In B2B BNPL, the buyer check at checkout decides most of your loss rate. Verify that the retailer or distributor is a real, active business, with consistent GST, MCA, Udyam, and footprint signals, before a limit is granted, and keep RBI digital lending duties with the regulated lender, not the platform.
Embedded trade credit has become a standard feature of Indian B2B commerce. A kirana store ordering from a distribution app, a contractor buying materials on a procurement platform, and a pharmacy restocking from a wholesaler: each can now be offered 15, 30, or 60 days to pay, funded by an NBFC or bank behind the platform.
The economics look attractive until the first cohort ages. Fake buyers, shell retailers, and circular orders show up as early defaults, and by then the goods have shipped. This guide covers how the RBI’s Digital Lending Directions 2025 apply to B2B BNPL, why KYB at checkout is the control point, which business data signals help set limits, a checklist, a numbered approval flow, and the metrics to watch.
How B2B BNPL works and who is regulated
A typical B2B buy now, pay later structure has three parties. The platform runs the checkout and owns the buyer relationship. A regulated entity (RE), usually an NBFC or bank, extends the credit line. The supplier is paid upfront, and the buyer repays the lender on terms.
In regulatory terms, the platform is often acting as a lending service provider (LSP) to the RE. That matters, because the RE remains accountable for the loan and for what its LSP does on its behalf.
Some platforms extend open-account trade credit from their own balance sheet without a lender. That is commercial credit between businesses rather than digital lending by an RE, but the fraud risks below are identical.
What RBI’s Digital Lending Directions require
The Reserve Bank of India (Digital Lending) Directions, 2025, issued on 8 May 2025, apply to digital lending by commercial banks, co-operative banks, NBFCs, and all-India financial institutions (para 3). They make no carve-out for business borrowers. Read the points below with counsel, as RBI’s later consolidation into entity-specific directions may have changed paragraph numbering.
- LSP arrangements: the RE must have a contract defining roles (para 5(i)) and conduct enhanced due diligence before partnering with an LSP (para 5(ii)).
- Key Fact Statement: borrowers must receive a KFS before the contract is executed (para 8(i)).
- Disbursal: Loans are to be disbursed into the borrower’s bank account, subject to the exceptions the Directions list (para 9(i)). Trade-credit structures that pay the supplier directly should confirm with counsel which exception they rely on.
- App reporting: REs must report digital lending apps, including those of their LSPs, on RBI’s CIMS portal (para 17).
- Default loss guarantee: where the platform provides a DLG, total cover on an outstanding portfolio specified upfront cannot exceed five percent of the amount disbursed (para 23(i)), carried over from the 2023 DLG guidelines.
The DLG cap is the commercial point many platforms underestimate. If the platform’s first-loss exposure is capped at 5%, the lender carries the rest of the loss. Lenders therefore have every reason to insist on strong buyer verification at checkout, and platforms that offer it negotiate better terms. Our guide covers the full framework.
What weak buyer verification costs a B2B BNPL book?
Credit is where fraud losses concentrate. The RBI Annual Report 2025-26, as reported by Outlook Business on 29 May 2026, recorded 10,114 fraud cases involving Rs 48,021 crore in FY26; advances accounted for 8,640 cases and Rs 40,774 crore of that.
The supply of fake businesses is large. The Ministry of Finance reported on 7 January 2024 that its drive against fake GST registrations had detected 29,273 bogus firms since mid-May 2023, with suspected ITC evasion of Rs 44,015 crore. Entities built to generate invoices can also generate orders.
Illustrative: a platform originates Rs 50 crore of B2B BNPL a month at a 45-day tenor. If 0.8% of buyers are fraudulent and default on their first bill, that is Rs 40 lakh lost per monthly cohort. With a 5% DLG, the platform absorbs part of it until the guarantee is exhausted; the lender carries the remainder, and the pricing of the next tranche reflects it.
Why KYB at checkout is the control point
In consumer BNPL, identity is the main question. In B2B BNPL, the question is whether the business exists, operates as claimed, and can absorb the goods it is ordering. That has to be answered before the first limit, because three common fraud patterns all exploit the first order.
- Fake buyers: newly created accounts using a borrowed or recently registered GSTIN, placing one large order and disappearing.
- Shell retailers: registered entities with no real trading activity, often sharing directors, addresses, or phone numbers with other shells.
- Circular orders: related parties ordering from each other to manufacture repayment history, then scaling limits before defaulting together. The same pattern shows in bank data.
Each is cheap to stop at onboarding and expensive to recover after shipment. Collections against a shell entity rarely recover much. More patterns appear in our note.
Data signals for limit setting
Business data helps with two decisions: whether to offer credit at all and how large the first limit should be.
- GST registration and filing behavior: active status, registration age, taxpayer type, and regular GSTR-1 and GSTR-3B filing. Declared turnover in returns is a useful ceiling for limits.
- MCA status for incorporated buyers: active company, incorporation date, directors, and their other directorships. Directors linked to struck-off or many recently formed companies warrant a referral.
- Udyam registration: enterprise classification and declared activity; useful for checking whether the claimed business type matches orders.
- Digital footprint: a retail or wholesale presence consistent with the trade name and location. Absence is not disqualifying for a small kirana, but it matters for large first orders.
- Platform behavior: order history before credit, delivery addresses, payment method history, and links to other accounts.
A sound starting rule: first limits are a fraction of what filings support, rising with on-time repayment. Business verification determines eligibility and the starting band, and repayment behavior does the rest.
B2B BNPL KYB checklist
| Criterion | Why it matters | What to ask the vendor |
|---|---|---|
| Input at checkout | Many small buyers only share a mobile number. | Can a check start from a GSTIN or a mobile number? |
| Turnaround | Checkout cannot wait hours. | What is the time to a full report? |
| GST filing depth | Filing behavior drives limit bands. | How many periods of filing status are returned? |
| People data | Shell retailers share directors. | Do you profile directors of private limited buyers and their other companies? |
| Inconsistency flags | Fraud lives in mismatches. | Do you flag name, address, and registration-date conflicts across sources? |
| Evidence | The RE must justify decisions. | Can each report be stored with a timestamp against the loan file? |
| Data protection | Proprietor data is personal data under the DPDP Act 2023. | Processor agreement, storage location, and retention? |
| Delivery mode | Fits platform and lender systems. | Portal, API, or bulk: which are live today? |
Approval flow: from checkout to first limit
- Capture at checkout (platform). Collect the GSTIN or mobile number, business name, and delivery address, with consent for verification and credit assessment.
- Pre-screen the business (platform as LSP, or lender). Run a business risk check on status, registrations, people, footprint, and inconsistencies before any KYC.
- Route by result (lender’s rules). Clean: proceed. Inconsistent: refer to an analyst. Cancelled registration, shell indicators, or links to known fraud: decline.
- Complete KYC and bureau checks (lender). Verify the entity and its authorized persons under the KYC Directions applicable to the lender, and pull commercial and consumer bureau data where relevant.
- Set the first limit (lender). Use filing-based turnover, business age, and order history to pick a starting band.
- Issue KFS and execute (lender, via platform). Deliver the KFS before execution and disburse in line with para 9(i).
- Monitor and re-verify (lender and platform). Re-check business status on limit increases, address or bank changes, and missed payments.
A realistic pilot runs eight to ten weeks: two weeks to agree rules, one to two weeks for a POC on historical applications, and four to six weeks on a live segment with a control group.
Metrics that tell you it is working
| Metric | Why it matters |
|---|---|
| Approval TAT (checkout to limit) | Conversion falls as wait time grows. |
| Approval rate by segment | Shows whether checks are blocking good buyers. |
| First-payment default (FPD) | The clearest fraud indicator in trade credit. |
| Fraud rate (confirmed fraud as share of approvals) | Separates fraud from credit losses. |
| Referral rate and analyst minutes per referral | Measures review load. |
| DLG utilisation | Shows how quickly platform-first-loss is being consumed. |
Where BeFiSc fits: BizCheck produces a business risk report in under 60 seconds from a GSTIN or a mobile number, built on GST, MCA, Udyam and digital-footprint data, and it profiles the people behind Private Limited entities. Lenders screening merchant loans is a listed use case. In a B2B BNPL flow, it serves as the pre-screen in step 2: it sits before KYC rather than replacing it, and its signals inform, but do not set, the lender’s limit. The free POC runs on 20–50 of your own buyer applications over 5–7 working days, with a full report per case, a side-by-side comparison with your current screening, and a debrief call, with no setup fee.
What to bring to the POC: 20–50 buyer GSTINs or mobile numbers, your approval decision and limit for each, and repayment outcome where available (especially first-payment defaults).
Key Takeaways
- B2B BNPL funded by a bank or NBFC through a platform falls under the RBI’s Digital Lending Directions 2025; the RE stays accountable for its LSP.
- The 5% DLG cap means lenders carry most fraud losses, so strong buyer verification improves platform-lender terms.
- Fake buyers, shell retailers and circular orders all exploit the first limit; KYB at checkout is the cheapest point to stop them.
- GST filing behaviour, MCA status and directors, Udyam data and digital footprint together support eligibility and starting limit bands.
- Track approval TAT and approval rate alongside first-payment default and confirmed fraud rate, so controls do not quietly kill conversion.
Conclusion
B2B BNPL works when the lender and the platform agree on one thing: the first limit is the riskiest decision in the relationship. Everything after it can rely on repayment data. The first decision relies only on what you can verify about the business at checkout.
The regulatory structure reinforces this. The RE is accountable, the DLG is capped, and the losses that slip past onboarding land mostly on the lender’s book. A fast business pre-screen, clear routing rules, and disciplined starting limits are the controls that keep the product viable.
Start by testing those controls on decisions you have already made, especially the ones that defaulted on the first bill.

Top comments (0)