Embedded finance is the integration of financial products into non-financial platforms — a hospital offering a 0% EMI for a surgery, an EdTech platform offering a course loan at enrolment, or a D2C brand offering BNPL at checkout. The challenge is that the financial product must be explained, qualified, and approved within the natural flow of the customer's purchase journey — not as a separate banking application. AI voice is the highest-conversion channel for this because it combines the immediacy of real-time interaction with the scale of automation.

Kallix's embedded finance qualification flow:
1. **Trigger**: customer abandons checkout due to price, hospital admission clerk flags a patient for financing, or EdTech platform identifies a student who has viewed pricing but not enrolled.
2. **Outbound or inbound AI call**: within 15 minutes of the trigger, the AI reaches out: 'I'm calling from [Partner Platform] — I noticed you were looking at [product/course/procedure]. We have a financing option that may help. Do you have 5 minutes?'
3. **Eligibility check**: soft CIBIL pull (or bureau score from NBFC partner's bureau integration); income capture (verbal, with AA consent option for instant bank statement); FOIR calculation.
4. **Offer presentation**: if eligible, the AI presents the offer — interest-free BNPL (0% for 30/60/90 days), low-cost EMI (3/6/9/12 months), or longer-tenure product loan.
5. **KYC initiation**: for NTC customers or higher-value offers, V-CIP link sent via WhatsApp; for bureau-verified customers, Aadhaar OTP eKYC suffices.
6. **Lead handoff**: qualified and KYC-verified lead pushed to the NBFC/bank lending partner's LOS for credit decision — typically automated for under Rs 50,000, human-reviewed for higher values.

Kallix customers in the embedded finance stack report 52–64% lead-to-qualification rates (customers who complete the AI qualification call and meet eligibility criteria) and 28–34% qualification-to-disbursement conversion within the same day.

India's BNPL regulatory landscape changed significantly in August 2022, when RBI prohibited PPIs (wallets, prepaid cards) from being loaded with credit lines from banks or NBFCs. This effectively shut down the PPI-based BNPL model used by several fintech players. The compliant BNPL structure going forward is a direct loan from an NBFC or bank, which is then used to fund the purchase — the credit does not touch a PPI wallet.

Key RBI frameworks governing BNPL:

**RBI August 2022 PPI Circular**: Banks and NBFCs cannot load credit lines onto PPI instruments. BNPL structured as 'credit to PPI wallet, then used for purchases' is non-compliant. Kallix does not facilitate this structure — all leads are qualified for direct loan products (personal loan, product loan, consumer durable loan) from licensed entities.

**RBI Digital Lending Guidelines 2022 (DLG 2022)**:
- KFS delivery within 60 seconds of any loan offer
- LSP (Kallix) disclosure at first interaction: 'I'm calling on behalf of [NBFC/Bank Name] regarding a financing option'
- No auto-enhancement of credit limit without explicit consent
- 3-day cooling-off right stated at acceptance
- Grievance officer details available at any point
- First Loss Default Guarantee (FLDG) cap of 5% of portfolio — relevant to co-lending arrangements

**TRAI TCCCPR 2018**: BNPL outbound qualification calls classified as transactional (existing customer relationship with the platform triggering the call). For cold outreach to non-customers, promotional classification applies — DND restrictions must be respected.

**IRDAI regulations**: for BNPL products that include embedded insurance (e.g., purchase protection insurance at checkout), IRDAI regulations govern the insurance component. Kallix presents these as optional and separately priced per IRDAI Bima Sugam guidelines.

Kallix's compliance architecture for BNPL: every qualification call begins with LSP disclosure, KFS is dispatched within 60 seconds of any offer being presented, and the qualification flow never loads credit onto a PPI.

Speed is the defining constraint of BNPL credit assessment — the customer is mid-purchase-journey, and a 10-minute credit application form will drive them away. Kallix's BNPL qualification is designed for sub-3-minute credit decisioning while maintaining responsible lending standards.

BNPL credit assessment architecture:

**Soft bureau pull**: triggered from the customer's mobile number or PAN (if pre-collected by the platform). Soft pulls do not appear on the customer's credit report and do not affect their score. Bureau response time: 8–12 seconds (CIBIL API, Experian API). Score threshold is configurable — BNPL typically uses a lower floor (CIBIL 650+) vs personal loans (700+) because ticket sizes are smaller and risk is diversified across a portfolio.

**NTC handling**: if the bureau returns 'No Hit' (customer has no credit history), the AI does not disqualify — it transitions to the alternative scoring flow (covered in a separate question).

**Income capture** (verbal):
- 'What is your approximate monthly take-home income?' — single question, not a detailed financial intake
- For BNPL under Rs 25,000, verbal income capture alone suffices (no document required)
- For BNPL above Rs 25,000, Account Aggregator consent request sent for bank statement pull

**FOIR calculation**: existing EMIs (from bureau tradeline data) + proposed BNPL EMI / monthly income. FOIR threshold for BNPL: typically 50% (more lenient than personal loan 40–45% because BNPL tenure is shorter and EMIs are smaller).

**Active NPA check**: any active NPA on the bureau disqualifies for BNPL — the AI routes such customers to a 'credit improvement advisory' flow rather than a hard rejection.

**In-principle result**: 'Based on what you've shared, you appear eligible for [offer details]. To confirm this, we'll need a quick verification step — I'll send you a link to complete in 2 minutes.' No guarantee language — always 'appears eligible' or 'in principle eligible.'

India has an estimated 160–200 million creditworthy adults with no bureau history — largely young earners, rural customers, and homemakers. BNPL is the most accessible entry point to formal credit for this population, and AI-assisted alternative scoring is what makes responsible NTC lending viable at scale.

Kallix's NTC alternative scoring signals:

**UPI transaction data (highest weight)**:
Account Aggregator consent pulls the borrower's bank statement. Kallix's alternative scoring engine analyses:
- UPI transaction frequency: customers making 15+ UPI transactions/month show financial activity and digital literacy
- Average monthly balance over 12 months: stability signal
- Salary/regular credit consistency: monthly inward credit from a single source (employer) is an employment proxy
- Absence of returned debit entries: no bounced payments in the last 6 months

**Mobile recharge consistency** (for NTC customers without bank statement consent):
With user consent, mobile number age and recharge pattern (ARPU and frequency) provide a basic financial activity signal. Customers with 24+ months of consistent recharges above Rs 200/month have a statistically lower default rate.

**Employment verification**:
- The AI asks for employer name and either the employee ID or offer letter number
- For salaried NTC customers, the employer's HR portal or corporate email domain verification provides a strong income anchor
- For self-employed NTC customers: GST registration number provides business vintage and scale

**Rental payment history**:
Customers paying rent via bank transfer (not cash) have a 12-month rent payment history in their bank statement — the most reliable NTC credit signal (similar to a 12-month EMI payment history).

**NTC BNPL limit calibration**: NTC customers approved for BNPL receive a lower initial limit (typically Rs 5,000–15,000) with automatic limit enhancement after 3 months of on-time repayment — a responsible credit building ladder that Kallix explains during qualification.

**RBI DLG 2022 NTC disclosure**: the AI discloses to NTC customers that the BNPL is a credit product that will be reported to credit bureaus — helping them build their credit history if repaid on time.

V-CIP under RBI KYC Master Directions (amended 2021) is the primary KYC method for digital lending at scale — it eliminates the physical branch visit while meeting the 'face-to-face' equivalence requirement that some customer profiles require.

When is V-CIP required for BNPL?
- Aadhaar OTP eKYC: sufficient for most BNPL under Rs 50,000 and for customers with existing KYC with the same lender
- V-CIP: required for BNPL above Rs 50,000, for customers with PAN-Aadhaar linking issues, and for lenders whose RBI authorisation requires face-to-face equivalent for all credit products
- CKYC: if the customer already has a CKYC number (14-digit CKR from CERSAI), KYC is completed by fetching the CKYC record — no V-CIP needed

Kallix's V-CIP facilitation flow:
1. **Pre-V-CIP data collection**: the AI collects PAN number, Aadhaar number (last 4 digits only — full Aadhaar not captured per UIDAI guidelines), address, and occupation. This data is pre-loaded into the V-CIP system so the video officer does not need to re-collect.
2. **V-CIP booking**: 'I'll send you a link for a 5-minute video verification — it works on any smartphone. Would you prefer [time slot options]?' The booking is made in the NBFC/bank's V-CIP scheduling system.
3. **V-CIP session**: the Kallix AI does not conduct the V-CIP itself — V-CIP requires a live human official per RBI KYC Master Directions. The AI's role is pre-qualification and data preparation. Average V-CIP duration with Kallix pre-collection: 5–7 minutes vs 12–15 minutes without.
4. **Post V-CIP**: if V-CIP is successful, the AI triggers the loan offer and KFS dispatch. If the V-CIP attempt fails (technical issue, document mismatch), the AI reschedules within 24 hours.

**UIDAI consent handling**: Aadhaar OTP eKYC requires explicit consent before the OTP is sent. The AI collects this consent verbally and confirms: 'I'm about to send an OTP to your Aadhaar-linked mobile for identity verification. By proceeding, you consent to sharing your Aadhaar details with [Lender Name]. Do you confirm?'

Healthcare financing is one of the highest-impact embedded finance use cases — financial stress at the point of medical need leads to delayed treatment, worsened outcomes, and long-term debt burden. Kallix's healthcare financing qualification is designed around sensitivity (the patient is in stress) and speed (treatment cannot wait).

Healthcare embedded finance trigger scenarios:
- Hospital admission desk: patient's insurance claim is partially covered, leaving a balance above Rs 30,000
- Clinic/dental/ophthalmology: elective procedure estimate shared, patient asks about payment options
- Diagnostics/lab: high-cost scan or test package above Rs 10,000
- Oncology/cardiac specialist: treatment plan shared, patient asks how to fund it

Kallix healthcare qualification flow:
1. **Trigger**: hospital management software (Practo, eHospital, CliniQ) or receptionist flags the patient. AI call within 5 minutes of flag.
2. **Sensitivity calibration**: the AI acknowledges the context: 'I understand you're dealing with a medical situation right now. I'll keep this brief — I just want to make sure you have a financing option available if you need it.'
3. **Insurance coordination**: 'Has your insurance approved this treatment? What is the balance amount after insurance?' — pulls the out-of-pocket amount that drives the financing requirement.
4. **Quick eligibility**: soft bureau pull on PAN (provided during hospital registration) + verbal income capture.
5. **Offer**: 0% EMI for 3–6 months for treatments under Rs 1 lakh; longer-tenure medical loan for above Rs 1 lakh. Partner NBFC/bank pre-approved for the hospital's patient profile.
6. **Ayushman Bharat coordination**: for patients with Ayushman Bharat (PM-JAY) coverage, the AI checks whether the procedure is covered under the scheme before routing to financing — financing is offered only for the out-of-pocket balance not covered by PM-JAY.

Kallix healthcare financing partners are NBFCs with medical loan product expertise — such as Bajaj Finserv Health EMI Card, Tata Capital Healthcare Financing, and lender-specific medical products.

EdTech BNPL is India's fastest-growing embedded finance segment — driven by high course fees (Rs 50,000–5 lakh for professional certifications and upskilling programmes), young NTC student borrowers, and the strong correlation between course completion and income improvement. The qualification challenge is that students often have no income and no credit history at the point of purchase.

Kallix EdTech BNPL qualification architecture:

**Currently employed student** (upskilling, part-time course):
- Standard BNPL qualification: income capture + soft bureau pull + FOIR check
- Most common: 3-month 0% BNPL or 6-month low-cost EMI at 12% p.a.

**Pre-employment student** (full-time course, bootcamp, certification before job):
- Deferred repayment BNPL: repayment starts 3–6 months after course completion — aligned to expected employment start
- Offer letter as income proxy: for students who have a job offer letter (IT/BFSI/consulting), the CTC stated in the offer letter serves as the income anchor for EMI calculation
- Co-applicant path: parent or guardian qualifies as co-applicant — their income is used for FOIR; the student is primary borrower and CIBIL building begins
- Income Share Agreement (ISA) variant: some EdTech lenders offer ISA (a percentage of salary for a fixed period post-employment) — the AI captures interest in ISA and routes to the ISA-offering lender partner

**Course completion insurance**: for loans above Rs 50,000, some partners offer course completion insurance — if the student drops out due to illness or job loss, the insurer covers remaining EMIs. The AI presents this as an optional add-on.

**Platform integration**: Kallix integrates with EdTech LMS platforms (Coursera, upGrad, GreatLearning, BYJU'S) via API — when a student's enrolment intent is detected (course added to cart, pricing page viewed 3+ times), the qualification trigger fires.

**Affordability framing**: 'This Rs 1.8 lakh course costs Rs 5,000/month over 36 months — roughly the cost of 2 restaurant meals per month. And the average salary increment for this certification is Rs 4–6 lakh p.a. within 18 months, per [platform]'s placement data.' The AI frames the EMI in the context of the income return.

Co-lending has become the structural backbone of Indian embedded finance — it enables fintechs and NBFCs to offer bank-competitive interest rates by accessing the bank's lower cost of capital, while the NBFC retains customer ownership and the bank diversifies its loan book through the NBFC's sourcing engine.

RBI Co-Lending Model Guidelines (September 2020):
- NBFC takes minimum 20% of each loan on its books; bank takes up to 80%
- Single loan agreement with the borrower — the co-lending structure is not visible to the borrower
- Interest rate is blended: if NBFC rate is 18% and bank rate is 10%, the borrower pays ~11.6% on the blended portfolio
- Grievance responsibility: primary responsibility stays with the NBFC (entity with the customer relationship)

Kallix's role in the co-lending stack:
1. **Lead qualification**: the AI qualifies the borrower against the NBFC's eligibility criteria (which incorporates the co-lending bank partner's requirements). A borrower who qualifies for co-lending gets the blended rate; one who only qualifies for NBFC-only gets the NBFC rate — the AI presents the more favourable rate if co-lending eligibility is met.
2. **Co-lending eligibility check**: some co-lending arrangements have bank-specific additional requirements (minimum CIBIL score, minimum monthly income, no NBFC-only products). The AI checks these in real time.
3. **KFS for co-lending**: the KFS for a co-lending product must disclose the co-lending arrangement — 'This loan is co-originated by [NBFC Name] and [Bank Name] in a ratio of 20:80.' The KFS template is pre-configured per the specific co-lending agreement.
4. **Regulatory classification**: co-lending loans are classified under the bank's books for Priority Sector Lending (PSL) — education loans and MSME loans qualify. The AI captures the use-case information needed for PSL classification.
5. **FLDG limitation**: RBI DLG 2022 caps FLDG at 5% of outstanding. For co-lending, the NBFC's 20% at-risk position serves as the de facto first-loss — no separate FLDG required. The AI does not reference FLDG in customer conversations.

BNPL fraud rates in India are 2–4x higher than standard personal loan fraud rates — because BNPL's speed and low documentation requirements are exactly what fraudsters exploit. Kallix's BNPL qualification includes specific fraud detection logic.

Fraud signal detection during AI qualification:

**Account takeover signals**:
- The caller's mobile number does not match the registered mobile on the platform's customer record
- The caller cannot confirm basic identity details (name spelling, email domain, last purchase from the platform) that a genuine customer would know
- Multiple qualification attempts from the same mobile number for different PAN numbers within 24 hours

**Synthetic identity signals**:
- PAN number age vs reported age: a borrower claiming age 25 with a PAN registered only 3 months ago (PAN allotment date visible in bureau response) is suspicious
- Bureau 'thin file' with high stated income: a borrower with Rs 80,000/month income and only 1 small credit card tradeline is unusual — genuine high earners typically have multiple tradelines
- Address mismatch between Aadhaar and bureau record: if the addresses diverge significantly without a change-of-address event in the bureau, flag for manual review

**Social engineering signals**:
- Borrower is being coached: AI detects a second voice or prompting in the background
- Purchase category does not match borrower profile: a 65-year-old qualifying for a Rs 40,000 gaming laptop BNPL at a D2C platform — age-purchase incongruity flag
- Extreme urgency: 'Can you process this right now, I need it in the next 10 minutes' — unusual urgency in a purchase financing context is a social engineering indicator

**SIM swap detection**:
- Aadhaar OTP sent but not received: if the OTP is not entered within 60 seconds, and the borrower reports 'not received,' this may indicate a SIM swap. The AI pauses qualification and asks the borrower to verify their telecom provider's last activity.

**AI response to fraud signal**: the AI does not accuse the borrower of fraud. It says: 'I need to verify a few more details before we proceed. I'll escalate this to our verification team — they'll call you within 2 hours.' Qualification is paused pending human review.

The economics of BNPL in India are fundamentally driven by the merchant subsidy model for short-tenor offers — and the transition to interest-bearing products for longer tenors. Kallix's AI explains this clearly, which builds trust and reduces post-purchase disputes about unexpected charges.

Interest-free BNPL (0% MDR-subsidised):
- Tenure: 30, 60, or 90 days (sometimes called 'Pay Later' or 'Deferred Payment')
- Cost to borrower: Rs 0 — no interest, no processing fee
- Cost to merchant/platform: MDR of 2–4% of transaction value paid to the NBFC/bank partner
- AI presentation: 'This is completely free — you pay the exact purchase amount of Rs 8,400 in 90 days. No interest, no fees.'
- RBI compliance: still structured as a loan (not a PPI credit line) — KFS required; interest rate disclosed as 0%; cooling-off applies

Interest-bearing EMI conversion:
- Tenure: 3–24 months
- Rate: 14–24% p.a. (equivalent to 1.2–2% per month)
- Processing fee: 1–2% upfront (disclosed in KFS)
- AI presentation: 'Over 12 months at 16% p.a., your monthly payment is Rs 820. The total amount you pay is Rs 9,840 — Rs 1,440 more than the purchase price. If you can pay in 90 days for free, that's the better option financially.'

Kallix's AI always presents the comparison honestly — including stating when the interest-free option is financially superior. This counterintuitive transparency drives higher trust and repeat BNPL usage: borrowers who feel they were given honest advice are 2.4x more likely to use BNPL again within 90 days.

For borrowers who cannot pay in 90 days and choose EMI: the AI confirms whether prepayment is available (typically yes, at 0–2% charge) and how to do it — encouraging borrowers to close early if they receive a cash flow improvement.

E-commerce BNPL lead qualification is the highest-volume use case — major Indian e-commerce platforms (Flipkart, Meesho, Snapdeal, and thousands of D2C brands) process millions of abandoned carts daily, and BNPL conversion is the most effective recovery mechanism.

Integration architecture:

**Webhook triggers** (platform → Kallix):
- Cart abandonment with cart value above lender-configured threshold (typically Rs 1,500–2,000)
- Checkout page time > 180 seconds (pricing friction signal)
- COD order above Rs 3,000 (customer chose COD because they didn't want upfront payment — BNPL converts many COD orders to digital)
- Return visit to product page 3+ times (high purchase intent without conversion)

**Kallix outbound call** (within 15 minutes):
'Hi [First Name], I'm calling from [Platform Name]. I noticed you were looking at [Product Name] for Rs [cart value]. We have a 0% BNPL option where you can pay in 3 equal monthly instalments — no interest, no extra charges. Would you like to complete the purchase on these terms?'

**Real-time offer + checkout link**: if the customer agrees, the AI sends a checkout link via WhatsApp with the BNPL option pre-applied and the cart pre-loaded. This reduces the friction to a single tap — the customer doesn't need to re-navigate to the cart.

**D2C brand customisation**: for D2C brands (fashion, consumer electronics, beauty, fitness), Kallix configures brand-specific voice and language. The AI sounds like an extension of the brand's customer service team, not a generic lender call.

**COD to digital conversion**: for COD orders above the threshold, the AI calls after order confirmation: 'Your order has been placed via Cash on Delivery. Would you prefer to switch to a BNPL option? You'd avoid the need to keep cash on hand and get access to the product immediately without waiting for COD delivery confirmation.' COD-to-BNPL conversion rate: 18–26%.

**Privacy and consent**: the platform's customer relationship is the consent basis for the transactional call. The AI discloses the partnership: '[Platform Name] has partnered with [NBFC Name] to offer this option. I'm calling on behalf of [NBFC Name].'

Travel is the second-largest BNPL category in India after consumer electronics — holiday packages, international flights, and hotel bookings are high-ticket, emotional purchases where financing significantly improves conversion. The unique feature of travel BNPL is that the purchase precedes the service consumption — the borrower travels in October and can pay in November.

Travel BNPL qualification specifics:

**Standard flow** (checkout abandonment):
- Flight booking: price above Rs 8,000 abandonment trigger
- Hotel/resort: above Rs 5,000/night or above Rs 15,000 total booking
- Package: above Rs 25,000 per person

**Pay-after-travel structure** (0% for 30 days post-check-out):
AI script: 'You can book your trip to Goa now and pay after you return. No interest if you pay within 30 days of your check-out date. So for a trip in October, your payment is due in November.'

**Seasonal proactive campaigns**:
For Diwali (October) travel, the AI runs outbound campaigns in August–September to customers who booked Diwali travel in prior years: 'Planning your Diwali travel this year? We have BNPL for flights and hotels — book now while prices are lower, pay after Diwali.'

**International travel**: for international bookings above Rs 1 lakh, standard BNPL is complemented by a 6–12 month travel loan at 14–16% p.a. The AI explains the total cost vs the opportunity cost of delaying the trip.

**Cancellation interaction**: travel has higher cancellation rates than consumer goods. If a BNPL borrower cancels a trip, the cancellation refund is credited to the BNPL account, reducing the outstanding. The AI explains: 'If you cancel, any refund from the platform will be applied to your outstanding BNPL balance first, then the remainder is credited to your bank account.'

**Platform integration**: MakeMyTrip, Ixigo, Yatra, EaseMyTrip, and large OTAs have partner BNPL integration APIs. Kallix connects to these via the NBFC partner's platform API.

BNPL repayment collection is structurally different from home loan or personal loan EMI collection. The average BNPL ticket in India is Rs 3,000–25,000 — small enough that customers may simply have forgotten, not defaulted. The tone of outreach must reflect this: a customer who bought a Rs 5,000 gadget on BNPL needs a friendly reminder, not a collections call.

Day -3 pre-debit confirmation: the agent contacts the customer 3 days before the BNPL due date: 'Your BNPL payment of Rs [X] for your [product/platform] purchase is due on [date]. Your UPI auto-debit is set up — please ensure Rs [X] is available in your [bank] account linked to UPI ID [XXXX]. Shall I send you a reminder on WhatsApp on the due date morning as well?' This anticipatory confirmation reduces bounce rate by 28–36%.

Day 0 BNPL payment link: if the NACH/UPI debit fails, the agent calls within 2 hours and sends a one-click UPI payment link via WhatsApp. The link is pre-filled with the exact BNPL outstanding amount and the lender's UPI ID. 54–62% of customers who receive this link pay within 4 hours. The agent script is positive: 'Your auto-debit did not go through today — here is a quick payment link so you can complete it in 30 seconds and avoid any late charges.'

Day +2 overdue recovery: for customers who have not paid after Day 0 contact, the agent presents 2 options: (1) Pay the full outstanding now with the UPI link — no additional charge. (2) Request a 7-day payment extension — available once per customer for first-time overdue under RBI FPC framework. The extension option converts 22–28% of at-risk customers who would otherwise have entered formal delinquency. Late fee disclosure is mandatory: 'If the payment is not made within 7 days, a late charge of Rs [X] will apply.'

RBI FPC and DPDP compliance: BNPL collection calls must comply with RBI Fair Practices Code — no calls before 8 AM or after 7 PM, no threats, no contact with family or employer without borrower consent. DPDP limits purpose of contact to the customer's own overdue account — no using the collection call to cross-sell other products.

The Account Aggregator ecosystem is the most significant infrastructure development for embedded finance in India since UPI. It creates a consent-first, standardised data-sharing framework where a customer can authorise a lender to access their financial data from any AA-registered Financial Information Provider (FIP) — banks, mutual fund RTAs, insurance companies, pension funds, and GST portal.

AA consent flow in BNPL qualification: the AI agent initiates the AA consent request during the qualification call: 'To assess your BNPL limit without requiring documents, I will send you a consent request on your registered mobile number — it takes 90 seconds to approve on your bank app. This allows us to review your account activity and make a credit decision in real time. Shall I send it now?' The consent request is sent to the customer's AA app (Setu, Onemoney, Finvu, NADL — RBI-licensed AAs). Once the customer approves, the lender FIU (Financial Information User) receives the customer's financial data in a standardised format.

Data accessed via AA for BNPL: (1) Bank account transaction history (12–24 months) — for cash flow analysis, salary verification, and recurring expense identification. (2) Mutual fund holdings (from CAMS/KFintech via AA) — for net worth assessment and repayment capacity beyond income. (3) GST data (for self-employed BNPL customers) — annual turnover, filing regularity, and business cash flow. (4) NPS balance — for HNI BNPL customers. AA data access is purpose-limited: the FIU can only use data for the consented purpose (BNPL credit assessment) and cannot store or reuse it for other purposes under RBI AA Master Direction 2021.

NTC customer benefit: for customers with no CIBIL history (NTC — new-to-credit), AA cash flow data enables credit decisioning that bureau-only models cannot make. A 28-year-old salaried customer in Tier 2 city with no prior loans but 2 years of regular Rs 35,000 salary credits and low average monthly debit (50% of income) is a strong BNPL credit risk — but has CIBIL score 0 or -1. AA data enables this customer to access BNPL credit they would otherwise be declined for.

Consent withdrawal: the DPDP Act 2023 and RBI AA framework both give customers the right to withdraw AA consent at any time. The lender retains access to data already shared before withdrawal but cannot request new data pulls after withdrawal.

BNPL and credit cards serve overlapping but distinct customer segments. Understanding the differences — and knowing which product fits which customer — is a critical qualification skill that the AI agent exercises before routing to the appropriate product.

Key product differences: (1) Eligibility: credit cards require CIBIL 720+ at most banks; BNPL qualifies at 650+ (bureau-based) or uses alternative data (AA/GST) for NTC. (2) Approval speed: credit card — 3–7 business days + physical card delivery 7–10 days; BNPL — 4–8 minutes in-app or post-AI-qualification call. (3) Interest rate: credit card revolving balance 36–42% APR; BNPL 0% for 15–30 day interest-free period, 18–24% for EMI conversion. (4) Credit limit: credit card Rs 50,000–5 lakh; BNPL Rs 5,000–2 lakh per transaction (dynamic per transaction). (5) Bureau impact: credit card is a revolving credit line — high utilisation damages CIBIL; BNPL reported as consumer durable loan — less impact on revolving utilisation ratio.

AI qualification routing: customers with CIBIL 720+ applying for a purchase above Rs 25,000 → credit card with BNPL as fallback. Customers with CIBIL 650–719 → BNPL as primary product, credit card for future consideration after 12-month BNPL history. Customers with CIBIL below 650 or NTC → BNPL with AA-based assessment as sole option.

BNPL as credit-building tool: the AI agent positions BNPL for NTC customers as a credit-building step — 'Completing your BNPL repayments on time will be reported to CIBIL and can help you build a credit history. Many customers use BNPL for 12–18 months before applying for a credit card.' This credit-ladder framing converts 38–46% of NTC BNPL leads who were originally unsure about taking credit.

Regulatory note: RBI's digital lending guidelines (2022) and BNPL classification circular require BNPL lenders to report all BNPL accounts to credit bureaus. This bureau reporting is both a risk management tool and a customer benefit — customers build credit history with responsible BNPL use.

Merchant activation is the supply side of BNPL — without a merchant network offering BNPL at checkout, there is no embedded finance product to qualify customers for. BNPL merchant acquisition outreach is a B2B qualification call with a different decision maker (merchant owner/finance manager) and different qualification criteria than a consumer BNPL call.

Merchant qualification questions: (1) Monthly transaction volume: 'What is your approximate monthly sales volume — above Rs 5 lakh?' Below Rs 2 lakh/month, the BNPL economics may not justify integration cost. (2) Average transaction size: BNPL is most valuable for transactions above Rs 2,000 — below this, customers rarely need credit. (3) Customer demographics: 'What is the typical age range of your customers?' Younger demographics (22–38) have higher BNPL adoption. (4) Current payment methods: 'Do you currently accept EMI payments on any card?' Merchants already accepting credit card EMI are warm prospects — they understand EMI demand. (5) Technical capability: 'Do you have a POS system or website?' This determines integration method.

Integration options: (a) QR code BNPL (offline): the merchant displays a BNPL QR code at POS — customer scans, completes qualification in-app in 4–8 minutes, merchant receives confirmation before customer leaves. No technical integration needed. (b) WhatsApp BNPL link (service businesses): the merchant sends a WhatsApp payment link with BNPL option — used by clinics, coaching centres, and salons. (c) API integration (e-commerce): webhook integration into the checkout flow — OTP-triggered BNPL option at cart stage.

MDR (Merchant Discount Rate): BNPL MDR ranges from 0% (for BNPL-as-marketing — merchant subsidises the BNPL to drive higher cart values) to 2% (standard). The agent explains: 'Your MDR is 1.2% — for every Rs 10,000 BNPL transaction, you pay Rs 120 to us. In exchange, customers who cannot afford Rs 10,000 upfront can complete the purchase — increasing your average order value by an estimated 28–34% based on our merchant cohort data.'

Merchant activation campaign: BNPL lenders generate merchant leads from 4 sources: existing bank SME loan customers, DSA networks, marketplace seller lists, and outbound prospecting to offline retail clusters. The AI agent qualifies leads, books a demo, and follows up on demo-to-onboarding conversion.

BNPL credit limit management is one of the most impactful retention and monetisation levers in embedded finance. A customer whose Rs 15,000 BNPL limit becomes insufficient as their purchasing power grows will either request a limit increase or switch to a credit card. Proactive limit increase outreach captures this inflection point before the competitor does.

Limit increase eligibility triggers: (1) Repayment track record: 3 consecutive on-time repayments with zero bounce = automatic limit review flag. The agent calls: 'You have completed 3 BNPL repayments on time — congratulations. You are now eligible for a limit increase from Rs [current] to Rs [proposed]. Would you like to activate the higher limit?' The in-call activation takes 60 seconds with an OTP confirmation. (2) CIBIL improvement: if the customer's CIBIL score (pulled via a monthly soft check with consent) has improved by 50+ points since the last limit assignment, the agent flags a review: 'Your credit score has improved by [X] points — this qualifies you for a higher BNPL limit at a better interest rate for any EMI conversions.' (3) AA income increase: if salary credits via AA monitoring show a 15%+ increase (new job or increment), the agent calls for a limit review: 'We noticed your income has increased — would you like us to review your BNPL limit to match your current profile?'

Limit reduction management: if a customer's CIBIL drops or repayment behaviour deteriorates, the system flags for limit reduction. The agent communicates this proactively (not reactively) before the customer attempts a transaction that gets declined: 'We have reviewed your BNPL account and adjusted your limit to Rs [new] — this is to ensure your repayments remain manageable. Your limit can be reviewed upward in 3 months.' This proactive communication prevents the complaint and confusion of an unexpected transaction decline.

Cross-sell from limit review: customers who accept a limit increase are warm prospects for a credit card, personal loan, or full digital bank account. The agent presents this after limit increase confirmation: 'With your BNPL history, you now pre-qualify for a credit card with a Rs [X] limit — would you like a quick overview?'

The BNPL opportunity in Tier 2/3 India is larger than metro BNPL — but requires a fundamentally different product design and qualification approach. Tier 2/3 customers are often first-time credit users, may not have smartphones, and transact in vernacular. The standard app-based BNPL funnel fails for this segment.

Language-first qualification: the AI agent detects language from the first response and switches to Hindi, Tamil, Marathi, Bengali, or other regional languages immediately. Hindi BNPL qualification: 'Aap Rs [X] ka [product] BNPL ke zariye khareed sakte hain — aapko abhi kuch bhi nahi dena, [tenure] mein barabar hissa kar ke chukana hoga.' (You can purchase Rs [X] [product] via BNPL — pay nothing now, split equally over [tenure].) Language-matched explanation converts 2.4× better than English in Tier 2/3 markets.

Feature phone BNPL via UPI 123Pay: customers without smartphones can initiate BNPL via UPI 123Pay (*99# USSD) — the AI agent walks the customer through the USSD-based UPI consent and mandate registration step by step in the local language. This feature phone BNPL channel serves the estimated 250 million Indians without smartphones who are BNPL-eligible by income but excluded by app-first design.

Kirana-embedded BNPL: the agent onboards kirana store owners to offer BNPL via a printed QR code — customers scan with any UPI app. The kirana owner gets a daily settlement, the customer gets 30-day credit. Kirana BNPL typical ticket: Rs 500–5,000 (weekly grocery + household). The agent explains to the kirana owner: 'Your customers can buy now and pay in 30 days — you get the money in 24 hours, we handle the credit risk.'

DAY-NRLM SHG BNPL: women SHG members with NRLM-linked bank accounts and 18+ months of regular saving and repayment (Grade A/B SHGs) are pre-qualified for group-guarantee BNPL of Rs 5,000–50,000 for productive assets (sewing machines, agricultural equipment, mobile phones for business). The AI agent outreach to SHG leaders in local languages achieves 68–74% group qualification rates.

BNPL delinquency management differs from personal loan or home loan collection in 3 fundamental ways: smaller ticket size (Rs 3,000–25,000 means customer psychology is different), shorter repayment tenure (15–90 days, so DPD escalates faster), and younger customer demographic (first-time credit users who need education, not just pressure).

DPD 1–7 bucket: soft recovery. The agent uses the same friendly tone as a reminder call: 'We noticed your BNPL payment of Rs [X] due on [date] has not been processed — are you facing any issue? Here is a quick UPI payment link to clear it today and avoid any late charge.' No threat, no urgency pressure. 62–72% of DPD 1–7 customers resolve within the same bucket with this approach.

DPD 8–30 bucket: structured resolution. Two options presented: (1) 7-day free extension — available once, converts 22–28% who need short additional time. (2) EMI conversion: 'Would you like to convert this BNPL outstanding of Rs [X] into 2 monthly EMIs of Rs [Y]? There is a Rs [Z] conversion charge but no further late fees.' EMI conversion converts 28–36% of DPD 8–30 customers and creates a structured repayment record that protects CIBIL score.

DPD 31–60 bucket: escalation-adjacent communication. The tone shifts: 'Your BNPL account of Rs [X] is now [Y] days overdue — this will be reported to credit bureaus as overdue on your next reporting date of [date]. To avoid a negative CIBIL entry, please make payment or contact us for a settlement today.' Legal mention is introduced only if the account qualifies for legal action under the lender's policy. Settlement offers (15–25% waiver on penal charges) convert 28–36% of DPD 31–60 accounts.

PMLA and data compliance: overdue BNPL accounts cannot be used to solicit other products. The collection call is single-purpose — recover the overdue amount. DPDP purpose limitation applies to collection data as well.

Grocery and utility BNPL addresses the most consistent financial stress point for lower-middle income India: the end-of-month cash crunch before the next salary credit. A household spending Rs 8,000/month on groceries that runs short by Rs 2,000 in the last week of the month is the precise use case that grocery BNPL was designed for.

Product mechanics: the grocery BNPL limit (Rs 2,000–10,000) is a revolving micro-credit facility — the customer can make multiple grocery purchases through the month, aggregated against the limit. On the 1st of the following month (or salary credit date if linked via AA), the full outstanding is auto-debited. No EMI — it is a 30-day interest-free facility. If repayment fails, a 30-day extension is available at Rs 15–50 late charge (vs Rs 500–1,200 for formal loan late charges).

Kirana integration via ONDC: the Open Network for Digital Commerce (ONDC) has enabled thousands of kirana stores to accept digital payments and offer embedded financial products. BNPL lenders integrated with ONDC can offer grocery BNPL at any ONDC-linked kirana with zero merchant integration cost. The AI agent activates grocery BNPL for customers during kirana visit recovery calls: 'You visited [kirana name] and did not complete your purchase — did you run short? You are pre-qualified for a Rs 5,000 grocery BNPL limit that resets every month. Want me to activate it for your next visit?'

Utility payment BNPL: BBPS (Bharat Bill Payment System) handles 9 billion+ utility payment transactions annually. BNPL for utility payments (electricity, gas, water, DTH, mobile recharge) prevents service disconnection for short-term cash-strapped customers. Kallix runs utility BNPL qualification during BBPS payment failure events: 'Your electricity bill of Rs [X] failed — you can pay it now via BNPL and repay next month. Want me to process it?' This in-moment qualification converts 62–74% of utility payment failure contacts.

Salary linkage auto-repayment: the most efficient grocery BNPL design auto-debits the outstanding on salary credit date (detected via AA monitoring of the linked bank account). The agent explains this at qualification: 'Your outstanding resets automatically when your salary credits — you never need to manually initiate repayment.'

Insurance bundling in BNPL serves two commercial objectives simultaneously: it protects the lender against involuntary default and provides genuine financial protection to the customer. Unlike credit card PPI (Payment Protection Insurance) which was mis-sold extensively in the UK and banned, BNPL credit shield in India operates under IRDAI POSP regulations with mandatory separate consent and transparent pricing.

Credit shield mechanics: the customer pays Rs 15–80/month (depending on BNPL limit and tenure) for a credit shield policy that covers: (a) Death — outstanding BNPL balance paid by insurer; (b) Hospitalisation above 3 days — EMI payment waived for the hospitalisation period; (c) Job loss (involuntary) — 2 BNPL cycles waived (conditions: EPFO-registered employment, documented retrenchment). Premium is disclosed in-call before consent: 'The credit shield cover adds Rs 35/month — in exchange, if you are hospitalised for more than 3 days, your BNPL payment is automatically waived for that month. Would you like to add this?'

Purchase protection insurance: for BNPL on consumer electronics, the AI agent offers purchase protection during qualification: 'Your Rs 18,000 phone purchased via BNPL can be covered against accidental damage and theft for 12 months for Rs 120. Would you like to add this?' Purchase protection is particularly resonant for first-time smartphone buyers in Tier 2/3 markets where repair costs are proportionally higher relative to income.

IRDAI POSP compliance: the BNPL AI agent must be registered as or operate under a POSP (Point of Sales Person) to solicit insurance. The agent identifies the insurance opportunity and presents it as optional — mandatory bundling (forcing insurance purchase as condition of BNPL approval) is prohibited under IRDAI Policyholder Protection Circular. Separate consent and premium disclosure are required.

NPA reduction impact: lenders with credit shield bundling at 60%+ penetration report 12–18% lower effective NPA rates for bundled accounts vs non-bundled. This is primarily driven by hospitalisation coverage — medical events are the #1 cause of short-term BNPL default among 25–45-year-old borrowers.

Earned Wage Access is one of the fastest-growing embedded finance products in India — positioned between BNPL (credit) and payroll (employer) as a real-time liquidity tool. The India EWA market is estimated at Rs 12,000 crore annually (Omidyar Network India, 2023), with primary use cases: medical emergency (42%), utility bill payment (28%), education fee (18%), and personal emergency (12%).

EWA vs salary advance loan: traditional salary advance loans charge 2–3% per month (24–36% APR) and require CIBIL checks and employer letters. EWA charges a flat service fee of Rs 25–99 per advance regardless of amount — effectively 0% interest. The distinction matters for RBI classification: EWA is not a loan if the employer is the counterparty (employer settles from payroll), but becomes a regulated lending product if a third-party NBFC is the funder (which is the case for most fintech EWA players operating without direct employer partnership).

Qualification flow: Step 1 — Employer check: 'Which company do you work for?' The system checks whether the employer is integrated with the EWA platform (EPFO-registered companies are detectable; payroll API integration is required for real-time salary data). Step 2 — Salary and payday: 'What is your monthly take-home and on which date does your salary credit?' Maximum advance = 50% of monthly take-home × (days worked / total working days in month). Step 3 — Advance amount: 'How much do you need — between Rs [min] and Rs [max]?' The maximum is calculated in real time.

Disbursement: approved EWA is credited to the customer's salary account via IMPS within 3–5 minutes. Repayment is automatic — deducted from the next payroll run (employer-integrated model) or auto-debited on salary date (NBFC model via NACH). No late fees for salary date repayment — the advance is designed to have no default risk since salary credit is the repayment source.

Cross-sell from EWA: EWA customers with 3+ usage cycles are warm prospects for personal loans, credit cards, and insurance. The agent triggers a cross-sell after the 3rd EWA disbursement: 'You have used salary advance 3 times in the last 3 months — you may benefit from a personal loan of Rs 50,000 at 12% per annum, which would be more cost-effective for larger needs. Would you like to check your eligibility?'

B2B BNPL addresses the working capital gap at the heart of Indian MSME finance — the 30–90 day credit cycle between purchasing inventory and receiving payment from buyers. India has Rs 30 lakh crore in SME trade credit outstanding, most of it informal (verbal credit terms with no documentation). B2B BNPL formalises this credit, creates a bureau-reported history, and provides the MSME with a formal credit facility that can grow over time.

Qualification criteria: (1) GST turnover — the primary income proxy for B2B BNPL underwriting. Businesses with Rs 25 lakh+ annual GST turnover and regular filing history qualify for B2B BNPL up to 10% of annual turnover. (2) Buyer-seller relationship — B2B BNPL is not open credit; it requires a documented buyer-seller transaction history. The agent asks: 'How long have you been purchasing from [supplier name] — and what is your typical order size?' (3) Invoice verification — the BNPL is disbursed against a specific supplier invoice; the agent collects the invoice number and amount. The BNPL lender pays the supplier directly; the buyer repays the lender in 30/45/60/90 days. (4) GSTIN verification — the agent confirms the buyer's GSTIN is active and E-invoicing compliant (mandatory for turnover above Rs 5 crore under GST e-invoicing rules).

Anchor-led model: large companies (anchor buyers) with 100–5,000 SME suppliers onboard their entire supplier network onto B2B BNPL. The anchor provides the supplier list and transaction history — dramatically reducing underwriting risk. Kallix qualifies and onboards anchor-nominated suppliers via outbound calls: 'Your buyer [company name] has recommended you for their supplier financing programme — you can receive payment within 24 hours of invoice submission instead of waiting 60 days. Shall I explain how?'

RBI supply chain finance regulations: RBI's Trade Receivables Discounting System (TReDS) is a regulatory framework for supply chain finance for MSMEs. B2B BNPL can be structured as invoice discounting on TReDS (factoring) or as a buyer-side credit facility. TReDS-registered transactions qualify for priority sector lending classification, improving lender economics.

Interest rate: B2B BNPL rates range from 12–18% per annum for 30–90 day tenor. For anchor-led programmes, the anchor may subsidise the interest rate partially — effectively offering 0% or subsidised B2B BNPL to preferred suppliers as a supply chain stickiness tool.

The BNPL-to-formal-credit graduation path is one of the most commercially significant opportunities in Indian embedded finance — it turns a Rs 10,000 BNPL product into a Rs 2–5 lakh personal loan customer relationship. This graduation path is also the reason BNPL is structurally valuable for lenders beyond immediate BNPL economics: it builds a creditworthy customer base that feeds higher-margin products.

Graduation milestones and cross-sell triggers: (1) 3-month milestone: CIBIL score has likely improved by 20–40 points. Agent call: 'You have completed 3 BNPL repayments on time — your credit profile is improving. You are now pre-qualified for a credit card with a Rs [X] limit. Would you like to check the features?' Credit card cross-sell at 3-month milestone converts 28–36% — customers are engaged and creditworthy but not yet saturated with credit products. (2) 6-month milestone: 6 repayments create a meaningful CIBIL history. Agent call: 'Based on your BNPL track record over 6 months, you now pre-qualify for a personal loan of up to Rs [X] at [rate]%. This could be useful for a larger purchase or emergency fund. Would you like to see your eligibility?' BNPL-to-personal-loan at 6 months converts 34–44%. (3) 12-month milestone: full bureau history established. Agent presents: personal loan, credit card (if not already taken), and savings account (if BNPL customer banks elsewhere) — full product suite introduction.

Cross-sell sequence design: the cross-sell is sequenced to avoid product overwhelm. Only one product is introduced per call — the most relevant based on the customer's usage pattern. A customer who used BNPL primarily for healthcare expenses at 6 months is offered health insurance, not a personal loan. A customer who used BNPL for electronics is offered purchase protection + credit card (for future electronics purchases).

CIBIL improvement reporting: the agent discloses CIBIL improvement to the customer as a benefit of BNPL repayment: 'Your CIBIL score has improved from [start] to [current] over your 6 months of BNPL usage — this improvement is what enables you to qualify for products you may not have been eligible for before.' This disclosure builds trust and reinforces the value of the BNPL relationship beyond the individual purchase.

Subscription BNPL solves a specific affordability friction: customers who can afford a service on a monthly basis but face a cash flow obstacle when annual or quarterly billing concentrates the cost into a single payment. A customer who spends Rs 500/month on OTT entertainment has no trouble with the ongoing cost — but when presented with a Rs 6,000 annual renewal bill, cash flow constraints may cause them to cancel.

OTT subscription BNPL: the trigger is the annual plan renewal notification. The agent calls: 'Your [OTT platform] annual plan of Rs 5,999 renews on [date]. Instead of paying Rs 5,999 at once, you can convert this to 6 monthly payments of Rs 1,050 — the Rs 300 additional charge covers the EMI cost. Shall I convert it?' 44–52% of OTT annual renewal contacts who are price-sensitive convert to subscription BNPL at this offer.

Gym and fitness membership BNPL: gym memberships are a high-ticket annual commitment — Rs 15,000–50,000/year at premium gyms. Payment failure at renewal is the #1 reason for gym membership cancellation. The agent calls at renewal: 'Your [gym name] annual membership of Rs 24,000 renews on [date]. We can split this into 12 monthly payments of Rs 2,150. Would you like to continue your membership this way?' Gym BNPL conversion: 48–58% of contacted at-risk renewals.

School/coaching fee BNPL: education fee BNPL is the highest-ticket subscription BNPL category — school fees Rs 15,000–2 lakh per term, coaching centre Rs 20,000–1.5 lakh per course. The agent targets parents at fee payment deadline: 'Your child's school fees of Rs [X] are due by [date]. We can process payment today and you can repay in [3/6/12] instalments — your child's seat is secured while you manage cash flow.' Fee deadline urgency + child's education stakes = 52–64% BNPL conversion.

Recurring subscription management: unlike one-time BNPL, subscription BNPL is a recurring facility — the same customer renews every year with the same provider. After the first year of on-time subscription BNPL repayment, the renewal is pre-approved: 'Your [OTT] BNPL renewal is pre-approved for this year — shall I activate it with the same 6-EMI structure?' Pre-approved renewal converts 72–82% vs 44–54% for first-time subscription BNPL.

BNPL regulatory compliance in India spans 4 overlapping frameworks — each governing a different aspect of the customer acquisition and credit lifecycle. Non-compliance in any single framework can result in regulatory action, customer complaints, or call channel shutdown.

TRAI TCCCPR 2018 for BNPL outbound: cold BNPL prospecting (calling non-customers to offer BNPL) is a Promotional category communication — requires DLT registration, template approval, and DND scrub. The PE (Principal Entity — the BNPL lender) and TE (Telemarketing Entity — Kallix) must both be registered on the DLT platform. DND scrub frequency: daily for high-volume campaigns, minimum weekly for moderate campaigns. Transactional BNPL calls (existing BNPL customer EMI reminder, limit increase notification, repayment follow-up) are DND-exempt under Transactional classification.

RBI Digital Lending Guidelines 2022: all BNPL products funded by an NBFC or bank must comply with RBI's Digital Lending Framework: (a) KFS (Key Fact Statement) must be provided to the borrower before disbursement — including APR, total cost of credit, and all fees. (b) Loan Service Provider (LSP) liability: Kallix as an LSP (lead generator and qualification platform) operates under the lender's licence — any regulatory violation by Kallix in the qualification process is the lender's regulatory liability. (c) Cooling-off period: borrowers must be given a minimum 3-day cooling-off period to cancel a BNPL disbursement without penalty — the agent discloses this at qualification. (d) Recovery agent code of conduct: all post-disbursement collection calls must comply with RBI's code of conduct for recovery agents — no abusive language, no calls before 8 AM or after 7 PM.

DPDP Act 2023 for BNPL qualification: credit bureau soft pull requires explicit consent before the check — the agent must state: 'I will now check your credit score — this does not affect your CIBIL score and requires your permission. Do you consent?' IVR keypress captures consent for DPDP compliance. AA consent flow under the AA framework already has consent capture built in — no additional DPDP consent needed for AA data access.

Payment link delivery and UPI mandate registration are the last mile of BNPL qualification — the customer has been qualified, the limit is approved, but disbursement requires completing the payment setup. Friction at this stage is the most preventable cause of BNPL drop-off post-qualification.

WhatsApp deeplink delivery: the WhatsApp payment link is a deeplink to the lender's app or web checkout — pre-filled with the customer's name, approved limit, and first purchase details. The customer taps the link and lands directly on the BNPL activation page with their information pre-populated. Zero-entry experience: the customer only needs to confirm their Aadhaar-linked mobile OTP. 72–84% open rate within 5 minutes of sending; 48–56% completion within 15 minutes of link receipt.

UPI NACH mandate registration: the most important technical step in BNPL setup is the NACH (National Automated Clearing House) mandate that enables automatic EMI debit on due dates. NPCI's UPI-NACH allows customers to approve a mandate directly through their UPI app (GPay, PhonePe, Paytm, BHIM) without visiting a bank. The agent sends the mandate request during the call: 'I am sending a UPI mandate request to your registered UPI ID [XXXX@bank] now — please check your UPI app, review the mandate details, and enter your UPI PIN to approve. This sets up your auto-debit for EMI payments.' The customer approves in 45–60 seconds.

Same-session vs post-call mandate: mandates approved during the call (agent waiting on the line while customer approves) have 58–68% same-session completion. Mandates sent after the call (customer asked to complete on their own) have 22–28% completion within 24 hours and 38–46% within 7 days. The 30–40% difference in completion rate is the commercial case for keeping the customer on the line through mandate approval.

VPA (Virtual Payment Address) verification: before sending the mandate, the agent verifies the customer's UPI VPA (e.g., name@oksbi) by initiating a Re 1 penny drop — confirming the UPI ID is active and linked to the correct bank account. This prevents mandate setup on an incorrect UPI ID and reduces future EMI bounce rate by 18–24%.

Bureau reporting transforms BNPL from a transactional product into a credit-building instrument. For the lender, bureau reporting reduces adverse selection — customers who know their BNPL behaviour is reported are more likely to repay on time. For the customer, responsible BNPL use creates a credit history that opens access to mainstream financial products.

Bureau reporting mechanics: BNPL lenders submit monthly credit data to all 4 licensed credit bureaus under RBI's CIR (Credit Information Report) format. Data submitted: account opening date, credit limit, outstanding balance, repayment history (on-time, days past due), and account status (active/closed/NPA). The first bureau submission occurs within 30 days of disbursement. Credit score impact is visible on the customer's CIBIL report within 45–60 days of first submission.

Positive reporting impact: for NTC customers, the first BNPL account typically creates a 'thin file' in CIBIL — a score between 0 and 300 with limited data. After 3 on-time repayments: score typically 600–640. After 12 on-time repayments: score typically 680–720. After 24 months of clean history: score 720–760, qualifying for most mainstream credit products. The agent communicates this credit-building journey at BNPL activation: 'Every on-time repayment you make is reported to CIBIL — in 12 months, your credit score could qualify you for a personal loan or credit card at a mainstream bank rate.'

Negative reporting impact: a single missed BNPL payment drops CIBIL by 30–60 points — more severe as a percentage for thin-file NTC customers than for established credit users. The agent uses this as a repayment motivation tool without being threatening: 'Your BNPL repayment history is reported to CIBIL — on-time payments build your score, missed payments reduce it. Given that this is your first credit account, on-time repayment will be especially impactful for your future borrowing options.'

Bureau dispute process: if a customer's BNPL repayment is reported incorrectly (e.g., a payment was made on time but not updated in the bureau), the agent guides the dispute: 'You can raise a bureau dispute on the CIBIL website using your payment proof — the lender has 30 days to respond under RBI bureau dispute rules. I can also raise an escalation from our end to correct the reporting.'

BNPL affordability assessment prevents the primary driver of BNPL delinquency: customers who accumulate multiple BNPL accounts across platforms (Amazon Pay Later, Flipkart Pay Later, Simpl, LazyPay, platform BNPL) until total monthly obligations exceed repayment capacity. Multi-BNPL stacking is the #1 risk factor in BNPL NPA — and FOIR-based qualification is the primary control.

FOIR calculation in BNPL qualification: Step 1 — Income confirmation: the agent asks for monthly take-home (salaried) or last month's net income (self-employed). For AA-integrated qualification, salary credits from bank statement confirm income without relying on customer-stated figures. Step 2 — Existing obligation check: the agent asks for existing EMIs (home loan, personal loan, vehicle loan, credit card minimum payment). For bureau-integrated qualification, existing liabilities are pulled from the CIBIL report via soft check. Step 3 — BNPL obligation across platforms: the customer is asked about existing BNPL accounts at other platforms. Platform aggregation (via AA — some AA FIPs report BNPL accounts) may capture this automatically. Step 4 — Maximum BNPL limit = (50% of income - existing obligations) × tenure months.

Multi-BNPL stacking detection: the single most important risk control in BNPL underwriting. A customer earning Rs 25,000/month with Amazon Pay Later (Rs 2,000/month), LazyPay (Rs 1,500/month), and Flipkart Pay Later (Rs 1,500/month) has Rs 5,000/month in BNPL obligations before applying for a new BNPL product. Their residual FOIR headroom is Rs 7,500 (50% × Rs 25,000 - Rs 5,000) — not Rs 12,500 as a single-platform assessment would suggest.

Affordability script: 'To ensure the repayments fit comfortably within your income, I need to ask: what is your monthly take-home after all deductions? And do you currently have any EMIs — home loan, personal loan, or vehicle loan? Do you already use BNPL or pay-later on any other platform?' This 3-question sequence is framed as borrower protection, not interrogation — 72–82% of customers answer all 3 questions willingly when framed correctly.

Self-employed FOIR: for self-employed customers without a fixed monthly income, the agent uses a 3-month average of the customer's stated income (or AA-verified bank credits) to calculate FOIR. The FOIR ceiling is lower for self-employed — 40% vs 50% for salaried — to account for income variability.

The ROI case for AI BNPL lead qualification is driven by three compounding effects: speed (qualification in 4–8 minutes vs 20–40 minutes manual), reach (unlimited simultaneous outbound calls vs human agent headcount), and conversion (personalised, context-aware outreach beats generic SMS/email recovery).

**Lead economics**:
- Manual or DSA qualification: Rs 800–1,400 per qualified-and-disbursed BNPL customer (agent time + overhead + drop-off cost)
- AI-assisted: Rs 180–320 per qualified-and-disbursed customer
- Saving: Rs 480–1,080 per disbursement

**Volume capacity**:
- Human agent: 40–60 BNPL qualification calls/day
- AI: 5,000–50,000 qualification calls/day (platform-dependent)
- For seasonal peaks (Diwali, New Year, back-to-school), AI scales instantly without hiring

**Platform-specific conversion**:
- E-commerce cart recovery: 22–30% BNPL conversion vs 8–12% email/SMS
- Healthcare: 45–55% of patients offered financing accept it (vs 20–25% when offered at counter by human)
- EdTech: 38–44% course BNPL conversion vs 22–28% without proactive outreach
- Travel: 28–35% booking completion with BNPL vs 15–22% without

**Deployment timeline**:
- Phase 1 (weeks 1–2): NBFC/bank partner LOS integration, bureau API setup, AA consent flow
- Phase 2 (weeks 2–4): platform webhook integration, script development per vertical, V-CIP booking integration
- Phase 3 (week 5): UAT — 100+ scenarios per vertical (healthcare, EdTech, e-commerce, travel)
- Phase 4 (weeks 6–8): pilot on 500–2,000 leads per vertical, conversion tracking, script refinement
- One-time setup: Rs 10–20 lakh; monthly platform: Rs 3–8 lakh (depends on call volume and verticals)