Product 02 — Payment infrastructure

A failed payment is
revenue you already earned.

Most businesses treat declines as a cost of doing business. They are not — a meaningful share are soft declines that would succeed on a different acquirer, at a different hour, or on a second attempt timed properly. This platform routes across acquirers, retries intelligently by decline code, and shows your revenue team exactly where the money is leaking.

Multi-acquirer routing UPI, cards, wallets, netbanking Mandates & subscriptions Tokenised, reduced PCI scope
Request a demo See the modules
What is inside

Six modules, one integration.

Your engineers integrate once. Adding an acquirer, a payment method or a new country becomes a configuration change rather than another sprint.

G/01

Smart routing

Each attempt scored on issuer, card network, amount, method and time of day, then sent to whichever acquirer performs best for that combination. Automatic failover when an acquirer degrades, so an outage on one rail does not become an outage for you.

Issuer-awareAuto failover
G/02

Retry intelligence

A fixed three-attempt schedule wastes most retries. Here the timing and rail are chosen per decline code — insufficient funds retried near payday, issuer timeouts retried within minutes on another acquirer, hard declines not retried at all.

Per decline codeTimed by pattern
G/03

Payment methods

Cards with 3DS, UPI collect and intent, wallets, net banking, EMI and pay-later options behind a single API and a checkout you can host or embed. New methods activate without your team writing integration code.

3DSUPI intentEMI
G/04

Mandates & subscriptions

NACH registration, UPI autopay and card-on-file mandates with lifecycle handling — registration, amendment, pause and revocation. Recurring debit runs on schedule with pre-debit notification and failure handling built in.

NACHUPI autopay
G/05

Settlement & reconciliation

Acquirer settlement files matched against transactions and bank credits automatically, with fees and taxes broken out per transaction. Refunds, chargebacks and disputes tracked through their full lifecycle rather than in an email thread.

Auto-matchChargeback tracking
G/06

Decline analytics

The dashboard your revenue team reads weekly. Authorisation rate by issuer, method, acquirer and hour, with declines grouped by reason instead of stopping at "failed". This is usually where clients first discover which bank is quietly costing them the most.

Auth rate by issuerReason grouping
Security & compliance

Four things your auditor will ask.

Payment work carries obligations that arrive whether or not you planned for them. These are designed in rather than remediated later.

The cheapest way to pass a PCI audit is to never touch a card number.

Tokenisation keeps cards out of your systems

Card details go straight from the customer's browser to the acquirer; your servers only ever hold a network token. That is the difference between a full PCI-DSS assessment and a short self-assessment questionnaire — a material saving in audit cost every single year.

Scope reduced, not just managed

Every transaction leg is logged immutably

Each attempt, retry, callback and settlement entry is appended with actor and timestamp, never overwritten. When a customer disputes a charge from eight months ago, the full sequence reconstructs in seconds rather than becoming an investigation.

Reconstructable, exportable

Idempotency and webhook integrity

Duplicate charges are the failure mode that costs trust fastest. Every request carries an idempotency key, and every inbound webhook is signature-verified and replay-protected — so a network retry never becomes a second debit on a customer's account.

No duplicate debits

Data residency by market

Indian payment data stored in India per RBI localisation requirements, UK and EU data in-region under GDPR. Serving a new market is a deployment configuration rather than an architecture project you discover you need mid-expansion.

RBI, GDPR, MAS aware
Proof

Routing logic that turned declines into revenue.

Subscription business · Multi-acquirer 🇬🇧 United Kingdom · 9 months · Under NDA

Roughly 6% of recurring charges were being written off

The problem

Every failure went to a single acquirer, retried on a fixed schedule, and dropped after three attempts. Nobody could see which issuers were declining or why, because reporting stopped at the word "failed". The finance team had accepted the loss as a fixed cost of the business model.

What we did

Built a routing layer across three acquirers scoring each attempt on issuer, card type, amount and time of day, then retrying on a different rail with a schedule tuned per decline code. Added a decline dashboard the revenue team reads weekly, and moved card storage to network tokens to cut PCI-DSS scope.

+4.1pt
Authorisation rate
£2.3M
Annualised recovery
SAQ-A
PCI scope reduced to
Before you call

Common questions.

Ask something else

No, and this is the most important thing to be clear about. We build and run the technology; the money moves through your own acquirer and banking relationships under your licence. We do not hold funds, we do not take a cut of transaction value, and we are not a party to the settlement. If you need an aggregator, you need a different kind of company.

The routing layer is acquirer-agnostic and we have integrated the major Indian and UK providers plus several regional ones. Each new acquirer is a defined piece of work, typically two to four weeks, quoted upfront. Keep your existing relationships and negotiated rates — this sits above them rather than replacing them.

It depends entirely on where you start. A single-acquirer setup with fixed retries usually has real headroom. A business already running two acquirers with tuned retries has much less. We will analyse a sample of your decline data before quoting and tell you if the gain does not justify the project — that conversation costs you nothing.

Usually one to two weeks of your team's time against our API and hosted checkout, with a sandbox from day one. Platform setup, acquirer integrations and configuration take eight to sixteen weeks in parallel depending on how many rails and methods you need live at launch.

Next step

Send us a month
of decline data.

Under NDA. We will show you the recoverable share, the issuers costing you most, and a realistic estimate of the lift — before you commit to anything.

Request a decline analysis See the BBPS Portal