EXL’s agentic AI and Google Cloud solution saves global financial services leader $12M through call containment
Client Overview
A global financial services leader managing over 100 million customer service calls annually sought to improve efficiency and reduce operational costs by enhancing their interactive voice response (IVR) system.
Despite significant investment in conversational AI, containment had plateaued at 70%, with roughly 30 million calls per year still requiring live agent intervention. These high-volume, low-complexity interactions presented a clear opportunity for automation.
Challenge
The client’s IVR struggled to accurately interpret varied customer languages, resulting in frequent misroutes and escalations to human agents. These low-complexity interactions were ideal for automation, yet the system couldn’t reliably identify intent. At an average cost of $4 per agent-assisted call, the financial impact was significant. To unlock further containment, the client needed a more adaptive, accurate solution that could scale across millions of interactions—without compromising compliance or customer experience.
Solution
EXL partnered with the client to design and deploy a next-generation IVR experience powered by Google Cloud’s Gemini platform. Built on an agentic AI architecture, the solution used Google’s generative AI to drive more flexible, accurate, and scalable selfservice interactions.
Rather than relying on a single model, the system introduced a coordinated set of AI agents—each handling a specialized function. These included intent recognition, subintent detection, and real-time response generation using Gemini Flash, along with a compliance guardrail to ensure regulatory alignment.
The modular design enabled faster decision-making, improved containment, and enterprise-grade transparency. Rolled out first for two high-volume use cases, the system was deployed as a lightweight overlay to the existing IVR, with minimal disruption and immediate impact.
Results
The agentic AI IVR delivered a 10% uplift in containment for the targeted use cases, successfully automating approximately 3 million additional calls annually from live agents. With each avoided call saving an estimated $4, this equated to $12 million in projected cost savings for just the initial implementation.
With full rollout across the 30 million applicable calls, the solution is expected to deliver over $45 million in annualized savings. The improved containment rate also reduced wait times, improved consistency of service, and allowed live agents to focus on more complex customer needs.
Next Steps
Based on the success of the pilot, the client has identified six additional use cases for expansion and is now exploring a full transition of its legacy IVR system to the new agentic AI model. Future iterations will continue to leverage Google Cloud’s generative AI capabilities and EXL’s delivery framework to drive further efficiency, agility, and customer satisfaction
Why EXL and Google Cloud
- EXL brought deep industry knowledge, AI operational design expertise, and proven success delivering scalable automation within regulated industries.
- Google Cloud offered the advanced generative AI infrastructure, including Gemini Flash, to enable secure, performant, and compliant conversational experiences.
Together, EXL and Google Cloud delivered a future-ready solution that not only reduced cost and complexity, but also redefined what intelligent self-service can achieve in financial services.