The ‘Return on Tokens’ Paradox: John Cutler on AI’s ROI Scrutiny

In a recent LinkedIn post, John Cutler discusses the shifting focus on Return on Investment (ROI) as Artificial Intelligence (AI) tools become more prevalent in the business world. Cutler highlights a perceived hypocrisy in how organizations measure value, suggesting that the intense scrutiny on AI’s financial impact is a relatively new phenomenon driven by potential cost-saving measures.

According to Cutler, the drive to quantify AI’s benefits, often termed “return on tokens,” is directly linked to organizational decisions regarding layoffs and restructuring. He points out that this sudden demand for measurable ROI stands in stark contrast to how investments in other areas have historically been evaluated.

Before AI …. oh, understanding ROI is just *so* hard …
After AI … we’re working feverishly to understand our return-on-tokens

John Cutler argues that the urgency to define and measure the ROI of AI stems from its potential to automate tasks and substitute labor. This capability, he suggests, makes AI a prime candidate for cost-reduction initiatives, thereby necessitating a clear justification for its adoption.

The Ambiguity of Past Investments

In his post, Cutler contrasts the current pressure to quantify AI’s value with the historical tolerance for ambiguity surrounding the ROI of various business functions and initiatives. He notes that investments in areas not directly tied to immediate labor substitution often faced less rigorous financial evaluation.

“Organizations tolerated enormous ambiguity around the ROI of collaboration tools, management layers, process changes, meetings, transformation programs, and even entire functions,” Cutler writes, emphasizing the selective nature of financial accountability.

A Shift Driven by Restructuring Potential

The core of Cutler’s argument is that the “measurability” of ROI for AI is not an inherent characteristic of the technology itself, but rather a consequence of the organizational decisions it enables. When AI presents a plausible path toward reducing headcount or reorganizing departments, the demand for concrete financial justification intensifies.

But once AI creates a plausible path to labor substitution or organizational restructuring, suddenly there is intense pressure to quantify the return.

As John Cutler observes, this pressure leads to a concentrated effort to establish metrics and track performance, particularly focusing on the ‘denominator’ – the costs associated with implementing and running AI systems. This focus on the cost side of the equation is crucial for justifying efficiency gains.

Watching the Metrics Closely

Cutler concludes with a cautionary note, urging observers to pay close attention to how these new metrics are being developed and applied. He implies that the framing of “return on tokens” might be heavily influenced by the desire to justify cost-cutting rather than a pure, objective assessment of value creation.

“They’re calling it ‘return on tokens,’ but watch carefully: a lot of the instrumentation is going toward the denominator,” Cutler warns. This suggests that the focus might be more on demonstrating cost savings than on uncovering new revenue streams or productivity enhancements that don’t directly lead to headcount reduction.