When Systems Are Running but Payments Are Not Moving

  • Antoine Cuypers, Director Strategic Alliances & Key Accounts at Vyntra

  • 03.11.2026 02:00 pm
  • #payments

The IT systems processing a payment can show a green light while a customer’s payment is going nowhere. The servers are running, there is enough processing capacity and no application has reported an outage. Yet somewhere along the journey, a transaction is waiting in a queue.

For a bank monitoring individual systems, everything may appear healthy. The customer is still waiting for their money.

 At Sibos this year, I discussed what this gap means for banks and their customers. It raises a question for financial institutions: how much does knowing that the infrastructure is working tell us about the service being delivered?

Understanding the Layers of the Payment Journey

Understanding that gap starts with the layers a payment passes through, from older mainframes to middleware and newer applications.

Silo-based IT monitoring tells the bank about each system’s condition, but does not necessarily reveal what happens between them. A transaction may require manual intervention from someone who is absent, or a queue may be moving more slowly than expected. Neither situation requires a system to go down.

Business activity monitoring follows the customer journey from beginning to end. It establishes what normal behaviour looks like across successive stages and identifies departures that need investigation.

Context matters here. The first Monday of the month, after salaries have been paid, can look different from the third. Retail payments and remittances have their own patterns too. Understanding these differences helps teams recognise unusual delays without treating every variation as a problem.

Customers Expect an Answer While They Wait

That understanding of the payment journey also helps banks answer a basic customer question: where is my payment? 

With instant payments, the genie is out of the bottle. Customers are accustomed to ordering something and tracking it to their front door, often with an estimated delivery time. They increasingly expect comparable visibility from their financial institution.

One customer example shows how different the experience can be within the same banking group. Following a merger, the bank had two operating centres with separate data environments. In the centre using Vyntra’s solution, a customer service employee could look up a transaction and answer a caller in around 30 seconds. 

For an enquiry involving the other centre, the answer could take three days. The front office had to contact operations, which might then need IT to retrieve archived information. The result travelled back through the same chain and almost four hours of staff time could be involved.

Giving customer service direct access to transaction information reduces the internal work required to answer a routine question. It also means customers can receive an answer during their call.

Detecting Delays Before Customers Report Them

Answering an enquiry quickly matters, but the same visibility can help a bank identify a delay before the customer calls. This becomes particularly important when a payment holds up the business activity it was intended to support.

In a trade finance example I discussed at Sibos, a container ship could not leave because a required payment had become blocked along the transaction chain. The captain contacted the corporate customer, which then contacted the bank. The customer’s complaint was the bank’s first indication of the problem.

This illustrates why the timing of detection matters. A payment delay can interrupt commercial activity even when the systems processing it remain available.

Real-time visibility gives operations teams an opportunity to investigate an abnormal flow earlier. It cannot guarantee that every disruption will be avoided, but it gives the bank a chance to respond before the customer has to report it.

A Passport for Each Transaction

Investigating a delayed payment also means understanding what has happened along its journey, including which checks it has passed through. We describe the record of those checks as collecting ‘Know Your Transaction’ passport stamps.

When a transaction passes through sanctions screening, its record captures the check, the result and the time. Other controls, including anti-money laundering and fraud checks, contribute their own stamps.

If the transaction reaches a stage without an expected stamp, the missing record can trigger an alert. Teams can investigate the gap and retrieve the history when asked how a payment was processed.

Integrity checks add another dimension by detecting changes to the account number, amount, currency or beneficiary. A change does not automatically mean wrongdoing. There may be a valid explanation, but the institution needs to detect and record it.

This brings together the transaction’s status, the checks it has passed through and any changes made along the way. It provides a single version of the truth that customer service, operations and compliance teams can use.

Making Transaction Intelligence easier to access

Once that transaction history is available in one place, the next question is how easily people can use it. A natural-language interface could allow customers and internal teams to ask questions without needing to understand technical transaction fields.

A corporate customer could, for example, ask how their payment activity across two countries has changed over the past three months. The underlying transaction data would supply the answer, with a language model presenting it in an accessible form and allowing follow-up questions. Audit and compliance teams could use the same approach to query the transaction history.

The value depends on the information behind the interface. Banks need a connected record of where a payment is, which checks it has passed through and what has changed along the way. Making that record accessible helps people answer questions and investigate problems using the same information.

Related Blogs

Other Blogs