Why Ill-Fitting Business Software Creates Hidden Work and When to Build Custom Tools
- Cargo
- Machine Learning
- Shipped
- Shipper
- Mukund Jha
- Readers
- 650
Workarounds can make imperfect software costly in staff time. AI-assisted development adds custom apps to the options for fixing persistent gaps.



Software that nearly fits can quietly become part of a company’s workload. Staff bridge gaps with spreadsheets, repeat data entry and extra steps that seem manageable one at a time. The cost is less visible than a subscription fee, but it can grow as the business adds customers, systems and employees.
A survey of business owners found that only one in seven felt well served by the software available to them. Among those who were poorly served, 27% relied on manual processes that had become harder to sustain as they grew, and another 27% used software-as-a-service products that didn’t match their needs. Those figures point to a practical problem: a tool can be affordable to buy while still consuming employee time every day.
That doesn’t make custom software an automatic fix. Building an application brings costs, delivery time and ongoing responsibility. The decision starts with understanding how work actually gets done, including the small accommodations employees may have stopped noticing.
Workarounds can hide the real problem
Smaller organisations often assemble their software stack gradually. One product handles an immediate need; another arrives when the business expands or a new task appears. Each choice can be reasonable on its own, yet the combined setup may leave staff transferring information between systems or remembering steps that no product handles.
These routines can become institutional knowledge. One employee knows which record must be updated after a sale, or which spreadsheet holds the numbers that another tool needs. Colleagues learn the sequence informally, and the extra work starts to look like a normal part of the role. That makes it easy for a persistent software gap to go unreported.
It also creates a continuity problem. When a process depends on one person remembering the details, their departure can take that knowledge with them. Before adding another product, owners and technology leads should map the full process and ask where staff spend time correcting, copying or reconciling information. The aim is to identify the source of the work, not just the most visible inconvenience.
That review can show whether a product already in use can be configured to cover the gap, whether a different standard tool fits better, or whether the business has a need that available products don’t address. In the survey, 33% of poorly served businesses said no workable solution was available to them. That finding supports exploring a custom build in some cases, while leaving open whether a particular company can justify the effort.
AI-assisted development changes who can shape an app
Custom development has typically meant hiring outside specialists or committing internal engineering time. Agency estimates for a basic build can range from $5,000 (£3,728) to $100,000 (£74,541), with a median of $20,000 (£14,903). Delivery may take several months or more than a year. For a small company, both the budget and the wait need to be weighed against the ongoing cost of the current workaround.
There’s another difficulty: business knowledge doesn’t always translate neatly into a specification. A founder may know that a task needs different handling in certain situations without having written down every condition. If those details are missed, a new application can reproduce the same gap in a different form.
AI-assisted development offers a more direct way for people who know the business to take part in shaping an application, even if they don’t know how to code. They can describe what a process should do, inspect the result and refine it as they spot missing cases. That involvement matters because the people closest to a workflow often understand its exceptions better than someone working from a brief alone.
Survey results suggest that businesses are already turning to alternatives: a third had built something to address their needs. Among the reported outcomes, 21% had retired a manual process, while a third had created a tool for which no workable solution had existed. The figures illustrate why building can appeal when standard products leave a persistent gap; they don’t show that every custom app will save time or money.
Development also continues after an app first works. Employees will encounter cases that weren’t anticipated, and business processes can change with growth. A way to adjust the software as those needs become clear can help keep the tool aligned with the work it supports.
A custom app brings operational duties
Once employees depend on an application, maintaining it becomes part of running the business. Every application analysed used a database by default, and about 59% connected to at least one external service. Owners therefore need to understand what information the app stores, which other services it relies on and what happens to day-to-day work if one component fails.
Access and change controls need attention, too. The business should know who can view information and who can alter the application. A named owner can keep it useful and coordinate changes, while another trusted person should be able to step in if that owner is unavailable. This basic ownership helps avoid a tool that works only because one employee knows how it works.
Documentation and operating procedures matter, but they should match the scale and purpose of the application. The goal is enough shared knowledge to maintain the tool, manage access and respond to problems without turning upkeep into a separate burden.
Buying, configuring and building are all practical options, depending on the gap. An established product makes sense when it covers the need. Configuration may be enough when the product is close. A custom application can be justified when a recurring problem has no workable alternative and the company can take responsibility for the software over time.
The useful measure is whether the change gives employees more time for their actual work. Before choosing a path, map the hidden steps and the people who keep them running; that evidence makes the cost of an imperfect fit—and the value of fixing it—easier to judge.