%201.webp)
LoanPro has introduced an AI-based summary interface for loan servicing agents, saying the tool has cut customer call times by up to 15%.
The system was developed with AllCloud, Amazon Web Services and Anthropic's Claude model.
Loan servicing agents often have to piece together a borrower's account history manually during live calls, moving between payment records, fee logs and prior notes. LoanPro says the new interface replaces much of that work by turning account data into a plain-language summary that agents can query directly while speaking with borrowers.
The tool lets agents "talk to the loan," drawing on a borrower's payment history, missed payments, returned fees and earlier interactions to provide immediate answers. It also records each action an agent takes during a call, creating an audit trail lenders can use for compliance reviews.
That record-keeping function addresses a common problem in consumer lending operations: firms must document not only customer conversations but also the steps staff took when handling an account. In regulated markets, those records can become important if a servicing decision is later challenged or reviewed.
LoanPro also says it has seen a decline in internal support ticket volumes since deploying the system, attributing the change to agents resolving more issues directly through the interface instead of passing cases to other teams.
Call handling
The software reflects a broader push by lenders and their technology suppliers to apply generative AI to front-line servicing work rather than limit it to internal experimentation. Here, the focus is on helping agents interpret fragmented servicing data in real time during customer interactions.
For loan servicers, call efficiency can affect staffing costs, wait times and borrower satisfaction. Agents handling distressed or complex accounts often need to explain missed payments, charges or earlier account actions quickly, and delays in finding information can prolong already difficult conversations.
"Agents shouldn't have to hunt for answers while a borrower is waiting on the line," said Colin Terry, Chief Product Officer, LoanPro.
"This gives them the full picture instantly, so the conversation can focus on solving the borrower's problem instead of finding the information to solve it," Terry said.
LoanPro did not provide an overall number of calls processed through the system, but said the reduction in handling times applied to agents using the interface. Lower support ticket volumes also suggest an operational effect beyond shorter calls, with fewer issues escalated for follow-up.
Partner roles
AllCloud worked with LoanPro on the design and deployment of the interface through the AWS Generative AI Innovation Centre, a program that links AWS customers with implementation partners for generative AI projects. Anthropic's Claude model interprets and presents the underlying account information in readable form.
The project focused on making servicing data usable in the middle of a borrower call. Instead of requiring agents to search across different systems, the interface presents contextual responses drawn from the borrower's account history.
That approach reflects a wider challenge in financial services technology. Lenders often hold extensive operational data, but much of it remains difficult to use quickly because it is spread across separate logs, workflow tools and historical records.
"Lenders are sitting on enormous amounts of data that's hard to act on in the moment," said Peter Nebel, CTO, AllCloud.
"We built this with LoanPro to close that gap - giving agents a real-time, conversational view into the loan itself, right when a borrower needs an answer," Nebel said.
Operational pressure
Loan servicing faces growing pressure to improve both efficiency and documentation as customer expectations rise and compliance scrutiny remains high. Borrowers increasingly expect immediate answers, while lenders need systems that show how staff reached decisions and what actions were taken on an account.
Tools that summarise records or support question-and-answer workflows are emerging as one way to address that tension. Their value, however, often depends less on the model itself than on whether the software can access complete account histories and fit into existing servicing processes without adding more layers of review.
LoanPro says its interface is designed to do that by combining borrower context and action logging in a single workflow. The result, it says, has been shorter calls, fewer internal escalations and a recorded history of every agent interaction.