How to manage fraud in the processing of credit applications?

Credit fraud attempts are becoming more sophisticated and undetectable to humans. Bleckwen relies on artificial intelligence and machine learning to detect them faster.

Processing credit applications and fraud: an endemic phenomenon

Many organizations must have a file handling system on a daily basis to carry out their activity:

• The loans to individuals and companies they are arousing growing enthusiasm, fueled in particular by low interest rates. According to the Banque de France, the outstanding balances managed by the main bancassurance groups are currently increasing by 5.5% per year (source: Bleckwen white paper).
• The car financing services experience the same progression of activity.

This directly exposes these different actors to fraud risksand certain scam techniques such as:

• AN identity theft allows a scammer to impersonate someone else, usually after stealing your personal information through a phishing attempt.
• The false statements constitute a fairly widespread type of fraud, particularly in the field of credit or tax and social fraud. The registrant will intentionally lie and provide false documents to get their money.

The fraud detection it becomes more and more difficult with the appearance of new false identities called “synthetic” or digital, manufactured from scratch and mixing the true and the false. Fraudsters also have a well-known ability to innovate and circumvent existing technologies. Faced with these different threats, the fight against fraud has become a priority.

Traditional and Emerging Fraud Threats: Some Key Figures

• Credit fraud is currently increasing by 25% per year according to the ASF, which is a considerable increase.
• Potential fraud is ten times more likely in the case of a file processed 100% online.
• Globally, financial fraud of all kinds could account for a total damage of €45 billion by 2023 according to WPI (source: Bleckwen white paper).

What solutions to prevent fraud?

With many financing or leasing specialists, the methods of fraud prevention are still traditional: it involves, in particular, requesting and verifying a large number of supporting documents that make it possible to establish, for example, the identity or the reality of the applicant’s income.

Funding agencies also rely on internal control mechanisms, with management of “white” or “black” lists to track scammers. They may also have some regulated files. Let us mention in particular:

• The record of incidents of amortization of personal credits (FICP);
• The Central Check File (FCC);
• The National Archive of Irregular Checks (FNCI).

These different files are automatically queried and only list individual cases that have already generated incidents. They will also be powerless in the face of carefully executed identity theft.

In practice, human resources to reduce fraud therefore remain limited by nature given the magnitude of the phenomenon. The detection involves crossing, for each file, tens or hundreds of variables in order to remove inconsistent elements: a difficult task even for controllers specialized in this type of transaction.

An external and automated control application, on the other hand, will give the possibility of generate alerts in real time and thus fight better against the new phenomena of financial crime.

Bleckwen: a smart real-time fraud detection and prevention solution

Bleckwen is a start-up founded in 2019 that has developed a patented anti-fraud technology in association with BNP Paribas. Its solution in SaaS mode, based on advances in artificial intelligence and machine learning, demonstrates remarkable performance in the detection, detection and prevention of fraud cases. A halving in the number of “false positives” and an increase in foiled fraud was observed compared to standard software.

A first application of the tool, from 2020, has allowed Carrefour Banque to have a reliable scoring model for the granting of consumer loans. Since 2021, the company has continued to develop its offering to target other players regularly affected by external or internal fraud, such as Renault Nissan Alliance’s RCI, one of the largest players in European auto finance on the car finance side. A major Spanish automotive finance company has adopted ML from Credit Fraud Services to combat credit underwriting fraud. Results: twice as many alerts were generated and the detection rate increased from 89% to 96%.

Bleckwen has the peculiarity of custom design of an AI model for each client company. These models are parameterized in such a way that they provide an explicit interpretation of each score obtained, with full transparency.

For credit institutions, this integration of AI into their processes is a double opportunity.
Opportunity to be more efficient and opportunity to launch other digital transformation projects exploiting existing data and tools for the fight against fraud.

The digitization of exchanges and many transactions poses new challenges in the fight against fraud and the processing of credit applications. A real-time data analytics tool like Bleckwen could provide a lasting solution to many private actors looking to limit fraud on a daily basis by allocating their activities and resources to larger projects.

The three key points to remember:

• Applying for a loan, renting a car… Online, false documents or false declarations are ten times more widespread;
• Many companies need to strengthen the nature and quality of their existing anti-fraud tools to ensure more effective internal control;
• Bleckwen has designed an artificial intelligence dedicated to the fast and efficient detection of fraud through file processing.