KeyBank Taps Conversational AI to Cut Call Center Costs

At KeyCorp, the parent of KeyBank, CEO Christopher Gorman this week highlighted how AI is already delivering per-interaction cost advantages during the company’s fourth-quarter 2025 earnings call. AI-handled calls cost roughly $0.25 each versus $9 for human-handled interactions.

KeyBank has also been steadily increasing its tech spend to fuel these efficiencies, rising from an $800 million to $900 million run rate in prior years to about $1 billion in technology and operations investment, including enhanced digital and AI capabilities. Gorman said that while it’s early to quantify broad AI-driven efficiencies, the bank has found roughly $100 million in annual savings through continuous improvement efforts, which will help fund ongoing digital transformation.

Conversational and generative AI can lower operational costs when deployed thoughtfully. Traditional call centers, long plagued by inefficient interactive voice response menus and staff bottlenecks, are being modernized with AI that can interpret customer intent and route or resolve inquiries faster than legacy systems could.

How KakaoBank and Lloyds Are Scaling Conversational AI

International banks also offer a look at how conversational AI is being scaled beyond pilots. In South Korea, KakaoBank has deployed conversational AI built on Microsoft Azure OpenAI to serve as a primary interface for customer inquiries. The system allows customers to interact naturally about account activity, transactions and financial services inside KakaoBank’s mobile app, reducing reliance on live agents while maintaining fast response times.

KakaoBank’s approach reflects a digital-native strategy where AI is embedded directly into the core customer experience rather than layered on top as a support tool. By resolving routine questions conversationally, the bank reduces service costs while keeping customers inside its digital ecosystem, where engagement tends to be higher and servicing costs lower.

In the United Kingdom, Lloyds Bank has taken a different but complementary path with its generative AI platform called Athena. Lloyds is using the tool to assist customers and employees, automating responses to common queries and helping staff access information more quickly. The bank has positioned Athena to improve service quality and productivity while easing pressure on contact centers.

Rather than replacing human agents, Lloyds has emphasized AI as an augmentation layer that speeds resolution and reduces manual effort. That model aligns with how several large banks are deploying conversational AI. They’re trimming costs through automation while preserving human oversight where judgment and empathy matter. Lloyds also plans to add an AI-powered financial assistant app this year that will provide personalized financial coaching.

Agentic AI and Workflow Automation Expand Savings Across Banking

Beyond frontline customer service, banks are deploying more advanced agentic AI systems that perform tasks across varied functions, from customer support to internal operations.

Wells Fargo, for example, has expanded its partnership with Google Cloud to deploy AI agents at scale. The tools automate tasks, including balance inquiries and debit card replacements, freeing human staff to focus on higher-value work and strategic customer relationships. The initiative also covers internal workflows, such as complex trade inquiries and document review, with AI agents helping employees find insights faster and complete tasks more efficiently, contributing to operational savings and improved agility.

These agentic systems go beyond reactive chatbots by synthesizing information across internal data sources and offering 24/7 personalized interactions, a capability that has the potential to reduce overall cost structures and time to resolution for many service functions.

While call center AI delivers immediate operational savings, banks are expanding the scope of these technologies to enhance customer engagement and retention. The PYMNTS Intelligence report “Beyond the Bot: Why Embedded Conversational AI Is Banking’s Next Strategic Advantage” found that 72% of bank customers say personalization influences where they choose to bank, and conversational AI is now being positioned as more than a cost-saving tool. It’s a value generator that deepens engagement and captures valuable insights.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.