office and middle-office operations, and more, executives said, adding that BPMs’ expertise in the operations of different domains gives them an advantage in driving AI adoption among enterprises.
Also Read: IT, BPM firms log rise in requests for AI chatbot pilots
A human in the loop is still essential to handle complexities, though the need for this will also go down with time, executives said.
At Genpact, the strategic decision to work on agentic architectures was made in mid-2024. It has an account-by-account plan to do this and the entire staff is being trained to adapt, chief technology and innovation officer Sanjeev Vohra told ET.
Invoice verification, for instance, can be automated by 80%. Internally, Genpact has a family of 20 AI agents that help in areas like IT helpdesk functions, which saw 50% more efficiency over the past year, he said.
Large language models are the “brains” of agentic solutions alongside access to enterprise data and platforms across silos, executives explained. Agentic AI can reason rather than sticking to set procedures, has short- and long-term memory to understand the context and needs of specific tasks, and the ability to talk to other agents to fulfil processes and drive better economics.
For instance, it can enable more proactive and personalised customer service and more effective upselling, pushing revenue growth and saving human agents’ time for more complex issues, they said, adding that routine and repeatable tasks can and should be done by autonomous agents.
Also Read: Tech to change job landscape in BPM industry over 3 years: Nasscom-Indeed report
“Jobs will go in the traditional areas of customer service, customer transaction processing, etc. Today, it cannot go to zero, but it will go down by 60-70%,” said Gaurav Iyer, senior vice president and global leader for AI solutions and digital strategy at EXL. “But I would also say the jobs in data and AI are going to grow by 100-120%.”
Reskilling in this direction is a pivot that the sector needs to make, he added.
WNS is “productising its services” and moving to an "AI+HI" (artificial intelligence+human intelligence) model, said Gautam Singh, global head of the firm's data, analytics and AI business unit, WNS Analytics.
The cost of AI going down presents a tailwind for the sector as more businesses will adopt AI and require BPM services for last-mile integration, executives said. Customers are increasingly expecting AI to be a standard part of BPM services rather than a premium value-add. They also want the benefits of the falling costs of AI to be passed on to them.
Also Read: ETtech in-depth: how AI is transforming India's BPM industry
While Genpact is moving all its existing work into the agentic architecture, EXL feels that 60-70% of jobs in traditional areas like customer service and customer transaction processing may go away. WNS said there is a move away from the labour-, time- and material-based commercial models towards value-based pricing, as customers focus on profitability, efficiency and outcomes.
The agentic AI use cases range across processes in customer service, order management in retail and ecommerce, accounts payable, insurance claims processing and underwriting, supply chain, back-office and middle-office operations, and more, executives said, adding that BPMs’ expertise in the operations of different domains gives them an advantage in driving AI adoption among enterprises.
Also Read: IT, BPM firms log rise in requests for AI chatbot pilots
A human in the loop is still essential to handle complexities, though the need for this will also go down with time, executives said.
At Genpact, the strategic decision to work on agentic architectures was made in mid-2024. It has an account-by-account plan to do this and the entire staff is being trained to adapt, chief technology and innovation officer Sanjeev Vohra told ET.
Invoice verification, for instance, can be automated by 80%. Internally, Genpact has a family of 20 AI agents that help in areas like IT helpdesk functions, which saw 50% more efficiency over the past year, he said.
Large language models are the “brains” of agentic solutions alongside access to enterprise data and platforms across silos, executives explained. Agentic AI can reason rather than sticking to set procedures, has short- and long-term memory to understand the context and needs of specific tasks, and the ability to talk to other agents to fulfil processes and drive better economics.
For instance, it can enable more proactive and personalised customer service and more effective upselling, pushing revenue growth and saving human agents’ time for more complex issues, they said, adding that routine and repeatable tasks can and should be done by autonomous agents.
Also Read: Tech to change job landscape in BPM industry over 3 years: Nasscom-Indeed report
“Jobs will go in the traditional areas of customer service, customer transaction processing, etc. Today, it cannot go to zero, but it will go down by 60-70%,” said Gaurav Iyer, senior vice president and global leader for AI solutions and digital strategy at EXL. “But I would also say the jobs in data and AI are going to grow by 100-120%.”
Reskilling in this direction is a pivot that the sector needs to make, he added.
WNS is “productising its services” and moving to an "AI+HI" (artificial intelligence+human intelligence) model, said Gautam Singh, global head of the firm's data, analytics and AI business unit, WNS Analytics.
The cost of AI going down presents a tailwind for the sector as more businesses will adopt AI and require BPM services for last-mile integration, executives said. Customers are increasingly expecting AI to be a standard part of BPM services rather than a premium value-add. They also want the benefits of the falling costs of AI to be passed on to them.
Also Read: ETtech in-depth: how AI is transforming India's BPM industry