Older applications stay in place because they still do a useful job, while newer platforms, including AI, are added around them.

Individually, those decisions are often perfectly reasonable. The difficulty comes from the connections between them. Systems do not always share information easily, data ends up in different places and the gaps between applications are rarely owned by anyone.

This creates a less obvious productivity problem. People end up filling those gaps themselves, moving information between systems, looking for data, following up on approvals and switching between applications to complete a process that should be relatively straightforward. None of these tasks takes a huge amount of time on its own, but across a business they can account for a significant amount of working time.

Our survey says...

For the Propel Tech Complexity Survey 2026 we asked 1,375 people working in UK organisations how complexity shows up in their working lives. More than half, 56.9%, said operational complexity had increased over the past two years, with 25.3% saying it had increased significantly. Just under a quarter said it had decreased.

There was no single source of the problem. Data spread across multiple systems was the most common answer at 16.6%, followed by disconnected systems at 15.7%. AI adoption and integration accounted for 13.5%, manual processes for 12.8%, compliance for 12.2% and communication between people for 10.4%. Legacy software was selected by 8.7%.

Taken together, the results suggest that the issue is less about one particular piece of technology and more about how the different parts of an organisation’s technology estate work together. Day-to-day operations were the area where respondents reported the most friction, at 21.2%, compared with 11.9% for working between departments.

How the hours add up

The amount of time people believe is being lost is significant. 68% of respondents estimated that their organisation loses six or more hours each week to manual processes, moving between systems, searching for information and similar activities. 43.9% estimated more than ten hours, while almost a quarter put the figure above 20 hours.

At an individual level, 56% said they personally lose six or more hours a week to operational complexity, with almost one in five estimating between 11 and 20 hours.

These are estimates rather than measured figures, so they should not be treated as a precise calculation of lost productivity. They do, however, give some indication of how much time people feel is being spent on the mechanics of getting work done. In a team of 50 people, six hours each would amount to 300 hours a week, equivalent to the working time of eight people on a 37.5-hour week.

When we asked what has the greatest impact on productivity, switching between multiple systems was the most common answer at 17.4%. Lack of integration followed at 14.5%, then limited use of AI and automation at 13.3%, repetitive manual tasks at 11.4% and waiting for approvals at 10.9%.

These are different symptoms of a similar problem. Information and decisions often have to be moved manually between people and systems.

Why it stays hidden

The cost of this kind of complexity is easy to overlook because it rarely appears as one large problem. Instead, people deal with it through small workarounds, such as reconciling reports, copying information between systems, checking several sources for the same data or chasing someone for an approval.

Over time, those workarounds become part of the normal process. People get used to them, new employees are shown how to do them and they stop being seen as something that could be improved.

That also makes the cost difficult to see in financial or operational reporting. Fifteen minutes spent reconciling two reports does not look significant on its own. Repeated across teams and every week, however, it becomes a considerable amount of time.

AI is making the problem more visible

Most of the organisations in our survey are already using technology to address some of this complexity. 89% said software had been used to reduce operational complexity, while 78.5% said the same of AI.

At the same time, investment in AI appears to be exposing problems elsewhere in the technology estate. When asked what their organisation's AI investment had uncovered, 25.3% identified systems issues, 20.4% data issues, 19.6% cost issues and 14.3% people issues. Only 9.6% said it had not uncovered any additional problems.

This is not particularly surprising. AI depends on access to reliable data and on systems being able to exchange information. Where those foundations are weak, introducing AI can make existing problems more apparent.

Starting with what you already have

When asked which investment would make the biggest difference over the next two years, connecting existing systems was the most popular answer at 20.4%. This was followed by embedding AI into existing software at 18.6%, automating manual processes at 15.7%, developing bespoke software at 13.2%, improving reporting and business intelligence at 12.4% and replacing legacy software at 11.6%.

There is also a clear willingness to invest. 87.8% of respondents are either exploring new software and AI capabilities or already investing in them, while 37.9% have budget allocated for the next 12 months.

For organisations trying to decide where to begin, it may not require a major technology project. A useful starting point is to take one process that regularly causes frustration and follow it from beginning to end. Look at how many systems are involved, where information has to be entered more than once and where work stops while someone takes action.

The people carrying out the process are often best placed to identify the workarounds that have developed around it. Their experience can help identify where connecting systems, automating a task or developing a small piece of bespoke software could remove some of the unnecessary work.

At Propel Tech, we work on these kinds of problems across technology estates, from integrating systems and modernising existing applications to developing software where an off-the-shelf product is not the right fit. 

If you are looking at where complexity is building up in your own technology estate, we can help you identify where there is a practical opportunity to reduce it.

Further reading

The hidden productivity problem isn’t always people (it’s technology) 

How much is organisational complexity really costing UK organisations? 

A complete guide to reducing business complexity with bespoke software