Documents read themselves. You only look at the doubtful ones.

Every incoming document is recognised, read and turned into data: lines and amounts for the ERP, deadlines and clauses for whoever needs to know them. When the system is not sure it says so, and that document goes to a person.

Sectors
manufacturing, logistics, construction, services
Used by
administration, purchasing, quality, legal
First prototype
2–3 weeks

The problem

Electronic invoicing solved part of the problem, not all of it. Order confirmations, delivery notes, certificates, contracts and tender specifications still arrive as PDFs or on paper, each in its own format.

Someone opens them one by one, looks for the data they need and copies it, or reads them in full to find the clause that matters. Nobody wants to do it and nobody can stop doing it. A mistake on a date or a penalty comes to light weeks later.

How the data flows

  1. Where it comes from: Email and PEC (incoming attachments); Scans and photos (crooked ones too); Shared folders (contracts, specifications).
  2. What happens, Document reading: AI, Recognises the document (type, sender, language); AI, Extracts the fields (lines, dates, batches, amounts); AI, Analyses long documents (deadlines, penalties, clauses); Compares (with orders and terms).
  3. Where it ends up: ERP (clean documents, registered); Analysis sheet (with links to the pages); Review queue (uncertain fields highlighted).
  4. every review becomes a rule for that supplier.
Steps marked AI are the model's. The rest is ordinary automation or people.

How we solve it

04
01

It recognises the document

Type, sender, language, even from crooked scans, photos and tables split across several pages. Each document type follows its own rules instead of one model for everything.

02

It extracts the data for the ERP

Order number, lines, quantities, dates, batches, amounts. Clean documents go in on their own; those with an uncertain field go to review, with the field highlighted and the original next to it.

03

It analyses the long ones

Contracts and specifications become a summary sheet with deadlines, penalties, renewals, requirements and non-standard clauses. Every point links back to the page it comes from.

04

It compares against what is there

The supplier's confirmation against the order issued, the contract against the standard terms. A difference becomes an alert, not a surprise.

What we measure

03

These are the numbers we set in discovery, before writing any code. The starting value is measured on your data, and the prototype is judged against it.

Documents registered without intervention

The share that goes straight into the ERP. It grows as rules are added for the most frequent suppliers.

Time per document

From arrival to registration or to the analysis sheet, including the review of doubtful cases.

Differences caught in time

Deviations on prices, quantities and dates, and non-standard clauses, found when the document arrives rather than when it is too late.

When it is not worth it

If almost all documents come from a few suppliers with fixed formats, a direct integration or an EDI feed is more reliable and cheaper. And analysing a contract does not replace a lawyer's opinion: it tells them where to look.

Services involved

Other use cases

11 scenarios
05 · RAG and knowledge base

Document assistant

Customers and colleagues ask questions, and the assistant answers using only your documents and the pages of your website. Every answer comes with its source, and when it does not know, it says so.

06 · RAG and knowledge base

Technical support

Manuals, technical sheets and resolved tickets become answers with the source next to them, for the technician on site and for whoever answers the phone.

Tell us the process. We will tell you if it makes sense.

A free 30-minute call, with no commercial follow-up unless you want it.

Request a call

Companies that have already worked with us

  • Velux
  • ISU
  • MovePay
  • AB Group
  • Archiva Group
  • Onda SpA
  • PDE
  • Bynary System
  • COOPI
  • Gimasi
  • Creactives SpA
  • Rev AI
  • Odditt Inc.
  • Si Sociale
  • Stones
  • Turboden
  • Schindler