PaySense automated data capture from loan applicants' bank statements and IDs
PaySense, a Mumbai consumer lending startup acquired by PayU, used Docsumo's APIs to read statements, cheques and IDs as applications arrived, instead of underwriters keying them in.
A personal loan application with two files: a six-month bank statement as a PDF, and a driving licence photographed on both sides. Docsumo read the statement's opening balance of ₹21,480.75, credits of ₹2,70,350.00, debits of ₹2,58,912.40 and closing balance of ₹32,918.35. The running balances added up and PaySense's rules passed, so the file was ready for review, and approved data went to the underwriting system by API.
Personal loans for India, in an app.
PaySense was a Mumbai-based consumer lending startup that offered personal loans to people across India through its app. In 2020 PayU bought a controlling stake, and PaySense became part of PayU Finance, PayU's lending arm.
- Personal loansB2CUnsecured loans for consumers, applied for in the PaySense app.
- Loan originationPayU FinanceBorrowers sourced and taken through KYC for PayU's lending arm.
Every application waited on an underwriter.
Before Docsumo, underwriters read every statement and ID by hand.
- Bank statements
- Cheques
- Identity documents
- A different layout for every bank
- Several kinds of ID to read
- Every field keyed in by hand
- Every check done by eye
- Loan decision
- Onboarded borrower
Built around their loan applications.
Docsumo is an intelligent document processing (IDP) platform.
- Many banks
Every bank's layout, one format
Statements from different Indian banks came out in the same fields.
- By API
Read by API, not keyed in
Statements, cheques and each type of ID were read by API as applications arrived.
- Rules
Checked by rules, not by eye
A rule-based engine validated each field and corrected errors in context.
What changed at PaySense.
| Before Docsumo | With Docsumo | |
|---|---|---|
| Statements and IDs | Read and keyed in by underwriters | Read by API as applications arrived |
| Data checks | By eye, one application at a time | Validated and corrected by a rule-based engine |
| The underwriters' time | Spent on data entry | Spent on loan decisions, with the data already read |
See how bank statement extraction and income verification work.
See it on your documents.
Bring three of your hardest bank statements and ID documents. We will run them on the call and show every field and every check.