AI investment doubled in the last year, according to the Royal Bank of Canada. Boards signed off, strategies were announced, and enterprise AI tools arrived.
Yet something unsettling is emerging: The signals below suggest that the humans and organizations responsible for leading AI transformation are struggling to keep pace. Although the machines are ready, workers are resisting, leaders are outsourcing human judgment, and CEOs are stepping aside.
Institutions are cracking under the very human friction we’ve been hoping AI would remove. The problem isn't technology; it’s turning out to be a very human one.
What: The ServiceNow Enterprise AI Maturity Index 2026 found that across 6,500 executives and employees globally, 59% of organizations now use agentic AI, “but only 9% have made meaningful progress building autonomous, multistep workflows.” AI-enabled workflows scored the lowest of seven pillars measuring AI maturity, even as confidence and spending surged.
So what: Most organizations are running agentic AI as a set of disconnected assistants, not as a coherent operating layer. The gap between deploying AI and transforming with it is significant and growing. Buying a capability and building the organizational, data, governance, and cultural foundations to use it are two different undertakings. Organizations that haven't grasped this are compounding risk at scale, running powerful tools on fragile infrastructure.
What: A report by generative AI company Writer and research firm Workplace Intelligence found that “29% of employees—including 44% of Gen Z—admit to sabotaging their company’s AI strategy,” including using unapproved tools, ignoring guidelines, and avoiding training.
So what: Workforce resistance is a signal that AI is arriving before the cultural and structural foundations are in place for humans to receive it. Just signing people up won’t cut it. Even training fails when it doesn’t sync to outcomes. Eighty-five percent of employees say their AI training isn't helping them do their jobs, according to research by AI workforce readiness company Docebo.
Consider augmenting human capabilities with role-specific, use-case-driven training, not just forcing AI onto a nervous workforce.
What: A Confluent study of 200 C-suite executives uncovered that 62% are using AI to make most decisions and that 140 of those surveyed reported second-guessing their own instincts when they conflict with AI's recommendations.
The cognitive stakes are rising. EEG research from MIT shows that heavy AI users exhibit persistently weaker neural engagement even after they stop using the tool (a phenomenon dubbed "cognitive debt"). Workers themselves are reporting what BCG researchers have documented as AI “brain fry”—acute cognitive fatigue from managing multiple AI systems simultaneously.
At the very top, the CEOs of Coca-Cola, Walmart, and Adobe have cited AI as a factor in stepping down, concluding they’re not the right leaders to finish the AI transformation their companies need. As Walmart's departing CEO put it, "I could start this next big set of transformations with AI, but I couldn't finish."
So what: The compounding dynamics here are worrisome: The leaders responsible for governing AI are increasingly relying on AI to do their thinking. The most candid among them are acknowledging they don't have what the next phase requires.
CEO turnover in S&P 1500 companies hit a 15-year high in 2025, according to leadership consulting firm Spencer Stuart. Companies named 168 new CEOs last year.
The leadership layer is under pressure in a way that’s only now becoming visible. This isn’t a warning about AI replacing leadership. It’s a warning about what happens when leaders don't actively develop and foster the human judgment and experience that makes AI business reinvention possible.
Researchers tracking AI-enabled pro se legal filings reported that AI is enabling people to file federal lawsuits without lawyers at an unprecedented rate. This is one of many systems whose continued operation depends on a mechanism to keep humans from overwhelming it.
Separately, an AI system denied a Medicare claim without adequate human review, triggering litigation. The case illustrates what happens when AI makes consequential decisions inside systems that were never designed for autonomous action.
So what: Legal filings, insurance decisions, hiring processes, and government services were governed, in part, by the assumption that scale required effort and that effort created natural limits. AI removes that assumption. The institutions and organizations that don't redesign for this reality face liability, loss of trust, and failure precisely when their systems are tested most.
What: A joint INSEAD and Harvard field experiment across 515 high-growth startups found that firms guided to discover broader AI use cases saw a 44% increase in applications and generated 1.9 times higher revenue, completed 12% more tasks, and were 18% more likely to acquire paying customers.
So what: The researchers found the central obstacle to AI adoption is not access to models or budget, but rather what they term a "mapping problem": the challenge of “discovering where and how AI creates value within a firm's production process.” Most organizations don't know where AI could create genuine value across their operations.
The firms that solve this, systematically discovering and connecting use cases rather than deploying isolated tools, compound their advantage over those that don't.
This mirrors the pattern that emerged from ServiceNow’s Enterprise AI Maturity Index. The organizations pulling furthest ahead aren't defined by their AI budget or industry. They're defined by connected data, governance built before deployment, and the organizational discipline to ask not just "where can we use AI?" but also "where should we redesign work and outcomes?"
These five signals show different aspects of the same challenge: Organizations have invested in AI capabilities, but they haven’t yet built the human and structural foundations to use them effectively.
AI spending is surging. Confidence has rebounded. Yet solving the hardest problems (mapping, governance, workforce readiness, and judgment) is being deferred in a race to simply add more AI.
What separates the organizations generating meaningful returns from those accumulating compounding risk is a series of slower, less celebrated choices. These organizations unify data before deploying agents, build governance into systems rather than bolting it on after, invest in the kind of ongoing human development that prevents cognitive debt, and redesign processes for a world in which AI removes the friction that once served as informal governance.
The machines are ready, but success won't come from the machines alone. Prioritize human investment if you want your AI investments to pay off.
Find out how ServiceNow can help you proactively solve problems and drive exponential productivity.
