As highlighted in the 2026 ITPro Future Focus Report released early this year, the most eagerly anticipated development in the IT world is the industry-wide shift from implementing AI to increase productivity and efficiency, to driving measurable business value.
But what originally began as a desire to ensure use case success and provide a competitive edge is now turning, for many, into a battle to keep costs from spiraling.
The tokenmaxxing trap
Much has been written about the recent case where Uber was forced to implement monthly user caps on AI use, after the company found its entire AI yearly budget spent in just four months, with one user reportedly spending $3,400 in just 24 hours. The news sparked the trend called ‘tokenmaxxing’, whereby tokens consumed by employees have become the main factor taken into consideration when it comes to measuring usage, and controversially for some companies, keeping tabs on employee ‘productivity’.
Therefore, in addition to concerns over ROI, or simply the odds of AI projects failing, there’s now a growing unease across the industry over the potential mounting costs of simply running AI projects at scale. This phenomenon is widely referred to as ‘tokenomics’, and it will only get more intensified as rising capabilities of new iterations of agents go hand-in-hand with rising costs.
For CIOs and CTOs, factoring AI in their yearly budgets is then becoming a real headache: tough to measure efficiency, tough to control spend, tough to anticipate use costs over time. So what’s a leader to do?
The Agentic Economy
The answer is, as with previous historical technological leaps, is the shift in workforce model. The companies that may be getting it right have used AI for pilot projects to strategically fine-tune their approach to a hybrid model, where AI becomes integral to staffing and is a consideration in OpEx budget-cycle level. Striking the right balance between headcount and AI is likely to be a different proposition for most businesses, though, as leaders strive to pinpoint the use cases where the added value is at its most business-relevant.
This shift would likely usher in a new era of value-centred strategy affecting the way CEOs determine in which strategic direction to take companies, COOs and CFOs to decide how to optimize their team build, and CIO and CTOs to decide on tooling and workflows.
Agentic AI adoption brings in a new learning cycle that will provide an ideal opportunity for companies, and vendors who work with LLM providers, to optimize business value and help making the right decision when it comes to selecting agents
In ITPro’s recent podcast, ‘The AI pricing timebomb’, Mendix CEO Raymond Kok spoke about a new learning cycle that will provide an ideal opportunity for companies, and vendors who work with LLM providers, to optimize business value and help making the right decision when it comes to selecting agents to build AI infrastructure and SaaS solutions as primary services, where so far companies tend to be oversubscribed with vendors and solutions.
Changing dynamics
So, where will this shift leave us as a human workforce?
Industry research predicts an evolution of the job market instead of a simple substitution of the current human workforce. The key for business leaders is efficiency: with an expected increase in productivity, a rise in demand and usage can also be accounted for. This is a phenomenon called Jevons Paradox, and it is now a popular byword in AI conversations.
With increased demand, simply substituting roles is counter-productive; rather, leaders are looking to ‘augment’ jobs with AI supporting and accelerating mundane tasks, freeing workers to expand their remits over entire portfolios and take on wider specialisms, wherein many are now siloed inside broken-up workflows.
The key focus for many companies revolves around AI literacy, as senior and manager level jobs will need to gear team dynamics towards humans and AI working hand-in-hand on projects and daily operations. This entails dedicated training programs aimed at increasing safety and productivity, but also a general reframing of the working environment where successful symbiosis is rewarded and cited as example. The aftereffects of this evolution are manyfold and will develop in time; seniority, performance, accountability and other important factors are still unknown factors.
For the tech industry, this is a period of momentous change with many businesses looking for expertise in business software, performance analysis, automation, talent management, training and upskilling to inform their growth strategy. And for vendors of IT solutions, this is a genuine opportunity to shape the way entire operational systems work, not only IT teams.
