Document RecognitionAI-based document recognition

Process creditor documents automatically

Is your accounts payable process tedious, slow, and inefficient? The digital mailbox, DeepBox Mailroom, has the solution! With intelligent document recognition and predefined workflows, documents can be entered and posted in accounts payable with minimal manual effort.

Receiving creditor documents

Creditor documents enter the company in diverse ways (letter post, e-mail, scan, electronic banking, eCommerce platform, ...) and in different formats. It is not always easy to allocate the creditor documents to the right department, cost centre or responsible person.


Forward documents to a digital inbox

Where and how the creditor document arrives at the company no longer matters. All creditor documents can be sent to the digital mailbox, DeepBox Mailroom, for further processing.


Analyse documents

Creditor documents are analysed in the digital inbox by industry-leading intelligent data capture AI (DeepO) and important information is extracted. The result is structured data: from the supplier address to the applied VAT, to individual invoice items.


Allocate receipts

The digital mailbox assigns the documents to a business area, cost centre or project based on the extracted information. The person responsible for the business area checks the document and can manually add more information if needed.


Enter documents in Abacus

A booking is automatically created in Abacus for the creditor document and the approval process for payment is initiated. Set rules allow direct booking into Abacus software (without step 4). In practice, this is often used for invoice amounts below a defined amount.

"DeepO has significantly improved our invoicing processes."

Transport AG Aarau is optimising its accounts payable process and relying on DeepCloud's DeepO artificial intelligence. Read in the interview with Dominik Stahel, Head of Finance and Accounting at Transport AG Aarau, how efficiency is increased in day-to-day work.


How is the document analysed?

The digital mailbox recognises individual invoice items as well as predefined keywords such as project numbers or cost centres. Therefore, automatic document recognition is ideally suited for the following use cases:

Recurring invoices

Recurring invoices with known invoice information, e.g., Swisscom invoices with the employees' telephone numbers, can be booked individually to the corresponding cost centre using item recognition.

Collective invoices

Collective invoices with many individual items, e.g., SBB invoices, can also be posted to individual cost centres.

Invoices with order numbers

Invoices with order numbers are recognised and can be linked to the order from Order Processing.

Define workflows and rules

Set your own rules and workflows for sending supplier documents directly to Abacus for further processing.

FAQ -  Frequently asked questions about document entry

What is DeepBox?

DeepBox is a Swiss cloud-based digital mailbox, with secure access that can be shared with selected people upon request.

How do I guarantee data security and data protection?

DeepCloud AG's cloud solutions are certified according to ISO 27001:2013 and your data is stored in Switzerland.

How is access protection regulated?

DeepBox is fully integrated so that it is enriched with content from the Abacus software. This enables the mapping of structures such as business units (BUs), projects or cost centres. Access rights and validation processes are therefore controlled centrally via the Abacus software.

Where are creditor documents released for payment?

For this purpose, the usual approval process in the Abacus software continues to be used. DeepO is used to automate routine tasks such as the entry of documents, account assignment and allocation to the respective persons.

Can I "train" DeepO?

You can continuously add new keywords and terms so that DeepO is constantly improving. The stored information is browsed for each invoice. The keywords are multilingual and can be defined per supplier, box or across multiple boxes.

What is the difference between OCR and AI document recognition?

Essentially, the difference between OCR and AI-based document recognition is that OCR (Optical Character Recognition) is mainly focused on extracting the text from an image or document, whereas AI-based document recognition can recognise and understand a wider range of document elements to better capture the meaning of the document.


With DeepBox you can process, store and share documents in a secure and automated cloud environment.

Learn more


DeepBox's industry-leading intelligent data capture AI.

Learn more