Chime Financial improves member experience and saves 250,000+ hours annually with AI-powered call summaries using Amazon Bedrock

Chime provides banking services to millions of members, and it prioritizes delivering a consistently excellent customer experience. But the company’s customer service agents had to juggle multiple tasks: listening to customers, taking notes, and doing post-call summaries. Chime wanted to streamline that work and more efficiently analyze the call data that it uses to further improve the customer experience. “Our ability to be efficient means that we can provide our customers with a better experience,” says Dennis Yang, generative AI product team lead at Chime.

As part of a hackathon in May 2023, the company built an AI-powered call summarization tool to address the time-consuming manual process of documenting customer interactions. After seeing the potential to empower agents and improve customer experiences, Chime chose to develop the proof of concept into a production-ready solution. The company was already using AWS for internal AI initiatives and was impressed by the rapid innovation of AWS in the generative AI space. It valued the enterprise-grade security of Amazon Bedrock and the ability to keep sensitive data within its existing AWS environment. Combined with ease of implementation and access to a wide range of industry-leading foundation models, Chime gained the security, innovation, and scalability it needed to bring AI into production using Amazon Bedrock.

The company tested multiple AI models in Amazon Bedrock before choosing Anthropic’s state-of-the-art model family, Claude. The company implemented Claude by Anthropic in Amazon Bedrock. Claude’s advanced reasoning capabilities and ability to process long-context call transcripts made it ideal for generating accurate, comprehensive summaries that agents could rely on in cases of repeat customer interaction. Chime tested the application throughout 2023 and in February 2024 launched it in production. Now, the tool is available to agents on all calls.

Chime summarizes each customer call in a matter of seconds by using Claude models in Amazon Bedrock. The company has cut the average agent handling time for a call by 18 seconds. That adds up to more than 250,000 hours saved per year by using AI-powered call summarizations (equivalent to approximately $700,000 in annual efficiency gains). Agents also use the tool to reduce manual note-taking and refer to summaries when speaking with previous callers.

Chime agents reported being impressed with the clarity of the new AI-generated notes, which they are using to provide context when talking to repeat callers. Many agents reported that the tool significantly enhanced their job performance, praising its ability to streamline workflows and improve their overall effectiveness. “Gone are the days of agents furiously typing notes while customers are on hold,” says Alexa Schaeffer, strategic initiatives lead for the operations and member experience team at Chime. “Now our agents, empowered by AI assistance built on Amazon Bedrock, can focus on customers, empathizing with their issues and finding solutions as quickly as possible.”

Internally, Chime is using the new summaries to expand its data analysis for improving services. “By using Amazon Bedrock, we’ve unlocked a treasure trove of actionable insights that were previously unavailable,” says Schaeffer. “It’s transforming how we serve our customers.” By harnessing its AI-powered solution, Chime has not only enhanced its understanding of its customers but has also increased its service net promoter score—a measure of customer satisfaction—by five points.

“We’re still finding applications for call summarization that we hadn’t envisioned,” says Schaeffer. “We’re uncovering innovative applications that push the boundaries of current practices, bolstering our customer service and efficiency.” In one example, one of Chime’s partners required Chime to send a summary of every call in a short time frame. By using the AI summarization tool built on Amazon Bedrock, an entire team could focus on customer service instead of note-taking. “This application continues to scale in terms of the internal solutions it can provide,” says Schaeffer.

Chime also automated call transcriptions by using Amazon Transcribe, an AI service that converts speech to text. Chime redacts any identifying elements from each transcript before sending it to Claude models for summarization. And the company is using Amazon Bedrock Guardrails, a feature that provides safeguards customized to application requirements and responsible AI policies, to help detect and redact sensitive information without significant overhead. “When we needed to redact sensitive information from a call transcript before sending it to the AI tool, we used Amazon Bedrock Guardrails,” says Mukund Sarma, product security lead at Chime. “We didn’t have to figure out our own process or dedicate engineering resources to building custom safeguards. We got enterprise-grade protection out of the box using Amazon Bedrock Guardrails, so we could move faster while maintaining our high security standards.”

With more accurate transcriptions, Chime is better able to investigate anomalies in customer calls. Investigators can simply read call transcripts to understand more quickly what’s happening. “Our application uses a transcription service and an AI service, and we can use AWS for both,” says Schaeffer. “For security purposes as well as ease of use, we can keep it all in one stack on AWS.”

Chime has shown how seamlessly a company can use AWS to implement and scale generative AI to improve its customer experience. Building on the success of AI-powered call summarization, Chime is exploring AI-driven improvements across its operations, including chat transcripts and proactive customer insights.

“AWS is our go-to provider for tackling a range of challenges,” says Yang. “We used Amazon Bedrock to go quickly from hackathon concept to production and scale AI across thousands of customer calls, delivering immediate business value with measurable cost savings and improved customer satisfaction. We’re using AWS to propel our innovation in generative AI, and we’re not just adapting to this technology—we’re setting the standard for how AI can enhance customer experience in financial services.”