There’s a Mass Rebellion Against AI in the Workplace

A WalkMe survey of 3,750 executives and employees found 54% of workers avoid their company’s in-house AI tools, while one-third never use AI at all. Only 9% of workers trust AI for complex business-critical decisions versus 61% of executives, and employees are spending eight hours per week cleaning up AI-related errors, equal to 51 work days per year. The article underscores growing workplace skepticism toward enterprise AI deployments and weak realized productivity gains.

Analysis

The market is still treating enterprise AI as a broad productivity unlock, but the more immediate read-through is margin leakage: when employees route around sanctioned tools, enterprises end up paying twice — once for software and again for manual rework, training, and governance. That shifts the investment debate from adoption to utilization, which is negative for vendors whose valuation depends on rapid seat expansion and high net retention. SAP is modestly exposed because weak AI satisfaction can slow incremental module attach rates and lengthen sales cycles for “AI-enabled” workflow upgrades.

The second-order effect is that this is more of a procurement and trust problem than a model-quality problem. If workers don’t trust outputs for high-stakes tasks, management will increasingly constrain deployment to low-value use cases, capping ARR upside and delaying the operating leverage narrative. That hurts the whole enterprise app stack: point AI tools, workflow automation, and consulting/service integrators that were pricing in a faster refresh cycle.

Near term, the catalyst path is asymmetric to the downside over the next 1-2 quarters if another survey or earnings commentary confirms that AI spend is not translating into measurable labor savings. The contrarian risk is that this becomes a configuration issue rather than a demand destruction story — if vendors pivot to narrower, embedded use cases, adoption can improve without headline enthusiasm. But until there is hard evidence of cycle-time reduction or headcount leverage, the burden of proof is on the bulls.

Consensus is likely overestimating the speed at which enterprise AI converts into budget expansion, but underestimating how quickly buyers can freeze discretionary AI spend after one bad implementation. That makes the setup less about a structural collapse and more about a pause in monetization slope; multiples that assume 20%+ AI-driven software growth are vulnerable if bookings commentary softens. The best opportunities are relative shorts in companies with the most aggressive AI monetization rhetoric and the weakest proof of workflow savings.

Key Decisions for Investors

• Short SAP into any strength over the next 2-6 weeks; downside is modest but the setup is for slower AI-led module attach and softer sentiment if enterprise buyers keep pushing deployments to the right.
• Pair trade: long mature workflow/software names with proven automation ROI, short AI-hype beneficiaries with limited evidence of conversion; favor a 3-6 month horizon where utilization data should matter more than narrative.
• Buy puts or put spreads on high-multiple enterprise software names that rely on AI upsell assumptions into the next earnings cycle; risk/reward improves if management commentary shifts from adoption to experimentation.
• Avoid initiating new longs in enterprise AI platform names until there is visible proof of labor savings in customer case studies; the upside can re-rate later, but current probability of disappointment is higher over 1-2 quarters.
• Use the next two earnings seasons as the key catalyst window: if productivity metrics remain absent, rotate out of AI-exposed software beta and into incumbents with clear cost-takeout use cases.