Businesses are beginning to take a harder look at their artificial intelligence spending as the era of heavily subsidised AI usage comes to an end, forcing companies to determine whether the technology is actually delivering enough value to justify its cost.
Stu Dorman, Chief AI Officer at customer experience technology company Sabio Group, said organisations are moving away from a period when employees could use AI extensively without much concern about the financial consequences. According to him, the change became more noticeable around late May and early June, when major AI providers began withdrawing some of the subsidies that had helped keep usage costs down.
For businesses that had become accustomed to widespread AI use, the impact is straightforward: more usage can now mean a more noticeable bill. AI models charge based on tokens, the units used to process information, making the cost of an organisation’s AI activity closely tied to how heavily its employees and systems use the technology. Brandspur Brand News reports that companies are consequently being forced to ask a question that was sometimes overlooked during the early AI boom: what exactly are we getting in return?
The shift comes at a difficult moment for the wider AI industry. Technology investors have faced a period of declining share prices, with a benchmark semiconductor fund falling by almost 10 per cent over the past month. At the same time, some of the boldest predictions surrounding AI, including claims that the technology would rapidly eliminate large numbers of jobs, have faced growing scrutiny.
Dorman believes part of the problem is that expectations were set extraordinarily high. Businesses were encouraged to imagine a future in which AI could transform almost every aspect of work, but adoption inside many companies has not moved at the same speed as the excitement surrounding the technology.
Competition is also making the calculation more complicated. New Chinese AI models are offering alternatives to American systems at substantially lower prices, giving companies more reasons to compare performance and cost before deciding which models to use.
For some organisations, the reassessment may be uncomfortable. Dorman acknowledged that companies spent heavily on AI in the early stages, with some even encouraging employees to increase their usage. More AI activity, however, did not always translate into better results.
That is why he believes the focus now needs to move from consumption to usefulness.
Instead of asking employees to use AI simply because it is available, businesses should help them identify specific parts of their jobs where the technology can save time, reduce repetitive work or improve the quality of a service.
The distinction is particularly important in customer service, where the benefits can be easier to measure. AI can reduce the amount of time an employee spends handling a customer interaction, or, in some cases, remove the need for that interaction altogether. For a business handling thousands of customer enquiries, even a small improvement can have a significant effect.
Also read: https://brandspurng.com/2026/08/12/nigerbev-names-chibueze-iwuoha-portfolio-marketing-manager/
The changing economics of AI may therefore force companies to become more selective. A worker who uses an AI tool dozens of times a day is not necessarily creating more value than one who uses it only a few times for tasks where it genuinely makes a difference.
Dorman remains optimistic about the technology despite the current uncertainty. He expects AI adoption to accelerate significantly over the coming years, although he believes that growth may be slower than major technology companies would like.
He also cited a Gartner forecast that AI could become a net creator of jobs by 2029, pointing to the possibility that increased productivity could eventually generate new employment opportunities rather than simply eliminate existing roles.
For ordinary workers, the changing approach could mean that the question is no longer whether their employer is using AI, but how intelligently it is being used. As the cost becomes harder to ignore, businesses will have to prove that their AI investments are helping people work better, serve customers faster and achieve results that can be measured.
