A borrower whose Aadhaar shows "Suresh K. Pillai", whose PAN shows "Suresh Krishna Pillai", and whose CKYC record shows "S K Pillai" is not three different people. They are one person whose name has been rendered differently across three government databases — a phenomenon so common in Indian identity infrastructure that any KYC system that cannot handle it is not fit for purpose. The KYC Verification Agent AI resolves conflicts, it does not simply report them.
The CKYC conflict landscape
Central KYC (CKYC) records were designed to eliminate repetitive KYC across financial institutions. In practice, they have created a new category of compliance complexity: the CKYC conflict. A borrower with an existing CKYC record at one financial institution presents to a new lender whose KYC captures slightly different identity data — because the borrower's name has changed (marriage), because their address has changed (relocation), because the original CKYC was entered with a data error, or because different institutions use different transliteration conventions for the same name.
Each of these scenarios requires a different resolution. Name change due to marriage is a legitimate update requiring a supporting document, not a rejection. A data entry error in the original CKYC is a CKYC correction process, not a borrower failing KYC. An address update is a CKYC modification, not an identity mismatch. Treating all conflicts as potential fraud — the default behaviour of binary KYC systems — generates false positives that damage borrower experience and cost disbursements. Treating all conflicts as benign — the outcome of insufficient scrutiny — creates compliance exposure.
The KYC Verification Agent AI classifies each conflict by type and applies the resolution pathway specific to that type.
The six conflict types and their resolution pathways
Name
Variant
Same person, different name rendering
Aadhaar "Suresh Krishna Pillai" vs CKYC "S K Pillai". Fuzzy match score above threshold — same underlying identity. Resolution: auto-reconcile using the Aadhaar name as primary (UIDAI biometric anchor is the most reliable identity document). Log the variant. No document request required. No borrower contact needed.
Name
Change
Different name — identity continuity via supporting document
Aadhaar "Priya Sharma" vs CKYC "Priya Agarwal" (maiden name on CKYC). Match score low due to surname change. AI classification: likely name change scenario (female borrower, CKYC older than 2 years, maiden name pattern). Resolution: request marriage certificate or gazette notification. Application held — not declined. Document receipt triggers CKYC update request on borrower's behalf.
DOB
Mismatch
DOB differs by 1–5 years — known government record issue
Aadhaar DOB: 14/08/1984. CKYC DOB: 14/08/1948. A 36-year discrepancy that is clearly a typo (84 vs 48 year inversion — common data entry error). AI classification: probable government data entry error, not identity fraud. Resolution: flag for manual review with discrepancy brief noting the likely error pattern. Reviewer requests DOB correction from original CKYC institution — does not reject borrower.
Address
Mismatch
CKYC address outdated — borrower has moved
CKYC address: Mumbai. Current Aadhaar address: Pune. Name and DOB match perfectly. Resolution: auto-reconcile on identity (name + DOB match is sufficient for KYC purposes). Initiate CKYC address update on borrower's behalf as a post-KYC compliance step — does not block the application. Log the address update for CKYC registry submission.
Duplicate
CKYC
Borrower has two CKYC records — different KIN numbers
KRA registry returns two CKYC records for the same PAN — one from 2018, one from 2021, with slightly different details. Resolution: use the most recent record as primary. Log both KIN numbers. Initiate deduplication request with CKYC registry — a regulatory obligation that the KYC AI handles automatically. Application proceeds on the more recent, complete record.
Identity
Mismatch
Identity data inconsistent — potential fraud or error — human review required
Aadhaar name, PAN name, and CKYC name all show materially different surnames with no recognisable relationship (not transliteration, not initials, not name change pattern). DOB mismatch beyond known error ranges. Resolution: application held. Detailed discrepancy brief generated for KYC officer. AML check triggered in parallel. No auto-resolution — human decision required at every step.
The conflict resolution outcome distribution
The conflict is the data — not the failure
A KYC system that treats every CKYC conflict as a rejection is not a compliance system — it is a risk-avoidance system that creates its own risk: the risk of rejecting creditworthy borrowers for bureaucratic data inconsistencies they did not create. The KYC Verification Agent AI treats conflict as information to be classified and resolved, not as a binary blocker. The 2% of genuine identity mismatches are caught and escalated rigorously precisely because the 98% of recoverable conflicts are not treated with the same alarm.