Who loses power when work is automated, and why that matters more than budget

When a small business decides not to automate a process, it is rarely because money is missing. More often, someone somewhere in the chain stands to lose something no spreadsheet captures: control, visibility or the reason their job is needed.

This is not a moral judgment about that person. It is a predictable mechanism, and understanding it changes how an automation project should be approached.

Technology that threatens expertise, not just a job

Harvard Business School researchers Das Narayandas and Shunyuan Zhang have studied why some technologies provoke especially strong resistance. Their conclusion is that certain tools do not fail because of their performance. They fail because they change the professional role of the person using them. The researchers call them self-disruptive technologies. They can improve a process while making the person who used to do the work feel less expert or less visible in their own job.

This explains something seen constantly in small businesses: the person who knows a process best is often the least enthusiastic about automating it. Not because they are ignorant or lazy. Often, that person is the only one who truly understands how the process works, and that opacity is part of their value to the company.

Power is more than hierarchy. It also means control over information

A recent academic review of resistance to digital transformation identifies a consistent pattern: resistance appears when employees believe a technology will reduce their power and autonomy within the organisation. That power does not always mean job title or salary. Often it means being the person everyone asks, the one who approves an exception, or the one who spots a problem before it becomes a fire.

Automating a process does more than save time. It redistributes who has the information and who can make decisions with it. That redistribution provokes a real emotional reaction: the same review says that a perceived loss of professional identity and recognition can reduce commitment to the company, even when the tool objectively works better.

Why this holds things up more than cost

The reason this resistance weighs more than the budget is simple: money is a visible, negotiable obstacle. You can look for funding, adjust the scope or spread the payments out. Fear of losing relevance in the company cannot be negotiated through a terms sheet. If it is to be managed, it takes a direct conversation and a clear role where the person can keep contributing what they know.

A World Economic Forum study cited in several 2026 HR analyses points in the same direction at a broader level: a large share of employers plan to reduce headcount because of AI in the coming years, while most employees admit they do not feel confident using these tools in their daily work. That combination of perceived threat and lack of confidence turns a technically simple project into a long organisational process.

What changes the outcome

The Harvard researchers sum it up in a line worth remembering: the machine does not need to be convinced or motivated, but the person does. The more time a manager spends considering how the tool benefits the person who will use it, the more likely the technology is to be adopted for real.

In practice, before automating a process, ask who knows it best today, how much of their value to the company depends on keeping it opaque, and what new role they can take on when that part of the process no longer depends only on what is in their head. Without that conversation, the tool can work perfectly in a demo and never be used in production.