Demo / Incoming orders and documents
The agent proposes.
A person approves.
Orders by email, customers’ PDFs, suppliers’ delivery notes: every day someone reads them, looks up the items, checks the prices and copies them into the business software. Here an AI agent does the first part and prepares a proposal. A person reviews it, corrects it if needed and approves it. Only then does the data go into the business software.
A simulation with invented companies and data. No document leaves this page and no AI model is called: it is a reconstruction of the workflow.
Try it with a document.
- 01Document
- 02Reading
- 03Review
- 04Business software
Document preview
Beyond the demo
How it works in practice.
The same workflow, connected to the documents and software the business already uses. You start with one type of document, measure the results, then widen the scope.
- 1
It reads where the documents arrive.
The orders mailbox, attached PDFs, scanned delivery notes. Even orders written in plain words, like a message to a sales rep.
- 2
It checks against your data.
Customer and supplier records, price lists, stock levels and order history. Anything that does not add up becomes a flag, never a silent correction.
- 3
It writes only after approval.
Into the business software you use today, with a log of who approved what and when. If the agent is switched off, the work goes back to being manual and nothing grinds to a halt.
Questions
Frequently asked questions.
Does it work with our business software?
It depends on how the software takes in data: an API, file import or database access. This is checked in the first analysis, before any quote.
We use Excel, not business software. Does it make sense?
Yes. We start from your spreadsheets: customers, items, price lists. For everyday work the tool takes the place of the spreadsheet, with checks and a record of who approved what; Excel stays for exporting the data when you need it. Writing straight into a shared file is only done where it is safe, because formulas and simultaneous edits make it fragile.
What if the agent misreads a document?
It can happen, especially with poor scans or ambiguous wording. That is why every field has a confidence level, doubts become flags and nothing goes into the business software without a person’s approval.
Where do the documents go?
Which AI services to use and where the data goes is decided at the start, case by case. In this demo no data leaves your browser.
How much time does it save?
It depends on how many documents come in, how they are written and on the business software. It is estimated on your real documents before starting, not in advance.
How many documents do you copy out every week?
A first conversation. A few sample documents, anonymised if you prefer, and the name of the business software you use, or the Excel files you work in, are enough. We’ll look together at whether an agent makes sense and which type of document to start with.