// the challenges of automatic supporting document verification
Regulatory inflation leads to a sharp increase in compliance expenses notably through the massive increase in staff numbers (control, risk and compliance functions).
Meanwhile, despite these investments, fines are on the rise. Between 2009 and 2018, this resulted in $145 bn being paid by the major European banks according to the BCG.
In parallel, AI applications are being developed. According to Gartner, 59% of global companies have made investments in AI. The reason for this success? More computing power, better algorithms in ‘learning’, and virtually unlimited access to learning data.
IN PRACTICE
- Manual review of supporting documents leads to disruption in the customer experience and/or delayed response, calling into question the file's transformation.
- The payroll increase devoted to controls poses recruitment, management, and, most significantly, training problems.
- Most of the these teams' time is spent on repetitive tasks that can be fully or partially automated.
” Provide users with real-time guidance on the submission, thereby ensuring a complete and compliant customer dossier.
Automated verification of supporting documents enables 100% automatic analysis of the submitted documents’ image quality, and their validity and consistency with regard to the information transmitted.
The authenticity of the received documents is thus checked (security elements, connection to external repositories, etc.) in order to detect attempts at documentary fraud.
Your regulatory obligations are thus automated and your risk of fraud significantly reduced.
// STEPS FOR AUTOMATED SUPPORTING DOCUMENT VERIFICATION
Image quality
The first step in the checks is to check the quality of the submitted images (resolution, blur, noise and contrast level).
If the image quality is not sufficient to be processed, a clear message will explain how the user can improve the submission.
Type and validity of documents submitted
The second step is to check that the submitted supporting documents are those expected and that they are valid. The documents currently recognised are: identity documents (national identity card, passport, resident card, driving licence), RIB statement of banking identity, pay slip, tax notice, telecom bill, invoice and energy payment schedule (electricity, gas and water), registration certificate, property tax.
File consistency
The third step is to compare the data retrieved from the documents with the information in the file.
On all documents, this is mainly the customer’s surname, first name and address, but also the date of birth and gender on identity documents.
After these first three steps, the records will be deemed complete and compliant.
Authenticity
A fourth step is to analyse the authenticity of the submitted documentation. These checks are specific to each document (security elements, connection to external repositories, etc.).
Artificial Intelligence has changed the game. Its integration into remote underwriting tools makes it easier, thanks to deep learning, to extract data, by adding its verification, even from documents poorly rendered by a photo taken on the fly with a smartphone.
Today, in fact, users want to be able to send all the documents necessary to finalise their transaction via the same channel. For them, there is no question of waiting: immediacy has become the rule.
Embedded in remote underwriting tools , AI has therefore created a real disruption in the customer experience, making it smoother, thus boosting digital underwriting in real time.
Operational efficiency
with greater visibility into the underwriting 'pipeline' and faster processing of customer files
Detect
attempted identity theft and documentary fraud
CLOSE UP ON TRUST AND SIGN
The solution for digital login and subscription
Trust and Sign responds to the main login and subscription challenges by offering the following solutions:
