ClickUp, the collaboration software startup once valued at $4 billion, laid off 22% of its workforce last Thursday, but CEO Zeb Evans insists this isn't about trimming costs. In a post on X, Evans framed the move as a bold pivot toward AI, claiming most savings would flow back to remaining employees through what he called "million-dollar salary bands." The idea: workers who leverage AI to generate outsized results would be compensated beyond traditional pay structures. "If you create outsized impact using AI, you'll be paid outside of traditional bands," Evans wrote. The company has already deployed roughly 3,000 internal AI agents across its operations, according to a Fortune report. Rather than completing tasks themselves, ClickUp employees now oversee these agents and verify their output meets quality standards. Evans has publicly set his sights on transforming ClickUp into what he terms a "100x org"—a company where AI dramatically multiplies what each worker can accomplish. Research suggests ClickUp isn't an outlier. A recent Gartner survey found approximately 80% of companies deploying autonomous AI technologies have reduced their headcounts as a result. However, Gartner's findings indicate these workforce cuts aren't automatically translating into stronger financial performance. Evans told TechCrunch via email that ClickUp is indeed seeing measurable productivity gains from its AI agents and plans to incorporate those efficiencies into a future product offering for customers. "Instead of gamifying token cost, we gamify value created and time saved," he wrote. The broader trend has sparked debate over how companies should track AI adoption. Some firms have begun monitoring employee "token consumption"—essentially how much AI processing power workers use—as a proxy for AI engagement. Critics have dubbed this practice "tokenmaxxing" and argue it incentivizes racking up AI expenses rather than delivering real value. The tension between measuring AI adoption and measuring actual business impact remains unresolved, leaving companies like ClickUp to navigate uncharted territory as they push deeper into an AI-first operational model.