Part 04 — Digital & Embedded Lending

OCEN, ONDC & Future

The future of embedded SME credit in India is not only private platform partnerships. It is increasingly shaped by public and open rails: Account Aggregator (AA) for consented financial data, Open Credit Enablement Network (OCEN) for standardized lending interactions, Open Network for Digital Commerce (ONDC) for unbundled commerce and financial services, DigiLocker/Aadhaar for know-your-customer (KYC), eNACH/eMandate for repayments and eSign for agreements. The promise is not that underwriting risk disappears. The promise is that lenders, loan agents, platforms and borrowers can connect with lower integration cost and better auditability.

OCEN: Unbundling The Lending Job

OCEN describes itself as a framework of application programming interfaces (APIs) for interaction between lenders, loan agents, collection/disbursement partners, derived data providers and Account Aggregators (OCEN). Its documentation explicitly reframes lending as a multi-party workflow: loan agents bring borrower context, lenders handle underwriting and capital, and other providers support data, disbursement and collections. OCEN now uses “Loan Agent” (LA) language and explains that LAs are intended to act as agents of the borrower, helping borrowers compare offers and understand rights, rather than merely acting as direct selling agents for lenders (OCEN introduction).

The practical value for micro, small and medium enterprise (MSME) lending is in small-ticket, short-tenor loans where custom bilateral integrations are uneconomic. OCEN’s earlier pilots show the pattern. GeM Sahay used Government e-Marketplace (GeM) as loan agent, with lenders offering loans to SMEs against government invoices. GST Sahay, launched in January 2023, used GST data and Account Aggregator data for working-capital loans, with SIDBI acting as loan agent according to OCEN’s pilot documentation (OCEN previous pilots).

That is a different architecture from a closed marketplace. In a closed model, each anchor negotiates and integrates separately with each lender. In an OCEN-style model, product discovery, offer, consent, sanction, disbursement and repayment messages are standardized enough that many lenders can serve many loan agents. The hard part shifts from integration to policy: which borrowers, which data, what limit, what price and who handles exceptions.

Account Aggregator As The Data Rail

AA is the consented data-sharing layer. The Department of Financial Services states that the AA network was introduced through the Non-Banking Financial Company - Account Aggregator (Reserve Bank) Directions, 2016, dated September 2, 2016, and that no customer financial information is retrieved, shared or transferred without explicit consent. As of March 31, 2026, it reports 179 financial institutions live as financial information providers (FIPs), 989 as financial information users (FIUs), more than 2.88 billion accounts enabled and 284.6 million linked accounts (DFS AA framework, updated May 27, 2026).

For SME lending, AA reduces dependence on uploaded PDFs and screen-scraped bank statements. A lender can receive bank-account data with consent, map inflow quality, detect bounced obligations, identify existing loan repayments and reconcile GST or platform turnover. But AA does not solve all data gaps. Many micro businesses use multiple accounts, cash, family accounts or platform wallets. A good underwriting system still needs entity resolution: proprietor PAN, business PAN, GSTIN, Udyam, bank accounts, platform seller IDs and guarantor bureau records.

ONDC Financial Services

ONDC began as an open commerce network but now includes financial-services categories. ONDC’s buyer-app page lists “Loan, Investments, Insurance & More” and includes personal loan and MSME loan categories, while making clear that ONDC is not a party to transactions through buyer applications (ONDC buyer apps). In August 2024, ONDC announced fully digital, paperless loans in about six minutes, integrated with nine buyer applications/LSPs and three lenders; the initial product was unsecured credit for salaried and self-employed individuals, using AA, DigiLocker/Aadhaar, eNACH/eMandate and Aadhaar eSign (ONDC press release, August 22, 2024).

ONDC’s financial-services resources now include credit specifications for personal loans, GST-based invoice loans and working-capital lines (ONDC financial services resources). That is important for SME lending because ONDC can separate roles: buyer app, seller app, lender, loan-service provider, data provider and reconciliation layer. A tyre dealer on one app, a pharmacy on another and a small restaurant on a third could theoretically access comparable credit products without each platform building a bespoke lender panel.

What Changes For Lenders

Open rails compress the cost of acquisition and data collection, but they also increase comparability. If multiple lenders appear in the same borrower journey, the Digital Lending Directions, 2025 require a fair digital view of matching offers with lender name, amount, tenor, annual percentage rate (APR), monthly repayment obligation and penal charges. That weakens opaque lead-routing economics and rewards lenders that can price accurately from standardized data.

Credit policy also has to become more modular. Instead of a human relationship manager collecting a full file, the policy engine may receive an OCEN/ONDC payload: GST invoice, bank AA data, Udyam, KYC, bureau, anchor-derived score and requested amount. The lender must answer quickly: decline, request more data, counter-offer, sanction with conditions or route to manual underwriting. System notes for this belong in integration catalog and state machines.

What Changes For Platforms

Platforms lose some exclusivity but gain faster access to lenders. A B2B marketplace that previously needed six months to negotiate lender integrations can expose standardized credit flows. Its moat moves to borrower context: quality of invoices, seller reputation, fraud controls, dispute history, return rates and ability to support collections without abusing customers.

Platforms must also decide whether they are a borrower-side loan agent, an LSP of the lender, or both under different contracts. That distinction affects disclosure, grievance responsibility, data use and fees. If the platform provides a default loss guarantee, the 5 percent DLG cap and disclosure rules apply. If it presents multiple lenders, dark-pattern and offer-comparison rules apply.

The Likely 2026-2028 Shape

The most realistic near future is hybrid. Large anchors and payment processors will keep private lender partnerships because they want control over experience and economics. Public rails will become the default for long-tail distribution, invoice-backed loans, GST-led working-capital lines and standardized small-ticket offers. Co-lending will remain important behind the scenes because banks want granular assets and NBFCs/platforms want capital efficiency.

The durable winners will not be the platforms with the most loan buttons. They will be the ones with clean consent, reliable identity resolution, high-quality transaction data, low dispute rates, transparent KFS journeys, fast reconciliation and disciplined collections. Open rails reduce integration friction; they do not forgive sloppy credit.