How to improve KYC conversion with optical character recognition

Regulated financial institutions, such as traditional banks and brokers, must perform Know Your Customer (KYC) and anti-money laundering (AML) checks before establishing a business relationship or onboarding new clients.

ID document and selfie authenticity checks, proof of address, and electronic signatures help protect both businesses and clients against fraud. But they can also be a huge source of friction during client onboarding.

Onboarding is a make-or-break moment. Do it well, and you have a new client; do it badly and you risk ending the relationship before it has even started. The good news is that a robust, detailed, and accurate KYC flow needn’t come at the cost of conversion. In this article, we will explore how Fourthline’s optical character recognition (OCR) technology can help you remove friction from the identification and verification process, ensuring you retain high conversion rates at this critical stage in the client journey.

A better way to capture and verify ID documents

Traditionally, the process of verifying ID documents was cumbersome, time-consuming, and prone to error. Clients had to type in their personal data character by character. If they made a mistake, they would have to delete it and try again, which is especially arduous on a mobile device. To help solve this, Fourthline has built an Identity Verification solution that uses our proprietary machine-learning models to extract the relevant data from ID documents within seconds, fully replacing the manual process.

Your client simply needs to take photos of their ID document for our OCR technology to perform the following steps within two-to-three seconds:

• Check whether a document is present in the picture.

• Confirm the detected document is an ID document.

• Extract and compare the data from the visual inspection zone (VIZ) and machine-readable zone (MRZ).

Our models have been trained on hundreds of thousands of ID documents from multiple countries. This ensures our solution can extract and read fields from a wide range of document types. We continually update our AI engine with the latest technology and newly issued ID documents to ensure the highest accuracy. Furthermore, our models are trained on our proprietary data sets, meaning we can perform more accurate readings compared to other solutions on the market. The whole pipeline shown in the video achieves over 95% accuracy for the set of documents currently supported.

In action: How OCR works with the Italian paper ID

The format of the Italian paper ID is much more diverse compared to other ID documents, which makes it difficult to verify. Despite being first issued in 1994, it is still very much in circulation, making it a central part of the onboarding flow for Italian clients. Here is how our system recognizes and verifies this document type.

1.The client takes a photo of the document.

Image source

2.Fourthline’s OCR technology detects and recognizes the document.

3.We extract all the relevant information:

Image source

4.Et voilà! An accurate client record in seconds without manual input from the client.

Fields: Surname, first name, birthday, place of birth, nationality.

Reading: Allen, Emerald Al, 06/02/1986, Louisiana, Stati Uniti D’America, Statuniense.

Learn more

Our solution is significantly faster and more accurate than manual data input. It can enhance your conversion rate and ensure you maintain accurate and compliant client records. Book a personalized demo to learn more about how our solution can help you achieve a high standard of user experience, while remaining compliant every step of the way.