↳ documents · service

Data extraction from PDFs, invoices and scans

We pull the data out of invoices, contracts and forms and put it into Xero, your CRM or a spreadsheet as a draft for you to approve.

↳ saves ~4 to 8 hrs/week · built in 1 to 2 weeks

Book your free 15-min call See pricing →

In practice

An example from a business like yours.

An example ↘

Say you run a small bookkeeping practice and you're retyping supplier invoices, receipts and bank statements into Xero for six clients, hours every week. We'd set up a dedicated email address per client. PDFs forwarded there get their line items read out, matched against that client's existing Xero accounts, and posted as a draft journal entry. You review and approve in batches. The typing stops, and the time goes back to clients.

What we'd build

What happens to one invoice once this is built.

A document arrives

By email, upload link, phone photo or a watched folder. Nothing to install.

PDF invoices into your accountsDraft bills in Xero, QuickBooks or Sage
Receipts from a phone photoPhoto in, booked with VAT and category
Contracts summarisedDates, amounts and notice periods in a tracker
Forms and applications processedFields typed in, missing ones chased
Scans and handwriting readOld files and handwriting made searchable
Checks against a checklistID, certificates and expiry dates checked

You

A batch to approve on Friday, with the exceptions flagged at the top.

↳ you approve. You stop typing.

↳ you approve. You stop typing.

↳ saves ~4 to 8 hrs/week · built in 1 to 2 weeks

Works with

Tools we work with.

If yours isn't on this list, ask. We can almost certainly work with it.

Xero QuickBooks Sage FreeAgent Notion Airtable Google Sheets Excel Dropbox Google Drive OneDrive Expensify Dext

Common questions

Questions about document & data extraction.

How accurate is the extraction?
It depends on the documents. After about a week of tuning on your specific suppliers and formats, clean invoices and statements are read correctly nearly every time. Contracts and scans vary with how consistent the source is. Every item carries a confidence score, and anything below the bar goes to you for review.
What document formats do you support?
PDFs (digital or scanned), JPG and PNG photos, Word documents, Excel sheets and handwritten forms. For scans and photos we use OCR plus AI vision models, which handle skew, shadows and poor lighting better than older OCR tools.
Can it handle multiple languages?
Yes. UK English is the default, but the underlying models read most European languages (French, German, Dutch, Spanish, Italian) and many others. Useful if you take supplier invoices from EU vendors.
How is the data secured?
Documents are processed in transit and at rest with encryption. We can self-host the extraction pipeline on your own infrastructure if compliance requires it. For most small businesses we use enterprise-grade cloud providers (AWS, Google Cloud, Anthropic, OpenAI) under DPAs.
What's the ROI math look like?
A typical project: £1,400 build, saving around 6 hours a week. At £40 an hour for your time that's £240 a week, or roughly £12,000 a year, so the build pays for itself within a couple of months.
Ready when you are ↘

Start small. Start with a call.

Tell us what's slowing you down. We'll tell you, in plain English, whether and how we can automate it. Free, 15 minutes.

  • Free 15-min call
  • £250 audit option
  • Fixed quotes
  • No long contracts
£250
audit · fixed · one-off
Book the audit → or Book a free 15-min call →