Open banking data now flows into accounting software every day, giving firms a daily view of their cash. We asked industry leaders what it changed and what still gets in the way, and the answer is that bank feeds still need regular checks.

Open banking data has become routine plumbing for small business finance. In the UK, the FCA says about 17 million people now use open banking, nearly one in three adults. Meanwhile, the volumes keep climbing. Open Banking Limited recorded 2.81 billion API calls in June 2026, a record, and 40.16 million payments. Yet the people who rely on open banking data every day still report gaps, duplicates and missing details.

We asked founders, product leads and partnership heads one question: how has open banking data sharing changed what your product can do for customers, and what still gets in the way? We also asked for real use cases over commentary. Three people answered, and we used all three. However, none of them is a lender or an open banking provider. Instead, they run bookkeeping, finance advisory and small retail businesses that use bank feeds every day. So we checked their experience against the data standards and the latest rules.

Open banking data by the numbers

First, reliability. In June 2026, UK open banking APIs posted 99.80% weighted availability and a 99.50% success rate, according to Open Banking Limited. Those figures look strong. Still, a 0.5% failure rate on 2.81 billion calls means about 14 million failed calls in a single month. Also, the figures cover only the nine largest UK banks, so smaller banks sit outside the statistics.

The US picture is less settled. The Financial Data Exchange says more than 114 million connections now run through its standard APIs. However, its own comments to the CFPB suggest that roughly a third of consumer-permissioned data sharing still relies on screen scraping. As a result, open banking data in the US depends heavily on which bank a customer uses.

No more month-end imports

Our first respondent founded Profitjets, a Delaware-registered bookkeeping firm with its team in Bengaluru, which serves clients across the US and Canada.

Automated bank data transfer changed our bookkeeping. We no longer import bank data to process transactions, which lets us focus on reviewing them instead. Because transactions now arrive daily rather than in a month-end batch, we no longer rebuild the financial records at the end of each month. Each transaction also comes with more detail attached.

From time to time, though, feeds become unreliable, so transactions are duplicated or don’t come through at all. The coverage and reliability of the bank connection decides whether the integration works. Some banks and account types are still not connected.

Automation removed the imports, but it didn’t remove the uncertainty in the records. That still needs human judgement. Some bookkeeping teams skip that step and build records without supporting documents.

Abhinav Gupta, Founder, Profitjets

Abhinav’s experience reflects the US market, where most bank feeds reach QuickBooks or Xero through data aggregators rather than a regulated open banking regime. His coverage point fits that picture. Open banking data from banks that still rely on screen scraping is more likely to break or drop out. We also changed his “real-time” to “daily”, since most accounting feeds refresh on a schedule rather than instantly. His final line is a criticism of rival bookkeepers, so treat it as his opinion. Profitjets is a bookkeeping firm, not a fintech.

A daily cash view, with gaps

Our second respondent runs a custom apparel business with companies in the US and Northern Ireland.

One feature I’d like to see more of from accounting software providers is open banking integration, so bank transactions pass straight through, ready to be categorised. For small businesses especially, it takes away some of the admin around bank reconciliation, such as downloading statements and reconciling them by hand. It also gives you a better idea of where you stand on cash on any given day, not just at month end. When you’re selling in several countries like we are, that helps you make better decisions on stock and spending.

There’s still room for improvement. Payment references aren’t standardised, between banks or even within them, so transactions can’t always be matched automatically. And although open banking integrations are generally reliable, connections can drop without warning. It isn’t a disaster, since you can always download statements, but I’d recommend that any business using this feature checks regularly that transactions are still coming in. My suggestion for accounting software teams would be to tell customers clearly when this happens.

Conor Smart, Managing Director, Arklavo

Conor’s reference problem sits in the system itself. In the UK, a Faster Payments reference allows only 18 characters, typed by the person paying. Open banking data can pass on only what the bank holds, so it cannot add an invoice number the payer never sent. His warning about silent drops needs one qualification. Xero’s UK open banking page says it prompts users before consent expires. Still, bank outages and broken connections can stop a feed with no clear alert, so his request is fair. Arklavo is an apparel firm, not a fintech.

Fintechbits analysis

Why open banking data feeds still go quiet

Our third respondent is a chartered accountant, CFO of Verloop.io and curator of CFO Matrix, a finance advisory practice in Bengaluru.

The biggest difference is the speed of the books. We receive transaction data every day, so we can categorise weekly. If something is duplicated or a charge is missing, we can ask about it quickly.

It can also cause problems. One client’s feed was switched off for several weeks because it needed re-authorising. Once we noticed the gap, we had to rebuild it from statements. Now we check every week that all feeds have posted recently.

There isn’t much data. The descriptions are quite generic, and confusion over whether items are pending or posted can create duplicates. Invoice numbers rarely come through, so we still match payments by hand.

It’s hard to know how much this varies between banks and countries, but mainly it’s a problem for the books.

Ankit Sarawagi, Curator, CFO Matrix

Ankit’s lapsed feed matches a common trap. Since 2022, the FCA’s PS21/19 rules have spared UK customers from logging in to their bank every 90 days. However, the app must still reconfirm consent at least every 90 days, and the feed stops if nobody does. The EU window is 180 days. His duplicates also have a clear cause. The UK open banking data standard marks transactions as “Pending” or “Booked”, and pending items can change before they settle. He did not say which country or software his client used, so his example may come from India rather than the UK.

What to watch next

Next, watch the US. A court has blocked the CFPB’s 2024 open banking data rule, and the bureau is reconsidering it. Meanwhile, JPMorgan now charges aggregators such as Plaid and Yodlee for data access, which could raise the cost of open banking data for apps. In Europe, the Commission has moved to withdraw its open finance proposal, while PSD3 awaits final publication. In the UK, the FCA plans to consult on a long-term open banking framework by the end of 2026, and its open finance roadmap puts SME credit first. Our guide to how open banking works covers the wider rules.

Above all, open banking data removes the manual import, not the need to check. Firms that rely on bank feeds should confirm weekly that every feed is posting, renew consent before it lapses and keep a manual route for matching payments.

FintechBits covers payments, banking and fintech developments for readers in the US and UK. This article is for information only and is not financial or legal advice. Views expressed are those of the FintechBits editorial team.