A Guth Labs publication ↗
Back to News

LoanPro details its Bedrock-powered loan servicing assistant

Cream cards inside a green glass archive cylinder with gold pathways through an access opening.
AI-generated editorial illustration

LoanPro and AWS published a November 17 case study describing a production AI assistant that summarizes borrower accounts and recommends next steps for loan servicing staff.

Built with AllCloud and the AWS Generative AI Innovation Center, the system uses Amazon Bedrock AgentCore and the Strands Agents SDK. A coordinating agent directs work to specialists handling account information and workflow recommendations, according to the technical account.

The assistant retrieves information across eight API endpoints covering around 1,400 fields per account. LoanPro says its permissions and compliance checks apply to each request, with AI actions recorded in an audit trail and attributed to a responsible person.

The case study reports call-time reductions of up to 15%, citing implementation partner AllCloud. It also describes an evaluation using 150 expert-labeled accounts and a 99% accuracy target. These are vendor-reported results, not independent validation or guarantees for other deployments.

The example illustrates how service teams can combine AI-generated account context with existing access controls. It documents an operating deployment, not a newly announced general-purpose model or an autonomous lending decision system.

Sources and citations

Revision history

Published .