The AI Expectations Bubble and What It Means for You

The bubble no one is talking about

The AI industry just crossed a strange milestone. McKinsey’s 2025 State of AI report found that 88% of organizations now use AI in at least one business function. That number sounds like a success story. It is not.

Only 6% of those organizations report capturing meaningful business value from their AI investments. Two-thirds remain stuck in what researchers call pilot purgatory — running experiments that never graduate to production. Only 39% say AI has affected their company’s earnings at all, and most of those report less than a 5% impact.

As Kiplinger recently reported, the real AI bubble forming right now is not a technology bubble. It is an expectations bubble — where boardroom claims about AI transformation are wildly detached from what companies are actually delivering.

What the expectations bubble looks like

The pattern is consistent across industries. A company announces an AI initiative. Leadership talks about transformation. Pilots launch. Then nothing scales.

McKinsey found that high-performing companies are 2.8 times more likely to have fundamentally redesigned their workflows around AI — not just bolted AI onto existing processes. Only 21% of all companies have done this. The rest are running AI tools on top of workflows that were never designed for them.

The blockers are not technical. They are organizational:

• Data quality and architecture — AI cannot produce good results from messy, siloed data
• Workflow rigidity — processes built decades ago resist the kind of changes AI requires
• Operating model inertia — teams keep doing what they have always done, now with an AI tool sitting unused
• Measurement gaps — no one tracks whether the AI pilot actually improved anything

Meanwhile, the capital pouring into AI infrastructure is staggering. Amazon plans to spend over $200 billion in capex for 2026. Alphabet has budgeted $175-185 billion. Meta is investing $115-135 billion. These companies are betting that enterprise AI demand will justify the spend. If most pilots never scale, that demand may not materialize the way investors expect.

Why small businesses are actually ahead

Here is the surprising part. While large enterprises struggle with pilot purgatory, small businesses are getting real results.

The SBE Council’s 2026 Technology Use Survey, released March 11, surveyed 517 small businesses with 2-99 employees. The findings tell a different story than McKinsey’s enterprise data:

• 82% of small business employers have adopted at least one AI tool
• The typical small business uses five different AI tools across operations
• 66% report revenue increases linked to AI — with 22% seeing gains above 10%
• Owners save a median 5 hours per week of personal time
• Businesses save a median 11.5 employee hours per week

Why the gap? Small businesses skip pilot purgatory because they cannot afford it. When a restaurant owner adopts an AI tool for review management or a contractor sets up automated dispatch, they need it to work immediately. There is no six-month pilot phase. No committee to approve scaling. No organizational inertia to overcome.

Small businesses adopt AI the way they adopt any tool — they pick something that solves a specific problem, try it, and either keep it or move on. That practical, problem-first approach is exactly what McKinsey’s high performers do, just without the bureaucracy.

How small businesses can avoid the hype trap

The expectations bubble is real, and small businesses are not immune to it. Adopting five AI tools does not mean all five are delivering value. Here is how to stay on the right side of the bubble.

Start with the problem, not the technology

The companies stuck in pilot purgatory almost always started with the technology. They asked “How can we use AI?” instead of “What is our most expensive problem?”

Flip the question. If missed calls are costing you leads, an AI answering service solves that. If scheduling chaos is burning staff hours, AI dispatch addresses it directly. If you cannot point to a specific problem an AI tool solves, you do not need that tool yet.

Measure from day one

McKinsey’s high performers tie AI to business KPIs. For a small business, that means tracking before-and-after numbers on the metrics that matter:

• Calls answered vs. missed
• Average response time to leads
• Hours spent on scheduling per week
• Review response rate
• Revenue per employee

If you started using AI tools recently, go back and compare this month’s numbers to last month. If you cannot see a difference, the tool is not working — or it is solving the wrong problem.

Cut what is not working

Only 9% of small business owners identify as pessimistic about AI, according to the SBE Council survey. That optimism is healthy, but it can also lead to keeping subscriptions that are not delivering. Audit your AI tools quarterly. If a tool has not demonstrably saved time or increased revenue in 90 days, cancel it and redirect the budget.

This is the opposite of the enterprise approach, where failed pilots linger for years because no one wants to admit the initiative did not work. Small businesses have a structural advantage here — fewer stakeholders means faster decisions.

Do not chase the next shiny model

Every week brings a new AI model release or feature announcement. Grok 4.20 launched this week. Apple is overhauling Siri. Microsoft just announced Copilot Cowork. These developments matter for the industry, but they rarely require you to change what is already working.

If your current tools are delivering measurable results, keep using them. Upgrade when a new capability solves a problem you actually have — not because a press release made it sound exciting. We have seen this pattern before: 40% of AI agent projects fail not because the technology is bad, but because the use case was not grounded in a real business need.

Practical AI adoption that actually sticks

The difference between the 6% of enterprises capturing AI value and the 66% of small businesses reporting revenue gains comes down to one thing: discipline. Not technical sophistication. Not budget. Discipline.

Discipline means:

1. One problem at a time. Pick your most painful operational bottleneck and solve it before adding more tools.
2. Clean inputs. AI tools are only as good as the data you feed them. If your customer records are a mess, fix that before deploying an AI CRM.
3. Human oversight. Every AI output needs a human checkpoint, especially early on. The businesses that get burned are the ones that set AI to autopilot and walk away.
4. Regular evaluation. Check results monthly. Adjust or cancel tools that are not performing.

The expectations bubble will burst for companies that confused adoption with transformation. It does not have to burst for you.

The bottom line

The AI expectations bubble is real — but it is mostly an enterprise problem. Small businesses that stay focused on specific problems, measure results honestly, and cut what is not working are already outperforming companies that spend 1,000 times more on AI initiatives.

The tool matters less than the approach. If you are not sure where to start, we can help you identify the right AI tools for your specific business — no pilots required.