AI cost pressures put focus on governance, executives argue
Teneo and Thoughtworks leaders say companies risk wasting AI budgets if they chase usage without tighter controls on value and cost.
By Daniel Okafor · Business Editor
3 min read
Corporate AI adoption has reached a more expensive stage, with usage-based pricing turning heavy use of large language models into a potential drag on margins, the chief executives of Teneo and Thoughtworks argued in a Fortune commentary. They said companies that respond with broad restrictions could undermine the productivity gains they have already built.
The executives focused on tokens, the units used to measure and price many AI services. They said enterprise AI costs have shifted away from fixed subscriptions toward pricing tied to consumption, making AI spending harder for companies to predict as models become embedded in daily work.
According to the executives, many companies have moved from proving they can use AI to proving they can use it profitably. They described the current period as a sustainable adoption phase in which companies will be judged by how well they control AI use while still creating business value.
Cost pressure follows early assumptions
The executives said two assumptions helped bring companies to this point. One was the belief that businesses could replace some labor costs with model costs and keep the savings as profit. They argued that the exchange has been less direct because some companies cut workers whose knowledge was needed to make AI systems useful.
As an example, they cited Bloomberg reporting that Ford needed to rehire hundreds of engineers after AI tools fell short in quality-control work. According to the commentary, Ford executives pointed to a shortage of veteran expertise as a factor that hurt product development and limited efficiency gains from autonomous systems.
The second assumption involved the timing of investor expectations. The executives cited Teneo’s annual CEO and investor survey, which found that 53% of investors expected returns from AI within six months, while 16% of large-cap CEOs said they believed they could meet that timeline.
Executives call for tighter AI controls
The Teneo and Thoughtworks leaders said companies should stop measuring AI primarily by spending levels and should instead ask where AI investment creates durable advantages. They argued that spending that produces revenue, improves customer experiences or builds company-specific capabilities should be treated differently from spending that automates lower-value work.
They also urged companies to match AI tasks with the least expensive model that can do the job reliably. Employees may prefer the newest models even when older or cheaper systems can produce the needed result, the executives said. One option they described is software that routes prompts to the appropriate model automatically.
The executives warned against rewarding employees for using AI as much as possible. They also said blunt caps on use can discourage experimentation. Instead, they called for incentives based on efficient use, or better outcomes with the right level of AI consumption.
Governance should be spread across companies rather than assigned only to a chief AI officer, the executives wrote. They said managers in different functions need responsibility for sustainable AI adoption, while employees need guidance on how their use is assessed and what skills they need to work effectively with AI.
The executives concluded that the companies best positioned in the next phase will be those that consistently convert AI consumption into economic advantage. Their argument puts governance, rather than speed or raw spending, at the center of the corporate AI race.
This story draws on original reporting from Fortune.