US IT leaders struggle to test faster cloud updates

Research from Original Software found that 75% of US enterprise IT leaders say updates to business-critical systems are arriving faster than their organisations can test them. The findings point to growing pressure on IT teams as cloud software changes release patterns.

The survey also found that 92% of respondents said cloud adoption had increased the amount of testing they needed to do, with 54% reporting a significant increase. The shift from on-premises software to cloud services has left IT teams dealing with more frequent updates that arrive on vendors' schedules rather than their own.

The report, titled The Enterprise Software Testing Gap, argues that testing demands have grown as business processes have spread across multiple applications and integrations. In that environment, a change in one system can affect another, making it harder for teams to trace problems before they disrupt operations.

That challenge is already showing up in the research. Some 59% of US IT leaders said it is becoming harder to identify issues before they affect business operations, while fewer than half, 48%, said they were fully confident their current testing approach would catch critical issues before impact.

Manual burden

The findings suggest many organisations are still relying on older methods to manage a rising testing workload. More than a third of respondents, 36%, said they remain largely dependent on manual testing using spreadsheets, documents, or email, while only 6% said they had highly automated testing across most software systems.

That gap between the pace of software change and the tools used to check it raises wider concerns for business continuity. If a software update affects finance, payments, warehouse systems, or customer processes, the consequences can quickly spread beyond IT teams into day-to-day operations.

Carl Andrews, Chief Executive Officer of Original Software, said the problem stemmed from a mismatch between modern software delivery and older testing practices.

"Enterprise software has changed enormously, but the way many organisations test their software hasn't changed at the same pace. You cannot keep adding testing workload without changing how it is handled. Core business processes can run across ERP, warehouse, payment, finance, and customer systems. A seemingly minor change in one application can have consequences somewhere completely different," Andrews said.

Vendor limits

The report also highlights a common assumption that a vendor's own testing is enough to protect a customer from disruption. That approach overlooks the fact that each enterprise has its own mix of configurations, customisations, integrations, and data flows, all of which can affect how an update performs in practice.

For IT leaders, that means a release that works in a supplier's environment may still create problems when deployed inside a customer organisation. The research suggests many respondents are aware of that risk but do not yet feel their current processes are robust enough to manage it consistently.

Andrews said organisations should focus not on slowing software updates, but on changing how they prepare for them.

"A vendor can test its software, but it cannot test your business. It doesn't have your exact configurations, integrations, data, or end-to-end processes. I'm not saying the answer is to slow down software updates - IT leaders need to change how they approach testing so they can keep pace without leaving the business, or themselves, exposed. That means planning earlier, automating repeatable processes, and treating testing as part of business continuity rather than simply another stage of an IT project," he said.

Automation role

The report presents automation as one way to reduce repeat work, particularly where the same core business processes need to be validated each time an update is issued. By recording tests around those processes, organisations can build repeatable test libraries and reduce the need for business users to carry out the same checks manually.

At the same time, the research does not present automation as a full replacement for human input. Instead, it points to a model in which repeatable tasks are automated while business users focus on new processes, exceptions, and exploratory testing.

Addressing that distinction, Andrews said: "Automation doesn't mean handing everything over to AI. Business users understand how these processes work and remain essential to testing. The opportunity is to stop asking them to manually repeat the same tests every time an update arrives. Start with the processes the business cannot afford to have fail and progressively automate the repeatable testing around them. As software change accelerates, testing needs to become a strategic capability for resilience."