Avant has rolled out an AI Virtual Agent across its inbound customer service operations for Avant Card. The system has been in use in the card servicing channel since January.
Developed with Replicant, the virtual agent is now handling a large share of incoming customer calls. About 62% of calls are completed without transfer to a human agent, and customers have given the system an average score of 4.6 out of 5 in post-call surveys.
The deployment marks a further use of artificial intelligence in day-to-day customer service operations in consumer finance, where lenders and card providers are trying to manage high call volumes while keeping specialist staff available for more complex cases. The system is designed to support routine interactions and gather structured information from customer conversations.
Operational shift
The virtual agent was built around five principles: round-the-clock availability, shifting routine tasks away from human specialists, structured data capture, the ability to absorb spikes in demand, and consistent, compliant communication with customers.
Data collected through those automated interactions gives Avant's teams clearer signals on what customers need and where service and product issues may require attention. The company said this kind of feedback loop is harder to achieve when calls are handled only by human agents.
For customer service teams, the change also alters the mix of work. Repetitive, high-volume requests can now be handled earlier in the process, allowing specialists to spend more time on cases that require judgment and empathy.
The approach reflects a broader pattern in financial services, where firms are using AI tools not only to ease pressure on contact centres but also to sort customer requests more effectively. In card servicing, incoming calls often include a mix of straightforward administrative enquiries and more sensitive issues that need human intervention.
Margaret Hermes, Chief Operating Officer at Avant, described the system as part of a wider operational design.
"Our goal is to build a more responsive operation where AI handles complexity at scale, our talented team of customer service human specialists focuses on the work that demands judgment and empathy, and the entire system generates better intelligence about customer needs over time," Hermes said.
Business scale
Avant offers credit cards, personal loans and a personal financial management tool to middle-income consumers in the US. It has connected more than 4 million unique customers to more than USD $13.2 billion in personal loans and more than 2.9 million credit cards since launching its lending and card products.
Those figures indicate the scale of the servicing operation behind the business. Large customer bases in consumer credit tend to generate substantial inbound traffic, particularly around account servicing, payments, balances and card-related support, making automation a practical area for investment.
Replicant, which developed the virtual agent with Avant, focuses on conversational AI for customer service operations. Financial terms of the partnership and the absolute volume of calls handled by the system were not disclosed.
Avant plans to expand its use of the platform in customer support. For now, the clearest measure of the rollout is the share of calls that no longer need to be passed to a live agent: about 62%.