- Bosses seem confused that mass AI in favor of human workers is not always a success
- AI operating costs are often higher than expected
- However, some businesses are happy to adapt and refocus if necessary.
Bosses are confused by the high cost of switching to AI-centric models, and many seem perplexed that replacing human workers with agents won’t instantly save them huge sums of money, according to a new study.
A new report from KPMG reveals that nearly a third of business leaders reported difficulty controlling the costs of operating AI in their organization.
The news comes as several major AI vendors, including Anthropic and OpenAI, have shifted some services toward usage-based billing, rather than flat-rate subscriptions, in recent months.
Making AI effective
The report enabled KPMG to survey 2,145 senior executives across 20 countries. It found that 29% struggled to understand rising operating costs as they sought to scale AI in their business, with a similar proportion also highlighting a limited understanding of AI costs and economics as a key challenge in deploying AI agents.
“As usage-based pricing models become more common, many organizations continue to develop the capabilities needed to effectively predict, monitor and manage AI spending,” KPMG said.
When things go wrong, the report highlights that executives are often unclear about who should take responsibility, particularly in cases of hallucinations or errors from AI models.
He emphasizes that the reason it is important for human leaders to be accountable is that “governance ultimately succeeds or fails through daily operational practices.”
“Organizations need clear rules about when employees can intervene, who bears AI costs, how AI results are reviewed, and what happens if systems fail. While most organizations report having at least some governance mechanisms in place, relatively few describe these practices as fully integrated,” the report states.
When costs exceed expected value, the report reveals a surprising degree of contrition from participants, with nearly half of organizations reporting they have rephased AI deployments in this case.
“These actions do not reflect a loss of confidence in AI,” the report warns. “Rather, they suggest a growing desire to assess where AI creates significant value and where it does not. Organizations appear increasingly focused on concentrating investments where the expected returns are strongest.”
“We see a clear divide between organizations with leadership responsibility at the top and those without it,” added Steve Chase, KPMG Global Head of AI and Digital Innovation, KPMG International.
“These companies achieve significantly better results across the board, such as greater trust, higher value realization, and established ROI.”
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.