Automated Invoice

Automated invoice is the solution which facilitates the automated management of the accounts payable process.

Automated Invoice


The Accounts Payable process involves both the posting of invoices for accounting purposes, and the relative reconciliation with the corresponding purchase orders/delivery notes/receipts. This verification involves the comparison of each individual line item contained in the two documents (invoice and purchase order/delivery note/receipt).

In order to identify possible differences this process is for the most part handled manually by companies, using paper documents for the posting phase, as well as for the comparison of line items of the invoice with those of the purchase order/delivery note/receipt.

The posting process is automated with the collection of data necessary for the entry into accounts of the invoices (for the total invoice amount and/or at the detailed row level) through appropriate Optical/Intelligent Character Recognition (OCR/ICR) solutions, with the provision of accounting flows towards the company's management/ERP system.

The automated reconciliation is carried out using a Machine Learning component (IRIDE – Intelligent Reconciliation of Invoice and DElivery Notes), which compares and matches the detailed invoice data with the relevant purchase order/delivery note/receipt details, producing the corresponding output (in data format as well as in report format).

The solution

Accelerator for end-to-end management of the billing process, from data entry automation for information contained in invoices, to the automated verification and reconciliation, using OCR and Machine Learning algorithms, of the invoices and the relevant delivery notes, with automated identification of the reasons for any discrepancies (i.e. prices and/or quantities different from those agreed upon, etc.), based on product codes in the materials database and any other relevant data.

Focus On

Maximum attention to costs, efficiency and quality of processes are some of the key drivers that have been guiding businesses in recent years. Within the Finance & Accounting sector, the Posting and Reconciliation of supplier invoices has come under particular focus.

Automated invoice is the solution that allows businesses to achieve these goals. It facilitates the automated management of the accounts payable process, from the posting phase to the reconciliation between invoices and purchase orders/delivery notes/receipts, highlighting the differences identified.

The current scenario - Critical aspects

The manual process is demanding in terms of manual effort and processing times, as well as being subject to errors and poor data quality. In addition to these inefficiencies, the manual process does not make it possible to capitalise on the knowledge and the added value derived from the repeated execution of the process.

High management costs, especially in scenarios where a significant number of documents are handled, generate high operating costs, which are not justifiable with respect to the added value produced.

Moreover, the manual process, which relies on the use of paper and is highly unstructured, is difficult to control and govern in terms of both efficiency and the quality of information handled.

Automated invoice – Solution

The solution can be applied to the management of different types of invoices, as well as to the posting and reconciliation processes. In particular, the solution is designed to handle the following types of processes:

  • Posting with reference to the purchase order/delivery note/receipt;
  • Posting without a purchase order;
  • Postings made on the total invoice amount (full VAT credit line);
  • Postings that take into account details related to every single line;
  • Postings by cost centre/point of sale (or other secondary accounting items).

A particular type of invoice whose posting can be automated consists of Utilities invoices for companies operating in the Retail sector. These are invoices for which the final balance must be calculated by point of sale in order to verify the store’s sales margin. These invoices are characterised by a relatively standardised format containing information related to each individual point of sale, and by a large number of pages, which increase with the number of stores managed. This automation allows companies to accurately and quickly report the final balances by point of sale, eliminating repetitive human activities with low added value.

The use of advanced technologies based on OCR/ICR makes it possible to extract and validate the document information, while the application of advanced Machine Learning technologies helps identify any balancing errors.

The solution outputs are:

  • Automatic posting flow (direct integration with ERP systems);
  • Automatic reconciliation flow (Discrepancy reports and/or direct integration with ERP systems).

The use cases developed

Automated Invoice Posting (goods/services) with reference to the purchase order/delivery note/receipt:

Through the use of OCR/ICR: validation of data automatically captured from the header and the lines of the paper document using management/ERP system databases (e.g.: supplier database, list of purchase orders, delivery notes, VAT rates, etc.) to generate an integrated accounting flow.

Automated Invoice Reconciliation for goods or services:

Using data from the OCR processing of the line items, the ML IRIDE solution makes it possible to identify differences between invoices and reference purchase orders/delivery notes/receipts, producing data flows towards the customer’s management/ERP system and/or reconciliation reporting, available for user processes and to be shared with other departments (i.e. sales, purchasing) or with the manufacturer for the resolution of anomalies.

Automated Posting of Utilities Invoices with posting to the ERP system:

Transformation of paper invoices into PDF files and, from this format, transformation into a deconstructed txt file. Through Smart OCR/ICR recognition of information (based on keywords, semantic logic, in some cases the application of rules and algorithms) at the single point of sale and the total invoice amount level for the relative correctness checks, and insertion of logic (using Python) for transcoding the information (from point of sale to cost centre, VAT code, accounting records, etc.) into the language of the corporate accounting system. Data generation for potential validation by the user or the direct generation of accounting flows in the accounting/ERP system.

The solution architecture

The architecture includes 4 macro-components:

  • OCR/ICR component: in general, the solutions are not based on predefined rules and on the definition of templates, but on semantic logic and self learning and adaptive technologies that capitalise on the related processing activities;
  • OCR/ICR LineItems component: solution for the recognition of line items, including when dealing with unstructured data, with self learning abilities and enhanced recognition performance over time;
  • Machine Learning Component (IRIDE): the IRIDE model is based on the Scikit-Learn machine learning libraries developed in Python;
  • Reconciliation Component (Discrepancy Report): the web interface for reconciliation and for the extraction of the reconciliation report is developed in the ASP.NET MVC language.

The solution is implemented on both Microsoft SQLServer and Oracle databases.


Data Robotics Accelerator

Employee Monthly Expenses

Employee Monthly Expenses is the solution that facilitates the automated creation of expense reports starting from the underlying cost items, quickly and without the need for manual intervention.

Data Robotics Accelerator

Know Your Orders

Know your Orders is the solution that makes it possible to create a simple interface which can be consulted by users using a natural language, thus facilitating access to information while ensuring consistency and accuracy.

Data Robotics Accelerator

Brick Machine Learning

Brick Machine Learning is the solution that makes it possible to simulate various configurations used in automated production lines, in order to recommend the optimal mix of devices required to achieve the overall performance requested by the end customer.