How UPI actually works, explained step by step

Hundreds of millions of people in India use UPI without ever wondering what happens after they press send. For anyone working in banking or financial technology, that gap is worth closing, because almost every digital finance product now assumes this infrastructure exists.

Here is the sequence, in order.

The participants

Five parties are involved in a typical UPI payment, and it helps to name them before tracing the flow.

• The payer, and the app they are using. The app may be run by a bank or by a third party.
• The payer's bank, which holds the account the money leaves.
• The payee, and their app.
• The payee's bank, which holds the account the money arrives in.
• The central switch operated by the national payments body, which routes the instruction between the two banks and handles settlement.

The virtual payment address

UPI's first useful idea is that you do not need to know anyone's account number. Each account is mapped to an identifier, commonly in the form name@bank, called a virtual payment address.

This matters for two reasons. It removes the error-prone step of typing an account number and branch code. And it means the payer never learns the payee's actual account details, which is a quiet but genuine privacy improvement.

The flow, step by step

When you send money, the following happens in a sequence that normally completes in under five seconds.

• Your app takes the payee's identifier and the amount, and asks you to authenticate with your UPI PIN. The PIN is entered into a secure layer, not into the app itself.
• Your app sends the payment instruction to the central switch.
• The switch resolves the payee's identifier to work out which bank holds that account.
• The switch sends a debit request to your bank. Your bank checks the balance, verifies the PIN, and debits your account.
• The switch sends a credit instruction to the payee's bank, which credits their account.
• Confirmation travels back along the chain, and both parties see the result.
• Actual settlement between the two banks happens separately, in batches, through the central settlement process. The money you see arrive is the banks honouring the instruction ahead of that settlement.

Why it is free, and who pays for it

For an ordinary person-to-person transfer there is no charge to either party. This is unusual: card payments carry a merchant fee that funds the whole network.

UPI was built as public infrastructure, with the deliberate policy choice that the rail should be free at the point of use so that adoption would not be limited by cost. The operating cost is carried by the banks and the central body, and the commercial return is indirect, in the form of accounts, deposits and the products that can be built on top.

This is the single most important thing to understand about UPI's economics, and the thing most often misunderstood.

What this makes possible

Once a free, instant, universally adopted payment rail exists, a large number of products become viable that were not before. Collecting a small recurring payment stops being uneconomic. A merchant can accept digital payment without buying a card terminal. Lending products can disburse and collect in seconds.

Much of what is described as Indian financial technology innovation is, underneath, a business model that only works because this rail exists and costs nothing to use.

Where the risks sit

Speed and irreversibility are a useful combination for commerce and an excellent one for fraud. Most UPI fraud does not attack the system. It persuades the account holder to authorise the payment themselves, through a convincing story.

That is why the controls that matter most are at the customer end: transaction limits, cooling-off periods on new payees, and the plain fact that no genuine institution ever asks you to enter your PIN to receive money.