What is Zelle and how does it work?

Canonical merchant metadata

• Maintain a merchant_profile that includes: MCC / NPCI category, verified flag (yes/no), sponsor bank/PSP IDs, and the official daily & per-txn caps for that category.
• Keep this data synced daily from your PSP or acquirer (don’t hardcode assumptions).

Keep real-time daily usage counters

• Track upi_volume_today per merchant (and optionally per merchant+PSP handle) in a fast store (Redis). Increment only after final success confirmation (webhook or settlement ack).
• Persist an authoritative copy in your DB for audits.

Pre-debit validation flow (run BEFORE initiating UPI collect)
When a subscription payment is due:

• Lookup merchant category & cap: cap_txn, cap_day = (verified ? high_cap : normal_cap).
• Read used_today = upi_volume_today and remaining = cap_day - used_today.
• If amount <= min(cap_txn, remaining) → proceed with UPI collect.
• If amount > min(cap_txn, remaining) → do not fire the UPI collect. Instead route to fallback (card, netbanking) or auto-split logic (if supported).

Support for splits / deferred attempts

• If you allow auto-split: compute compliant legs (each ≤ per-txn cap and within remaining daily allowance), create a payment_plan record with leg IDs, and execute legs sequentially (prefer sequential to avoid race conditions).
• If you allow scheduling: queue remaining amount for next valid window and notify customer (don’t auto-debit without consent if policy requires).

Real-time sync with PSPs & webhooks

• Rely on PSP webhooks for PAYMENT_SUCCESS, PAYMENT_FAILED, REFUND, REVERSAL. Only increment the upi_volume_today on a confirmed success webhook (or settlement confirmation).
• If webhooks are lost, reconcile via periodic pull or settlement file to avoid drifting counters.

Race conditions & concurrency protection

• Use optimistic locking or atomic Redis ops (INCRBY with check) when reserving remaining capacity just before initiating collect: reserve → call PSP → on success commit, on failure release.
• Idempotency keys per subscription+attempt so retries won’t double-count.

Fallback UX & orchestration

• If pre-check fails, surface a clear, friendly message: UPI limit reached for today — choose card, netbanking, or schedule. Offer one-tap fallback with invoice/policy prefilled.
• For partial-success (split legs), show progressive updates: ₹X of ₹Y paid via UPI; remaining ₹Z — pay now with card or we’ll retry tomorrow.

Mandate / autopay rules

• For UPI AutoPay mandates, verify that mandate is valid and that the bank supports multi-leg/large value AutoPay (some banks differ). If a mandate would breach daily cap, don’t attempt — schedule or prompt for alternate method.

Reconciliation & accounting hooks

• Persist PSP txn refs and map every successful leg to the subscription invoice. Reconcile by UPI ref/UTR.
• Distinguish Received (Pending settlement) GL vs Settled GL; only finalize revenue when settlement/confirmation arrives (or per your revenue rules).

Notifications & audit trail

• Send pre-debit notification where required. Log pre-check result, decision, API payloads, PSP response codes, and who/what fallback was chosen. Keep this for compliance and dispute resolution.

Monitoring & alerts

• Alert when failed due to limit events spike. Add dashboards: daily cap usage per merchant, split success rate, number of fallbacks, and reconciliation mismatches.

Edge cases & tests

• Test concurrent debits that together exceed remaining allowance.
• Test midnight boundary (cap reset) and cross-day splits.
• Simulate webhook loss and settlement lag to ensure counters reconcile.
• Test refunds: refunded amount should decrement upi_volume_today if PSP semantics require (verify with PSP whether refunds free up cap).

Policy & UX decisions to finalize

• Decide if you will auto-split or force fallback when over cap. (Auto-split is smoother but more complex.)
• Decide whether to block subscription processing or schedule retries when cap is hit.

Following the NPCI regulations that came into effect on September 15, 2025, your POS system must validate insurance premium UPI payments by checking both per-transaction and cumulative daily limits for the specific merchant category. This is distinct from the lower, standard UPI limits.

Start by tracking cumulative UPI transaction volume per store per day in your payment service or ERP. Every time a UPI payment is processed, log the transaction amount along with the merchant VPA, date, and category. Then, have a daily counter that resets at midnight (or whenever NPCI’s cycle resets).
Next, define category-wise cap thresholds in your config, for example, ₹10L for verified merchants in approved categories and ₹2L for everyone else. Then, set trigger points like 80%, 90%, and 100% of the limit.
At the 80% mark, you can send a soft warning via Slack, email, or dashboard alert, something like:
Heads up! Store #45 has reached 80% of its ₹10L UPI cap for today (₹8L processed). Monitor closely to avoid declines.
At 90%+, escalate it as a high-priority alert, possibly tagged to both finance and store managers so they can proactively switch to alternate payment modes (like netbanking or card) before UPI declines start.
To make this real-time, plug it into your payment processing middleware or gateway webhooks. Each successful transaction updates the cumulative total immediately, no waiting for end-of-day reports.
If you want to get fancy, you can integrate a visual dashboard showing daily cap usage as a progress bar per store. Many finance teams do this in Metabase or Grafana, it’s an easy way to see which merchants are nearing their limits at a glance.
Finally, make sure your alert system includes category validation logic. You don’t want to alert a store using a ₹2L cap if it’s actually verified for ₹10L. So sync your merchant category and verification status daily with your PSP or NPCI data source.

Ride-hailing apps can simplify split fares using UPI by leveraging the new, higher daily aggregate limits for split payments and offering a streamlined user experience. They can enable users to pay their share of a fare directly through a single, higher-value split payment, rather than making multiple smaller transactions. The app can also streamline the process by automatically calculating and suggesting the split amount for each user's UPI payment, making the process faster and simpler.

First, hardcode category-based caps in your ERP’s payment configuration. Tag insurance premium payments with a merchant category like MCC 6300 (insurance carriers) and set a max UPI limit of ₹5L (or ₹10L if verified). This ensures any transaction above that instantly throws a validation error before hitting your payment gateway.
Second, add a daily aggregation check. Your ERP should maintain a running total of all UPI transactions made to that insurer per day per payer account. If a new premium payment pushes the total beyond the daily limit, the system should automatically block it or split it into compliant chunks (like ₹5L + ₹3L).
Third, enforce real-time warnings and soft limits in the UI. When a finance team member tries to initiate a large UPI payment, show an alert:
UPI daily cap for insurance payments is ₹5,00,000. Please use NEFT or RTGS for higher amounts.
This prevents staff from wasting time on transactions that will fail anyway.
Fourth, include policy-level validation in workflows. For example, if the premium amount exceeds the cap, automatically switch the default payment mode from UPI to bank transfer. Your ERP can trigger this rule in the payment gateway selection logic.
Fifth, implement audit logging and exception reports. Track all over-limit attempts, even blocked ones, so finance teams can identify if users are consistently trying to make non-compliant payments. This helps refine training and controls later.
Finally, make sure your UPI integration uses NPCI’s error response codes (like U68 for limit exceeded) to handle rejections cleanly in case any slip through.