A U.S. consumer lender that offers personal loans runs applicant documents through super.AI, and its staff review the results in their own portal.
When someone applies for a personal loan with this lender, they upload supporting documents, like an insurance card or a vehicle registration. Before anyone can review the application, each of those documents has to be identified and its details pulled out.
That sorting and reading is repetitive, and it pulls staff away from the parts of the job that need a person.
The lender was clear with us about the goal. They wanted their people back on the work that needs them, and they weren't trying to replace anyone.
The workflow runs on our platform. An applicant uploads a document, super.AI classifies it and pulls out the data, and the results go back to the lender's own review portal by webhook.
The lender's staff still review the data there. They work in the portal they already use, and the classification and extraction are done before anything reaches them.
In fall 2026, paper documents started running at production volume too, including vehicle titles.
The lender has been a customer since 2024, and the production workflows went live on April 16, 2026.
From April 16 to October 20, 2026, we counted 565,538 workflow runs and 6,260,069 automated task executions in production. A task execution is one automated step inside a run, like classifying a document or extracting its fields, and each run averaged about 11 of them.
Volume has grown. Monthly runs went from 78,735 in May 2026 to 109,411 in September 2026, which is 39% growth. The busiest week, starting August 17, 2026, had 27,908 runs, and the recent pace is about 3,750 runs a day.
If your applicants upload supporting documents and your team spends its time sorting and reading them before the real review starts, that's what this lender's workflow handles. Here's more on document workflow automation.
