Name Match in KYC: Fuzzy Matching Explained
Why "RAHUL KUMAR SHARMA" and "Rahul Sharma" should still match — and how a name-match API decides.
Names in Indian KYC rarely match character-for-character. The same person might appear as "RAHUL KUMAR SHARMA" on their PAN, "Rahul Sharma" in your form, and "R K Sharma" on their bank account. A strict string comparison would reject all of these — which is why KYC uses fuzzy name matching that returns a similarity score instead of a yes/no.
What trips up a simple string check
- Word order — first/last name swapped or middle name present.
- Initials — "R K Sharma" vs "Rahul Kumar Sharma".
- Spelling and transliteration variants of Indian names.
- Extra titles, spaces, punctuation and case differences.
How a name-match API helps
A name-match API normalises both names and returns a match score (and usually a pass/review flag) that tolerates these variations, so you can auto-approve confident matches and send only genuine mismatches to manual review. It is commonly used to match a customer's entered name against the name on their PAN, Aadhaar or bank account.
Good practice
- Set a score threshold for auto-approval and route borderline cases to review.
- Normalise inputs (trim titles and extra spaces) before matching.
- Log the score with each verification for your audit trail.