Agentic AI excitement runs ahead of results

BCA Research warned that investor enthusiasm for agentic AI is outpacing measurable commercial returns: a cited 2025 MIT study found 95% of corporate AI pilots failed to reach deployment with positive returns within six months, while a 2026 McKinsey survey found only 6% of respondents attributed at least 5% of operating earnings to AI. BCA estimated AI systems with human oversight could perform some tasks at about 75% lower cost than U.S. employees, but noted validation work and that only 20% of tasks completed by leading models meet reasonable client acceptance standards. The report said Claude Cowork's launch contributed to a 15% January decline in the MSCI U.S. IT Services index, but called the reaction excessive and said workflow restructuring could take at least a year.

Analysis

The key investment distinction is between AI capability and captured economics. Token growth can support infrastructure demand while simultaneously overstating end-user productivity; if validation remains labor-intensive, customers may resist paying for agents or reduce usage after pilots. That weakens the near-term case for monetization-led AI premiums, but does not by itself establish a durable decline in incumbent software value.

For CRM, INTU and WDAY, the second-order risk is less immediate displacement than pricing and seat pressure as buyers use AI to challenge renewal costs. Conversely, established workflow, data, controls and distribution may let incumbents bundle useful features and retain smaller customers while standalone agents struggle with reliability. The crucial evidence is renewal behavior and realized customer ROI—not launches or token consumption.

Over the next 1–3 months, sentiment and positioning can dominate; over 6–18 months, workflow redesign and reported AI contribution to revenue or operating earnings matter more. A genuine assistant breakthrough or independently verified reductions in labor and software spend would reverse the cautious thesis. Given the article provides no company-level adoption, churn or valuation data, avoid treating the sector move as proof of mispricing.

Key Decisions for Investors

• No immediate directional trade: the evidence supports skepticism toward AI monetization claims, but not a company-specific earnings downgrade. Verify CRM, INTU and WDAY disclosures on AI attach revenue, renewal rates, seat counts and customer retention before sizing exposure.
• Watch for a relative-value entry in established workflow software versus agent-led AI exposure (ANT), rather than shorting incumbents outright. Consider only if incumbent shares weaken without deterioration in renewals or guidance; define the thesis as workflow entrenchment and delayed substitution, not immunity from pricing pressure.
• Catalyst watch over the next 1–3 months: earnings commentary that quantifies paid AI adoption and customer ROI. Falsify the cautious view if deployments translate into measurable paid revenue or customer productivity gains; strengthen it if pilots fail to convert and customers cite validation costs.
• For CRM, INTU and WDAY, treat sustained seat contraction, weaker renewals, or AI-driven pricing concessions as the key downside alerts. A product launch or rising usage alone is insufficient confirmation of either durable displacement or successful monetization.