C’est moi Wizidee, votre solution-experte de la collecte et du traitement des documents ! Faisons connaissance autour d’un sujet essentiel : le traitement de documents en temps réel. Le terme est souvent utilisé dans le jargon des solutions digitales et applications, mais qu’est-ce que ça veut dire exactement ? Je vais tout vous expliquer.
What does processing documents in real time bring to the table?
The notion of processing in real time is about instantaneity. It’s the fastidious job of recognising and processing, performed by a smart solution which “reads” and checks all the documents provided via mobile. Processing is automated and performed locally. But the most important thing is that all these stages are performed in a matter of milliseconds. No more waiting around looking at a screen!
Conversely, it’s easy to see that there is no real time when front and back offices are completely siloed and an officer must intervene physically to approve the documents. In this case, the client supplies the list of documents required, then an officer takes the time necessary to process and check the documents. Even if the officer gets to work straight away, the client has lost both time and autonomy before gaining access to their service. There is also a risk of introducing human error, that will need to be corrected later.
It’s not real time when processing time takes longer than 30 seconds per document, whether or not the process has been automated. The client pipeline is fully automated, but the processing time for each document is so long that completing your file becomes really fastidious.
In other words, processing in real time makes for a swift, easy and reliable user experience. You can send documents from a remote location in a matter of seconds. They are processed, checked and approved instantaneously, without needing to consult external sources.
To get real time, you need a smart way of reading documents.
How does it work?
There are various ways to process, check and collect data from a remote document. What you need to remember, is that artificial intelligence, including machine learning, plays an important role.
With these methods, document recognition is optimised, with the use of an entire library of pre-registered document templates. The algorithms can thus identify the type of document received instantaneously. So they can determine whether it’s an ID card, proof of address or a sign-up form.
And to go even further than automated document recognition, the latest deep learning technology applied to optical character recognition(OCR) helps to read the document and only collect the data needed, based on the library of pre-registered document templates. This technology can read any information on a document, from IBAN and MRZ codes and barcodes to classic and handwritten fields.
In the case of ID recognition, deep learning combines facial recognition with ID document checking. The user’s face and the ID photo supplied are also compared in real time.
Who do real-time processing solutions cater to?
Often, the industries involved are BtoC services where you need to juggle regulatory compliance (KYC, GDPR etc.), risk control and the capacity to respond ever more quickly and remotely to their clients’ individual needs. But generally speaking, all industries may need this type of solution: Banking, Insurance, Telecoms, Energy, Services, e-Commerce etc.
Let’s look at an example from an industry where access to the product or service requires an ID check. Opening a bank account, requesting a loan, signing up for an insurance policy or a telephone plan: these tasks are often performed remotely, but still need to be performed meticulously, to ensure that the documents supplied are both valid and compliant.
In the case of a vehicle rental service, the remote rental service will ask the client to send their driving licence in order to move on to the next stage. With real-time document processing, formalities are handled faster because the documents are approved and certified without any delays caused while the rental firm processes the documents. No more waiting around when you bring the vehicle back!
Another example to show how pertinent a real-time document processing solution can be: when checking someone’s age and more specifically whether they are a minor. This is a regulatory requirement in many industries including the sale of alcohol or vaping fluids (online or at self checkout terminals), content platforms that are unsuitable for minors, online gaming etc. Moreover, nowadays, checking if someone is a minor has become a sensitive subject, in France for example there are associations saying it’s too easy to access porn websites.
How can you be sure of choosing the right real-time document processing solution?
To be sure of getting it right, take care to check that it offers not only all that I’ve already mentioned but also these two features:
First, the ability to extract precisely what you need to know. You can avoid drowning in data if your solution is purely operational and only retrieves the information you actually need.
The second feature follows on from the first: information extraction must be updated in keeping with new business regulations and other legal obligations.
Now you know all about the issues inherent in processing documents in real time. If you’d like a more in-depth exploration of the subject and Wizidee, contact me!