The power of automated document verification for KYC/AML compliance in African markets

As more businesses look to expand across Africa, they’re faced with the reality of the continent’s fragmented regulatory landscape. KYC/AML regulations have become more critical across Africa as governments look to combat fraud, terrorism financing, and other financial crimes.

According to a 2021 report from the Financial Action Task Force (FATF), 48 of Africa’s 54 countries have criminalized money laundering and set up some legal frameworks for AML/CFT measures. Although specific regulations can vary between countries and industries, customer identification remains at the heart of all AML/CFT efforts. Know Your Customers (KYC) procedures usually involve collecting and verifying certain information about customers, such as their name, address, date of birth, and government-issued identification. This information helps financial institutions and other regulated entities identify customers and assess their risk profiles.

For most KYC processes, after collecting information there are three applicable methods to verify the customer’s information:

• ID Number Verification: This requires verifying the ID number on any government-issued ID. While this can be a great solution, the lack of centralized databases in many countries and issuing institutions means that it is not a readily accessible option.

• Manual Document Verification: This requires a review of the ID by a specialist who can manually identify discrepancies and verify the authenticity of the documents. This method is significantly limited by scale. With 54 different countries and over 200 acceptable ID types across the continent, it would require a significant number of hires to complete KYC reviews across markets.

• Automated Document Verification: This involves the use of machine learning techniques like optical character recognition (OCR) to analyze documents for signs of forgery or tampering. This kind of verification can be highly efficient at fraud detection and is typically scalable in multiple markets.

This information helps financial institutions and other regulated entities identify customers and assess their risk profiles.

What is Automated Document Verification?

As highlighted earlier, Automated Document Verification refers to the use of machine learning algorithms to verify the authenticity and validity of a document piece, in this case, ID documents or any other type of KYC document.

The process typically involves identifying the documents that need to be verified, labelling important information from the document, gathering a large dataset of the specified document, and training an algorithm with the dataset. The algorithm then learns the differences between authentic and fraudulent versions and is faster able to spot them than a regular human observer.

Smile Identity also employs Optical Character Recognition (OCR) to extract the personal information from the document and return it in the results. The data could then be compared against other information sources for a very high security check.

Take the Kenyan National ID for instance which is issued to citizens who are 18 years and above:

A typical Kenyan ID contains multiple data points which are extracted and analyzed by a computer. This process is repeated with multiple, well-labelled samples of authentic and forged IDs, and the algorithm is taught to look out for markers like font type, size, spacing, watermarks, and lots more.
Automated document verification is the go-to method for many organizations today for a number of reasons including:

• Scalability: Provided a large enough dataset for a document can be gathered, any ID document can be verified with high accuracy using automated verification. This allows organizations to be able to verify identities across borders.
• Accuracy: If well-trained, a document verification model can reach very high levels of accuracy and often recognize fraudulent documents better than a manual reviewer would be able to.
• Speed: Since machine learning algorithms can work faster than humans, they%E2%80%99re able to verify significantly more customers within a short time. For organizations that require a smooth onboarding process, this is an important feature to have.

Despite all its advantages, automated document verification is not an infallible solution. There are some common failure points to watch out for when evaluating a document verification provider:

• Unrealistic Image Quality Requirements: Many document verification providers require extremely high resolution photos to be machine readable. This requires high bandwidth to upload the photo and a device with a high quality camera, conditions that can be difficult to meet for many users.Smile Identity has built an image capture SDK that is optimized for resolution vs image size. We have conducted thorough testing to find the perfect balance to minimize upload failure and maximize document verification accuracy.
• Weak Counterfeiting Detection: The advancement of technology means that counterfeiting documents has become more sophisticated; certain counterfeits may be too well-made for some automated systems to detect.Smile Identity combats fake documents in three ways:
  1. Document Authenticity Check: Our algorithm picks up patterns, not only with authentic IDs but with counterfeits, constantly identifying new ways of spotting fakes.
  2. Review of OCR Information: First is our use of an adaptive algorithm i.e. an algorithm that learns from data to improve its performance over time.
  3. Trained Expert Reviewers: Our networks of ID experts study document security features and can identify new counterfeits and label them so the algorithm is able to spot them in the future.
• Frequent changes in document design: Many issuing institutions routinely update the designs on their IDs. For example, Ghana’s Electoral Commission issued new voter IDs to eligible voters every election year to enhance security features. These frequent changes need to be kept up with to make sure the most updated documents are recognized by the computer. Smile Identity’s local presence in major markets on the continent allows us to respond quickly to changes in ID designs, protecting the interests of our customers.

Smile Identity has built an automated document verification solution specifically for Africa that takes into account challenges from capturing document images on older devices to keeping security checks up to date with frequently changing documents. Our Document Verification product is able to verify over 230 ID types across 52 African countries and supporting markets in Europe and North America.

Curious about our Document Verification Product? You can check it out here.

Looking to learn more about document verification? Send us a message.