KYC Flow
Last updated
Last updated
KYC Flow is mainly divided into 3 parts: ID verification, face verification and data verification. AML screening is optional.
Either users take ID image realtime or submit from your gallery, our OCR AI will start verifying it. First it checks if the submitted image contains an ID and what kind of information is on it. Then it transforms image into text. We display the transformed data(OCR Data) and let users clarify it(User Input) if necessary.
In the second step, when user takes or submits a selfie, we have two AIs working for identification and forgery detection.
First, Face Compare AI compares the two faces, one acquired from ID and one from selfie. The second AI(Liveness AI) detects that the submitted face photo was not stolen from others. Trying KYC with pictures or videos of others is the most common forgery we filter out.
After face verification step finishes, verifying data process will follow. We compare OCR data and user input data acqired at 1st Step.
At this moment, Text Matching AI works during the name comparison step. This AI reads the similarity between the two texts and detects differences, such as the same name but written in different languages or middle names. For example, the similarity between the name Joseph Robinette Biden and Joe Biden is highly appreciated and regarded as the same person.
AML screening is conducted using the identity information obtained through KYC. Text Matching AI also works to reduce the possibility of missing dangerous people.
