È possibile una riconciliazione digitale e… smart? – Data Manager Online

In the financial and administrative jargon, “reconciliation” can take different meanings; In this article I’ll focus on the meaning about money collection, i.e. the linking of financial flows with the respective titles of credit, and the recording of such collection in a management system. This is an intrinsic process of the active cycle of companies, but it is also common in the finance world realities that deal with specific forms of credit, such as FACTORING and ADVANCES on INVOICES.

Analyzing in greater detail the financial and accounting entities the reconciliation deals with, we can see, first and foremost, the financial flows and, more specifically, the component concerning the revenues: sums received from the customers of a company to pay off a good or service.

The other fundamental entity is constituted by the titles of credit: documents univocally coded by companies, attesting the right to demand, within a certain date, a certain sum of money from a debtor: invoices, bills, taxes, etc.

The challenge

Reconciling, in practice, means answering to two simple questions: Who pays? What is he/she paying?

It is possible to answer without significant complications, if that the financial flows are composed of structured forms of payment: those that, in relation to the subject of payment, report in a positional way the identifiers that lead directly to the title of credit: Sepa Direct Debit, MAV, RIBA, pre-compiled bulletins and PagoPA (the Italian Payment System for the Public Administration) are examples of structured forms of payment.

Answering becomes more complex when the financial flows are composed of unstructured forms of payment, in which the payer’s informations and the purpose of the payment are provided in free text fields. The SEPA or SWIFT transfers and the infamous hand-written white bulletins are examples of unstructured forms of payment. For these, it is necessary to analyze the paying party to identify the debtor, then comprehend the meaning of the purposes of payment, and extract from these the object of payment.

Possible approaches

The realities that carry out manual or semi-automatic reconciliation face significant costs, since they have to assemble teams and resources dedicated to this purpose. These costs persist even if one relies on companies that perform outsourced administrative services, including reconciliation. The goal is to be effective, to reconcile correctly and on time, but often the backlog is cumbersome or the solution not efficient.

An alternative is the digitalization of the process that automatically solves repetitive activities, generating efficiency: we are in the field of DIGITAL RECONCILIATION.

The majority of the adopted solutions at a global level are of the “traditional computing” kind: algorithms hard-wired with known cases to successfully correlate payments and titles. But this is the limit of “traditional computing”: the known cases, because the presence of payments with uncommon characteristics requires the supervision of an operator to solve the unknown or doubtful case.

Luckily, in support of solving real life problems, software technologies of cognitive computing are taking hold, that emulate human reasoning to solve problems with high variability: we are talking about Artificial Intelligence, and more precisely of artificial neural networks: software with performances definitely better that traditional algorithms. In fact, in those situations in which a neural network has to solve unknown cases, it is able to give a correct answer by approximation, having been trained on a large number of cases from which it has drawn the experience that will allow it to generalize. Just as our brain will allow us to read this sentence without too much effort: “Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn’t mttaer in waht oredr the ltteers in a wrod are”, a properly trained neural network will be able to understand correctly the content of a text field.

With artificial intelligence, the ambition is to minimize the cases that need human supervision, thus increasing the level of efficiency, with an additional very interesting ability of self-learning.


Mario S. Farris – Fintech Entrepreneur

* * *

K Linx is the software platform of the innovative Milanese startup of the same name, devoted to the automatic reconciliation of payments. It deals in the automatization of those processes typically carried out manually in the back offices of companies or in the financial world.

K Linx is already successfully used in top credit institutes, financial firms and other enterprise companies. The startup is successfully experimenting with artificial neural networks to improve the platform, surpassing the already excellent performances obtained with traditional computing algorithms.


#artificialintelligence #startups #accountreceivables #reconciliation #riconciliazione #efficientamento #incassi #pagamenti #cashmanagement #reconciliation #accountreceivables #payments #fintech #startupstrategies