AI Agent Profile · LendingIQ · Bengaluru
KYC Verification Agent AI
DivisionLending Operations
Resume
What this agent does
The KYC Verification Agent AI performs the automated checks that underpin every customer's identity verification at LendingIQ — Aadhaar authentication via the UIDAI API, PAN validation against NSDL records, Aadhaar–PAN linkage status, CKYC registry matching, and cross-field consistency verification across all identity documents. It produces a structured KYC verification report that tells the onboarding team exactly what was checked, what was found, and whether the KYC file is complete, needs additional information, or cannot proceed. It does not accept a customer's KYC — that is a human-authorised act recorded in the CKYC registry.
Primary functions
Aadhaar / PAN Verification
Every individual applicant — synchronousInvoked when: applicant submits Aadhaar and PAN details during onboarding and provides OTP consent for Aadhaar authentication
- Authenticates the Aadhaar number via the UIDAI OTP API — the applicant receives an OTP on their registered mobile, enters it in the journey, and the agent verifies the OTP with UIDAI. A successful authentication confirms the Aadhaar number is valid and belongs to the person who received the OTP on the registered mobile number. It does not verify that the person using the device is the person whose Aadhaar was authenticated — that is the role of liveness detection in the Fraud Detection Agent AI.
- Validates the PAN against NSDL records: confirms the PAN is active (not deactivated or surrender-marked), retrieves the name as per PAN records, and checks whether the PAN is linked to the submitted Aadhaar number — which is a mandatory requirement under the Income Tax Act and a strong identity corroboration signal when the linkage is confirmed.
- Cross-checks the name on the Aadhaar demographic data against the name on the PAN record and against the name on any other submitted document — salary slip, bank statement, or company registration. Minor name variations (initials vs full name, spelling transliterations) are classified as minor discrepancies requiring human review; material name mismatches (different names entirely) are classified as a verification failure.
- Does not store Aadhaar biometric data. It uses the OTP-based authentication pathway, which does not involve fingerprint or iris data. The only Aadhaar data accessed is the demographic fields (name, date of birth, gender, address) returned by UIDAI on a successful OTP authentication — and only the verification outcome, not the full demographic data, is retained in the KYC record.
CKYC Match
Every applicant — existing KYC check before fresh collectionInvoked when: PAN number is available — CKYC check runs before requesting fresh KYC documents
- Queries the CKYC registry with the applicant's PAN to check whether a CKYC record already exists from a prior interaction with another financial institution. If a CKYC record exists, it retrieves the KYC status (KYC-compliant / KYC-non-compliant), the KYC date, and the institution that performed the original KYC — reducing the need to collect and verify documents that have already been verified by another RE.
- Compares the key demographic fields in the CKYC record (name, date of birth, address) against the applicant's current submission to detect material discrepancies — an address change since the original CKYC, a name change, or a date-of-birth mismatch. Minor discrepancies are flagged for the KYC officer to assess; material discrepancies require fresh verification regardless of CKYC status.
- Determines whether the existing CKYC record satisfies LendingIQ's KYC requirements for the applicable customer category and product — because different products may require different KYC levels (simplified / normal / enhanced) and an existing CKYC record at a lower level does not satisfy a higher-level requirement.
- If no CKYC record exists, the agent proceeds to fresh document collection and triggers the full verification suite. At the end of a successful fresh KYC, it produces the CKYC registration payload for the human KYC officer to submit to the CKYC registry — the submission itself is a human-authorised act.
Video KYC Outcome Integration
For applicants completing VKYC pathwayInvoked when: Video KYC Moderator AI has completed a VKYC session and the outcome needs to be integrated into the overall KYC verification report
- Reads the structured output from the Video KYC Moderator AI — session pass/refer verdict, compliance checklist outcome, face match result, liveness check outcome, and any moderator notes — and integrates it with the Aadhaar/PAN verification results and CKYC match results to produce a complete KYC file status.
- Checks that the VKYC session outcome satisfies the RBI V-CIP (Video based Customer Identification Process) requirements for the applicable customer category: that the OVD presented during the session matches the OVD on file, that the customer's face was visible throughout, that the session was conducted in a single sitting without interruption, and that the date and time watermark from the VKYC system confirms the session was live.
- Identifies any remaining KYC gaps that the VKYC session did not address — for example, a VKYC session that verified identity but did not collect the address proof separately required for the applicable KYC category. Flags these as open items before the KYC file can be considered complete.
Knowledge base
RBI KYC Master Direction (RAG)
Full KYC Master Direction with all amendments — OVD categories, simplified/normal/enhanced KYC tiers, CKYC obligations, Aadhaar authentication provisions, V-CIP requirements. Always retrieved live, never cached.
UIDAI / NSDL API Responses
Real-time authentication and validation results from UIDAI (Aadhaar OTP) and NSDL (PAN verification). Pulled at invocation per applicant — not stored beyond the KYC file record.
CKYC Registry
Central KYC record data retrieved per applicant by PAN. The starting point for every verification — checking what is already known before requesting fresh documents reduces customer friction and processing time.
Applicant Document Set
OVDs submitted by the applicant — Aadhaar card image, PAN card image, address proof, photograph. Provided at invocation; not stored beyond the KYC record retention period.
Video KYC Moderator AI Output
Structured VKYC session outcome — pass/refer, compliance checklist, face match result — for integration into the complete KYC verification report.
KYC Verification Knowledge
Pre-training knowledge of Indian KYC framework, Aadhaar ecosystem, CKYC architecture, V-CIP compliance requirements, and digital onboarding KYC practice up to knowledge cutoff.
Hard guardrails
Known limitations
Important Reads
Learn more about how to deploy KYC Verification Agent AI to your lending workflow.
- Use case #0001How KYC Verification AI processes 50,000 verifications per day without errorsA KYC verification that takes 4 minutes per applicant is acceptable at 200 applications a day and a crisis at 2,000. The KYC Verification Agent AI processes 50,000 Aadhaar, PAN, and CKYC checks daily — simultaneously, in parallel, without the batch delays, manual queues, or error rates that define human-scale KYC operations.Read article →
- Use case #0002CKYC matching: what KYC AI does when records conflictA 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.Read article →
- Use case #0003KYC rejection handling: how AI communicates reasons to borrowers clearlyA borrower who fails KYC has not necessarily failed to qualify for a loan — they may simply have a name spelling difference between their Aadhaar and PAN, a CKYC record that needs updating, or a document uploaded at insufficient resolution. The KYC Verification Agent AI communicates the specific, correctable reason for every KYC hold or rejection, in plain language, with a clear next step. "Your KYC could not be verified" is not a communication. It is a closed door.Read article →
