AI in action: How enterprises are scaling AI for real business impact

To capitalize on the enormous potential of artificial intelligence (AI) enterprises need systems purpose-built for industry-specific workflows. Strong domain expertise, solid data foundations and innovative AI capabilities will help organizations accelerate business outcomes and outperform their competitors.

Enterprise technology leaders discussed these issues and more while sharing real-world examples during EXL’s recent virtual event, “AI in Action: Driving the Shift to Scalable AI.”

“The key to driving real impact lies in seamlessly integrating data and AI into the way businesses work,” said Rohit Kapoor, chairman and CEO, EXL. “It’s not just about implementing technology. It’s about orchestrating data, digital solutions and human intelligence to optimize decision-making and unlock new opportunities.”

The year of agentic AI

Agentic AI holds the key to unlocking these opportunities. With autonomous, self-regulating AI agents, enterprises can create automated workflows that adapt to real-world business complexity and augment their human experts to boost efficiency, accuracy and innovation.

Kevin Ichhpurani, president of global partner ecosystem with Google Cloud, shared an example of a mutual client and how EXL and Google have helped them with customer service agents. The agents understand the consumer’s intent when they call, make      educated decisions through complex reasoning and then take action, such as initiating a product exchange or ordering a replacement unit.

“We see [2025] as the year of delivering agentic experiences for clients, where we automate complete end-to-end business processes,” Ichhpurani said.

To achieve this goal, EXL last month launched its agentic AI platform, EXLerate.AI. It orchestrates AI models alongside human expertise and analytics “to help businesses harness AI without getting slowed down by technical complexities,” Kapoor said.

The virtual event also featured demos of EXL Code Harbor, a generative AI-powered code migration tool, and EXL’s Insurance Large Language Model (LLM), a purpose-built solution to the industry’s challenges around claims adjudication and underwriting.

The Insurance LLM is trained on 12 years’ worth of casualty insurance claims and medical records and is powered by EXL’s domain expertise. Built on NVIDIA’s AI stack, the LLM delivers 30% greater accuracy and 30% lower costs than general-purpose models.

“Insurance LLM assists claim adjusters to be more productive and accurate in a shorter time period,” said John Fanelli, vice president, enterprise software, NVIDIA. “It also delivers the best outcomes for both the insurers and the insured. Insurance LLM is a fantastic example of what we call an agentic AI system.”

AI in the wild

In two event panels, enterprise AI practitioners shared the trends they’re seeing this year and how they’re adapting. The first conversation focused on the evolving symbiosis between data and AI.

“There used to be a discussion about how much data you have,” said Sidd Kuckreja, CTO with TruStage. “Now it’s about the quality of data as you think of the regulatory landscape, bias mitigation, privacy and ethical considerations.”

Randy Huang, vice president and chief data scientist for U.S. business with Prudential, emphasized the importance of security and governance, because more people are using AI platforms to access and use sensitive data.

“The focus on data is really changing based on how the data is generated and how the data is used,” Huang said.

And Preetha Sekharan, vice president of Unum’s digital incubator, noted that while data can fuel AI innovation, the inverse is also true.

“What is really interesting with genAI and newer technologies is how AI can accelerate how you generate, how you transform, how you understand data,” Sekharan said. “That is really a fascinating twist in how we think about data.”

The second panel focused on how AI helps enterprises maintain a competitive advantage. NRG Energy uses AI to conduct ongoing scenario modeling, analyzing weather and forecasting its effects on customer demand and energy prices.

“There’s a lot of data points, and … there’s a really good opportunity to use that to do better prediction,” said Dak Liyanearachchi, chief data and technology officer.

Sarthak Pattanaik, head of the artificial intelligence hub at BNY, discussed the bank’s internal platform, which enables employees to build AI-powered systems while ensuring security, privacy, fairness, ethical usage, accountability, and transparency.

“It democratizes access to AI in a responsible fashion, so it helps innovation at scale,” Pattanaik said.

And Dr. Ashish Atreja, professor of medicine at University of California – Davis Health, spoke about AI improving patient access to care.

“The biggest value for patients that’s going to happen is moving healthcare fundamentally from one-to-one care, where you have to be with a physician and a patient in the same space and time, to one-to-many care — how you can automate digital care pathways through digital avatars, through digital apps, through digital therapeutics,” Atreja said.

A fundamental transformation

Simply adopting AI is no longer enough. As industry leaders emphasized during EXL’s event, success requires integrating AI with high-quality data and deep domain expertise — while rethinking and optimizing business processes.

“AI is not just a technological shift,” Kapoor said. “It’s a fundamental business transformation.”

To learn more about what agentic AI and EXL can do for your business, visit here.