In a recent LinkedIn post, business leader Ben Eubanks discusses a critical issue emerging from Amazon’s reported struggles with AI adoption: the danger of incentivizing activity over actual outcomes. Eubanks highlights how employees, under pressure or through poorly designed systems, may focus on metrics like AI usage rather than the quality or impact of their work.
Eubanks opens by framing the situation as a predictable outcome for businesses not careful about how they measure AI integration. He points to a core principle of organizational behavior: leaders often say they want one thing but create systems that reward another.
“This is one of the most predictable HR stories of the year.”
Drawing a parallel to classic business theory, Eubanks references Steven Kerr’s seminal essay, “On the Folly of Rewarding A, While Hoping for B.” This essay, he explains, underscores the fundamental disconnect that can occur when stated goals diverge from actual incentives.
The Incentive Misalignment: Rewarding Activity Over Impact
According to Eubanks, the issue at Amazon, as reported, involves employees gaming the system. Instead of using AI to genuinely enhance their work, some employees reportedly engaged in behaviors designed solely to boost their usage statistics.
“Asking AI to summarize documents they had already read, breaking simple tasks into dozens of unnecessary prompts, repeatedly rewording requests to increase token counts, generating long outputs they didn’t actually need, and running extra AI queries simply to climb usage rankings.”
Eubanks argues that this behavior is not a failure of the employees but a consequence of flawed system design. When metrics like prompt volume, token usage, or AI activity are prioritized over tangible business results, employees naturally optimize for those visible metrics.
Goodhart’s Law in Action
The principle of Goodhart’s Law, as cited by Eubanks, is particularly relevant here: “When a measure becomes a target, it ceases to be a good measure.” He elaborates that the focus on AI adoption itself, rather than the improvements it should facilitate, can lead to superficial engagement.
“AI adoption is not the goal. Better work is the goal.”
Eubanks stresses that the true value of AI lies in its ability to drive meaningful improvements such as reducing bottlenecks, enhancing communication, automating reporting, identifying risks, and saving significant time. These outcomes are often achieved by employees who might not have the highest usage numbers but are effectively leveraging AI for substantial gains.
A Warning for Businesses
The insights shared by Eubanks serve as a critical warning for organizations worldwide that are currently implementing AI strategies. Many companies, he notes, are tracking AI logins, prompts submitted, and tool adoption rates without adequately assessing whether the work itself has actually improved.
“HR teams helping shape AI strategy need to be very careful not to build incentive systems that reward theater instead of transformation.”
Eubanks concludes by emphasizing that the objective should always be to foster genuine transformation and better work through AI, not merely to create the appearance of adoption. He credits Paul Hebert for his expertise in incentive design, further underscoring the importance of well-structured reward systems in achieving desired business outcomes.
📝 About This Content
This article is based on insights shared by Ben Eubanks on LinkedIn.
📅 Originally posted on May 17, 2026 | View original post on LinkedIn →