SETI · Email document OCR auto-entry

Background

Quotes and receipts from suppliers usually arrive as email attachments, either PDFs or photos. To enter them into the company’s books, someone opens each one and copies the numbers.

Plenty of tools offer an upload screen. But uploading still means a person collecting files and submitting them. The files are already in the inbox, and uploading adds one more step.

The problem

The goal was to remove every step that needs a person.

Better reading accuracy alone is not enough. However well the text is read, if someone still has to find and upload the file, the work stays with a person.

What we built

• Inbox monitoring — Attachments arriving in the inbox are picked up automatically. No one uploads anything
• OCR reads the text — PDFs or photos, scans or phone shots, the text inside is extracted. It is designed for documents with a set layout, such as quotes and receipts
• AI turns the text into fields — OCR returns a block of text, not a table. AI works out where the items end, where the amounts begin and which line holds the total. This is where documents in different layouts end up in the same columns
• You decide the format — Every company keeps its books differently. Instead of asking you to adopt our template, it exports to Excel in your own format
• An option for sites without email — Where email integration is difficult, such as on a secure network or in the field, documents enter the same pipeline another way
• No originals kept — Once the data is extracted, the source files are deleted. Not storing other companies’ quotes on our servers comes first
• Built for peak load — Reading takes time. Documents are queued and processed one by one in the order received, so a single stuck document does not hold up the rest

Results

When an email arrives, the rows fill in with no one lifting a finger.

It is currently free of charge during the pilot period.

See more products we build and run ourselves under In-house products.