Balkrishan Kalra on accounts payable
From From Pilots to Production: Who Owns the AI Workflow? CEOs of EXL, Genpact, FirstSource & Microsoft · · Zinnov
“Within finance operations — take accounts payable: we were in the human‑processed, human‑validated zone for nearly 25–30 years. Just last year we moved the entire accounts payable solution to machine‑processed, human‑validated 'responsible AI' lab; that's what we call agentic AI.”
On , Balkrishan Kalra, Chief Executive Officer, President & Director at GENPACT LTD, spoke about accounts payable during From Pilots to Production: Who Owns the AI Workflow? CEOs of EXL, Genpact, FirstSource & Microsoft on Zinnov.
About Balkrishan Kalra
On April 7, 2026, Balkrishan Kalra participated in a panel discussion titled "From Pilots to Production: Who Owns the AI Workflow?" alongside CEOs from EXL, FirstSource, and Microsoft. During the conversation, Kalra noted that less than 10% of AI pilots reach enterprise-scale production and discussed the challenges of moving from pilots to production in industries such as retail, CPG, pharma, and BFSI. He stated that applying agentic AI to an end-to-end process requires upfront consideration of process maturity, and that in many large companies, technical architecture is "a little bit in mess" due to acquisitions without integration. Kalra said that "agentic or AI is not a magic pixie dust that you just sprinkle and magic will come out." Kalra also provided an example from Genpact's own operations, stating that within finance operations, the accounts payable workflow had been "human processed, human validated" for 25-30 years, but that in the past year the company moved the entire solution to a "machine processed human validated responsible AI lab," which he described as agentic AI. He emphasized that an organization's "moat" has shifted to its people, who understand the brand and culture, and that technology and platform can be agnostic. He added that over the past two years, conversations have moved from guard rails and safety to full production deployments, and that a "chasm" is growing between leaders and laggards in AI adoption.