Across industries, a new behavior is emerging inside boardrooms and operating teams: “tokenmaxxing.” The term captures a growing tendency for businesses to aggressively deploy AI tools—not because they are strategically aligned, but because they signal modernity, efficiency, and innovation.
On the surface, these organizations look AI-forward. They have copilots embedded in workflows, dashboards tracking “AI usage,” and leaders publicly touting productivity gains. But beneath that veneer, many are facing a harder truth:
They are not generating measurable ROI—and in some cases, they are quietly eroding it.
This gap between perception and performance is becoming one of the defining execution failures of the early AI era.
The companies that ultimately win will not be the ones that generate the most tokens, deploy the most copilots, or claim to be the most AI-forward. They will be the ones that treat AI not as a technology layer, but as a lever inside a clearly defined growth model.
Because AI, on its own, does not create value.
It only amplifies it.
The Data Is Clear: Most AI Efforts Are Not Delivering Real Returns
The numbers behind enterprise AI adoption are increasingly difficult to ignore. Multiple large-scale studies converge on the same conclusion: most AI investments are not translating into business value.
A 2026 PwC CEO survey found that 56% of companies reported no revenue growth or cost reduction from AI over the prior year. A Forbes survey of over 1,000 executives showed that less than 1% had achieved significant ROI (defined as 20%+ impact), with most reporting only marginal gains in the 1–5% range. [forbes.com]
At the project level, the picture looks even worse. RAND and Gartner analyses consistently place the failure or underperformance rate of AI initiatives at 70–80%, meaning the majority never deliver the outcomes they were designed for.
But this does beg the question, if 56% of companies reported no revenue growth or cost reduction from AI, what are the other 44% of businesses doing right?
It comes down to strategy.
Why “More AI” Is Not Producing More Value
If you look beneath the surface of these numbers, a consistent pattern emerges. Most organizations are not failing because they are using AI incorrectly in a technical sense, they are failing because they are using it contextlessly.
The dominant model of adoption has been horizontal and tool-driven. Companies roll out general-purpose AI assistants across functions—sales, marketing, legal, operations—without defining an overall growth goal and redesigning the underlying workflows those tools are supposed to improve. The result is localized efficiency gains that rarely compound.
This is why so much of AI’s impact shows up in ways that never reach the P&L. Companies are optimizing for the wrong metrics. Instead of measuring revenue per workflow, cost per transaction, or throughput improvements, they track proxies: adoption rates, usage frequency, token consumption. These are easy to measure but weakly correlated with value.
The outcome is what McKinsey researchers have described as “AI theater”: widespread deployment without corresponding transformation of how the business actually operates.
Tokenmaxxing is simply the behavioral manifestation of that phenomenon.
The Core Failure: AI Without a Value-Creation Model
What separates the small subset of companies achieving meaningful returns from the majority that are not is surprisingly consistent across studies.
High-performing organizations do not just adopt AI, they anchor it to a clear business objective. McKinsey finds that companies seeing the greatest impact are far more likely to tie AI initiatives to growth and innovation goals, not just efficiency. [mckinsey.com]
They also do something most companies avoid: they redesign workflows. Only about 20% of organizations have fundamentally reworked processes to integrate AI, yet this is the single factor most correlated with value creation.
This distinction is critical. AI does not behave like traditional software. It does not simply layer onto existing systems and improve them incrementally. In many cases, realizing its value requires rethinking how work gets done in the first place—how decisions are made, how data flows, how tasks are sequenced.
Adopting AI without redesigning workflows is like introducing electricity into a factory that was built for steam power and expecting productivity gains without changing the layout of the machines.
Most companies have not made that leap. They are still operating steam-era processes with AI layered on top.
From Tokenmaxxing to Strategy-Led AI
The next phase of AI adoption is already beginning, and it will force a separation between companies that continue to signal and those that actually perform.
The shift is subtle but fundamental. It requires moving from AI-first thinking to strategy-first thinking.
Instead of asking where AI can be used, companies need to start with a different question: where is value created in the business, and how can AI amplify that?
In practice, that means identifying the handful of workflows that actually drive revenue, margin, or asset utilization—and concentrating AI investment there. It means designing systems, not deploying tools. And it means attaching every initiative to a measurable financial hypothesis before it is built.
This is exactly what the data shows high performers doing differently. They are more selective, more disciplined, and more integrated in their approach. They scale fewer use cases, but they scale them deeply.
