Lending
0 articles and 2 community signals on Lending for people running AI in production, written and curated by Wren with every source linked.
What practitioners are saying
All signals →Harmoney, an Australasian online lender, says an agent that reads bank statements cut loan decisions from 24 hours to 15 minutes, on a Google Cloud customer page
On a Google Cloud customer page, Harmoney says its 'Bank Bot' classifies and reads applicants' bank statements and runs affordability checks without a person, work that used to take two service reps and a credit officer. It reports drop-off down 20% to 30% and a build by one developer and one product manager. The page gives two document volumes: up to 20,000 a month and up to 1,000.
Why it matters For lending operations and compliance leads: the check that once needed three people now runs unattended. Ask who samples its decisions and how often it is wrong; the page gives no error rate, cost or regulator view.
AWS blog: LoanPro says an account-summary assistant cut loan servicing call times by up to 15% across more than 800 staff at early lender customers
A post co-written by AWS and LoanPro, whose software lenders use to service loans, says an assistant that summarises a borrower's account and suggests a next step cut call times by up to 15%; staff had spent 1 to 2.5 minutes a call finding their place. LoanPro also reports nearly 8% fewer internal tickets and 99% accuracy on 150 expert-labelled accounts. No lender is named and no cost is given.
Why it matters For leads of loan servicing and other contact centres where staff open several screens before they can help: time spent orienting per call is the baseline to measure before a pilot. The 15% is an 'up to' figure from the vendor's partner.