Last year, your company increased its AI budget. So did 92% of companies surveyed by McKinsey. This year, PwC surveyed 4,454 CEOs and found 56% report AI has delivered neither increased revenue nor decreased costs.
The spending went up. The results didn’t follow. And now the trust is collapsing.
Between May and July 2025 alone, employee trust in company-provided generative AI fell 31%, according to workplace research from Sweep. Trust in agentic AI systems dropped 89% in the same period. The speed of that decline matters. This isn’t gradual skepticism building over years. This is a correction happening in months.
The pattern should look familiar. Companies adopted AI the same way they adopt most technology: tool first, strategy never.
The Maturity Illusion
The gap between AI ambition and AI reality is staggering.
McKinsey’s 2025 “Superagency” report found 92% of companies plan to increase AI investments over the next three years. Only 1% of leaders call their company “mature” on the AI deployment spectrum — McKinsey’s word, not mine. What they actually mean is that AI is integrated into workflows and driving measurable outcomes. Nobody is “mature” in a technology that reinvents itself every quarter. But only 1% have even reached the point of operationalizing it.
That’s a 91-point gap between intention and capability.
The numbers cascade from there. According to Index.dev, 82% of organizations plan to adopt AI agents within one to three years, but only 2% operate agents at full enterprise scale. Alteryx surveyed 1,400 business and IT leaders and found fewer than one in four AI pilots successfully reach production. Gartner predicts organizations will abandon 60% of AI projects due to lack of AI-ready data.
The pattern is consistent across every study: massive investment, minimal operationalization. Companies are buying AI like they buy martech. As I explored in “Your Martech Stack is a Frankenstack,” tool proliferation without strategic integration creates expensive paralysis. The AI stack is becoming the new Frankenstack.
Why the Trust Is Breaking
The trust collapse has three layers, and each one compounds the others.
The first is the ROI vacuum. PwC’s 2026 Global CEO Survey found 56% of CEOs report AI has delivered neither revenue growth nor cost reduction. Only 12% report both. Forrester’s 2025 analysis found fewer than one-third of AI decision-makers can tie AI value to P&L changes. When executives cannot connect AI spending to business outcomes, confidence erodes fast. Forrester now predicts enterprises will defer 25% of planned 2026 AI spend into 2027 as the hype fades.
The second layer is the output trust problem. The Content Marketing Institute’s 2025 research found only 4% of B2B marketers report high trust in generative AI’s outputs. Twenty-eight percent report low trust. The people using AI daily don’t trust what it produces. They’re running it through manual review cycles that eat the efficiency gains AI was supposed to deliver.
The third layer is employee resistance. SHL surveyed over 1,000 U.S. workers and found only 27% fully trust their employers to use AI responsibly. Fifty-nine percent believe AI is making bias worse. McKinsey found C-suite executives estimate only 4% of employees use GenAI for 30% or more of daily work. The actual number is 13%, three times higher. Leadership doesn’t know how their teams are using AI, and the teams don’t trust leadership’s AI decisions. That’s a governance vacuum.
The Strategy Gap, Again
This is the same dysfunction I wrote about in “AI Didn’t Replace Your Marketers. It Exposed the Strategic Gap.” The companies failing at AI aren’t failing because the technology is broken. They’re failing because they deployed tools without direction.
The 1% who have actually operationalized AI share a pattern. They did the boring work before the exciting work. They defined what AI should do before they bought the tools to do it. They established governance, data quality standards, and clear use cases tied to business outcomes.
Everyone else treated AI like a magic prompt. Install the tool, see the results. When the results didn’t come, they bought another tool.
IDC and SAS found 46% of organizations admit a “trust gap” between what AI promises and what it delivers. Only 40% have deployed safeguards such as explainability, audit trails, or ethical frameworks. Only 25% have fully implemented governance programs despite 78% using AI.
Over 50% of enterprise IT leaders cite the lack of explainability as a critical barrier to scaling AI projects, according to IBM’s Global AI Adoption Index.
The trust gap is a governance gap. And the governance gap is a strategy gap.
The Customer-Facing Cost
The internal trust collapse creates an external one. When companies deploy AI-generated content without governance, customers notice.
Adobe’s Digital Trends 2026 report found a third of customers would disengage upon discovering content is AI-generated. Thirty-seven percent say the same if they learn they’re interacting with AI when expecting a person. Edelman’s 2025 Trust Barometer found three times as many Americans reject the growing use of AI as embrace it.
For B2B companies, this matters more than most realize. LinkedIn’s 2025 B2B Marketing Benchmark found 94% of B2B marketers agree trust is the key to success. They’re deploying the one technology their customers trust least to build the one thing they say matters most.
The efficiency gains from AI-generated content look compelling on a dashboard. The trust erosion shows up 18 months later when pipeline quality declines and nobody can explain why.
What This Actually Requires
The fix is unglamorous. It looks a lot like what should have happened before any AI tool was purchased.
Start with a use-case audit. Map every AI deployment to a specific business outcome. If a tool can’t be connected to revenue, retention, or operational efficiency, it’s decoration. The 60% project abandonment rate Gartner predicts isn’t inevitable. It’s the result of deploying AI for its own sake.
Establish governance before you scale. Lucidworks’ 2026 research found 83% of AI leaders report major concerns about generative AI — an eightfold increase in two years. They aren’t worried about the technology. They’re worried about the absence of guardrails. Governance isn’t bureaucracy. It’s the operating manual that prevents your AI investments from becoming expensive liabilities.
Define the human layer. AI handles execution at scale. It does not handle judgment, strategy, or trust-building. The companies in that 1% understand this distinction. They use AI for research, drafting, data processing, and pattern recognition. They keep humans in charge of decisions, relationships, and quality control.
Accept that AI is a tool, not a strategy. The 92% increasing spend are treating AI as a destination. The 1% who’ve operationalized it treat it as infrastructure. Infrastructure serves a purpose defined by something else. That something else is strategy. Without it, AI is just expensive electricity.
The Uncomfortable Parallel
Every technology cycle follows the same arc. Massive hype. Massive investment. Disappointing results. Blame the technology. Then, years later, the companies that did the boring foundational work quietly pull ahead.
Cloud computing followed this arc. Digital transformation followed it. Martech followed it. AI is following it now.
The 56% of CEOs reporting zero AI returns aren’t victims of bad technology. They’re experiencing the predictable consequence of skipping the foundation. Truth before tools. Strategy before scale. Governance before deployment.
The 1% already knew this. The other 91% are about to learn it the expensive way.
