Identity verification (IDV) is software that checks that an identity document is genuine, and that the person holding the phone is the person on it. It usually compares a photo of the document with a selfie. Banks, brokers, lenders, insurers and apps call it through one API: their app sends the two images to the IDV service, and the service sends back a verdict a few seconds later. IDV is the core of digital KYC, the identity checks that regulated companies must run on new customers.
IDV runs four checks. First, it reads the document. Software finds the card in the photo and reads the name, the date of birth and the number. This reading is called optical character recognition, or OCR. Second, it checks that the document is genuine. It looks for signs of editing, such as a changed font or a pasted photo. It also checks the number with the official source. For a PAN card, that source is the Income Tax Department's database. Third, it compares the face in the selfie with the photo on the card, and gives a match score. Fourth, it confirms that the selfie shows a living person, not a printed photo or a video on another screen. This is called liveness detection.
Here is a typical case. A buyer at a two-wheeler showroom in Coimbatore wants a loan for a ₹95,000 scooter. The lender's sales executive opens the loan app on a tablet. The buyer photographs her PAN card and takes a selfie. In about eight seconds, the IDV service replies. The PAN is valid, and the name matches the official record. The face match score is 94 out of 100, above the lender's limit of 80. The liveness check passes. The app moves on to the loan decision, in the lender's loan origination system (LOS).
If any check fails or is uncertain, the case goes to a person. The lender's team reviews it, or asks the customer to complete video KYC, a live call with an employee.
IDV must work in difficult conditions: a cheap phone, poor light and a weak mobile connection. Indian IDV companies built their models for exactly these conditions.
IDV is now in a contest with AI. AI tools can produce fake faces and edited documents that look real. So many IDV services add deepfake detection, which decides whether an image or video was ever real. After IDV, the customer's name usually goes to sanctions and PEP screening.
The same checks appear far from banking. Background verification uses them to check job candidates, and online proctoring uses them to check that the right student is taking an exam.