AI document extraction with human review

We build AI extraction for purchase invoices, freight invoices, bills of entry, purchase orders and bank statements. An LLM reads the PDF directly, with a prompt per document type. 88–95% of documents need zero correction, and every one still waits for a person to approve it before anything is posted to the books.

Who it's for: Trading, import-export and manufacturing businesses whose accounts team spends its day typing GST invoices, customs paperwork and bank statements into Tally.

How it works

1. 01 A PDF arrives, uploaded in the browser or forwarded from a phone.
2. 02 An AI model reads the PDF directly, with instructions for each type of document.
3. 03 A person checks the extracted fields on screen and approves them.
4. 04 Approved documents post into Tally as vouchers. Nothing is posted without approval.

The numbers

Purchase order, zero-correction rate

95%

Purchase invoice, zero-correction rate

94%

Freight invoice, zero-correction rate

91%

Bank statement, zero-correction rate

89%

Bill of entry, zero-correction rate

88%

Documents posted without human approval

None

What to require from any vendor for this: 2 lines on the ERP requirements checklist, each with the test to ask for in a demo.

Questions buyers ask

How accurate is it?

88–95% of documents need no correction, depending on the type: purchase orders highest at 95%, bills of entry lowest at 88%.

Which documents can it read?

Purchase invoices, freight invoices, bills of entry, purchase orders and bank statements.

Does anything get posted on its own?

No. Every document waits for a person to approve it before it reaches the books.