The AI metric one company says managers should stop caring about
Companies are pushing workers to use AI. Synthesia's Laura Gonzalez says measuring token usage can reward activity over outcomes.
Companies are pushing workers to use AI. Synthesia's Laura Gonzalez says measuring token usage can reward activity over outcomes. This report comes f
Read Full Story at Business Insider Mkt →Why This Matters
The fixation on token usage as a performance metric for AI adoption reflects a dangerous misalignment between corporate incentives and actual productivity gains. While executives chase measurable activity, workers may be churning out unchecked AI outputs that lack real business value—mirroring the empty "vanity metrics" that plagued early social media marketing. The debate cuts to the heart of whether companies are investing in transformation or just performing the illusion of it.
Background Context
The pressure to integrate AI into workflows has accelerated alongside the rise of no-code and low-code tools that democratize access to generative models. Yet many organizations still cling to legacy productivity frameworks, where usage metrics like tokens consumed or hours spent in AI platforms serve as proxies for efficiency—despite their poor correlation with outcomes. This phenomenon isn’t new; similar misallocations of resources emerged during the cloud computing boom, where "migration counts" often eclipsed actual performance improvements.
What Happens Next
As skepticism grows about token-based evaluations, companies will likely pivot toward outcome-driven KPIs—though defining and measuring these will prove contentious. Regulators may eventually weigh in if inflated AI usage statistics mislead investors about operational efficiency. Meanwhile, employees caught between old metrics and new expectations could face renewed scrutiny over their actual contributions versus perceived productivity theater.
Bigger Picture
This reckoning with AI metrics is part of a broader correction in how organizations measure digital transformation, moving beyond vanity counts toward tangible business impact. The trend mirrors past cycles where metrics like "page views" or "followers" yielded to deeper engagement or revenue-driven assessments. As generative AI matures, the industry may finally confront the overarching question: Are we optimizing for real progress or just the optics of innovation?
