Companies shifting to artificial intelligence-led operating models are facing bigger-than-expected cost burdens.

Tech outlet TechRadar reported on July 9 that a survey by British management consulting firm KPMG found 29 percent of corporate executives said they found it difficult to understand rising AI operating costs within their organisations.

The survey covered 2,145 senior leaders in 20 countries. It found companies feel burdened when trying to gauge AI operating costs and economics as they scale AI across their organisations. A similar share of respondents also cited a lack of understanding of costs and economics as a key challenge in adopting AI agents.

The trend coincides with a period when major AI companies such as Anthropic and OpenAI shifted some services from flat-rate pricing to usage-based billing. Companies still lack sufficient capabilities to predict, monitor and manage AI spending.

Cost issues were also linked to responsibility for technical operations. In many cases, executives had not clearly decided who should be accountable when AI models hallucinate or produce errors. KPMG viewed human leadership accountability as important, but said governance ultimately hinges on day-to-day operating practices.

In practice, many workplaces had not fully embedded standards for intervention, cost ownership, output review procedures and responses to disruptions. KPMG said, "Clear rules are needed on when employees can intervene, who is responsible for AI-related costs, how AI outputs are reviewed, and what actions are taken in the event of system disruptions." Many organisations said they had at least minimal governance mechanisms, but not many said they had fully put them in place.

When costs exceeded expected benefits, companies also adjusted their adoption strategies. The survey found nearly half of organisations rewrote AI deployment schedules or adjusted the pace of implementation. KPMG said such steps did not mean weakening trust in AI. KPMG said, "There is a growing movement to more strictly distinguish between areas where AI creates real value and where it does not."

The survey also found an assessment that gaps between companies are widening based on responsible leadership and the ability to control investment. Steve Chase (스티브 체이스), head of global AI and digital innovation at KPMG International, said a clear difference is emerging between organisations where accountability is clear at the top and those where it is not. He said such companies show higher trust, value realisation and established return on investment.

As AI adoption spreads, companies' challenges appear to be shifting from technology adoption itself to cost forecasting, operational accountability and setting investment priorities. As a result, it is expected that how precisely companies manage profitability and governance, rather than the scale of AI adoption, will become the benchmark that differentiates corporate performance.