There is a number that should stop every business owner cold. According to McKinsey’s 2025 State of AI survey, 88% of organizations now use AI in at least one function. According to BCG, 74% of companies generate no tangible value from it, despite $252.3 billion in collective spending in 2024.
Read those two figures together and you have the defining business challenge of this era. Almost everyone is using AI. Almost no one is getting paid back for it.
The gap is not the technology
It is tempting to blame the models. That is not where the failure lives. RAND Corporation studied the question directly and found that more than 80% of AI projects fail, at roughly twice the rate of other IT projects. The root causes were not about model quality. They were about design:
- Misunderstood problem definition. Teams build before they know what problem they are solving.
- Inadequate data. Systems are built without the real operational context they need.
- Technology-first mentality. Tool selection leads, strategy follows.
- Insufficient infrastructure. Experiments stay disconnected from the business.
- Problem too difficult. The level of autonomy does not match the use case.
Every one of these is a design problem, not a tooling problem. That is the good news. Design problems are solvable.
What the winners do differently
The most consistent finding across McKinsey, BCG, MIT, and RAND is almost boring: workflow redesign, not tool selection, is the primary driver of AI value. Organizations that redesign workflows before selecting tools are 2x more likely to achieve significant financial returns.
The resource split tells the same story. The proven pattern for successful AI is roughly 10% algorithms, 20% technology and data, and 70% people and processes. Most companies invert that. They spend on tools and hope the processes catch up. They never do.
When it works, the payoff is real. Early generative AI adopters return $3.70 for every dollar invested. Top performers return $10.30. The difference between those two numbers is not better software. It is better design.
Where small and mid-sized businesses actually sit
If you run a smaller business and feel stuck, you are not behind. You are in the largest group. The PayPal and Reimagine Main Street survey found that 51% of small businesses are “Explorers”: interested, experimenting, but not yet committed. Their top barriers were time and resources (37%) and not seeing a clear use case or ROI (34%).
Those are not skeptics. They are capable businesses waiting for a structured path. The thing standing between them and value is rarely intelligence or effort. It is a guide and a plan.
The first move
You do not close a 74% value gap with another tool subscription. You close it by defining the right problems first, redesigning the workflow around them, and matching the right level of human oversight or AI autonomy to each task.
That is exactly what the AI Audit is for. It is a set of short exercises, offered as a courtesy, covering the people and workflows carrying the work, the tools and subscriptions you already have, the business foundation and goals, and the data conditions around it, starting with AI Land, an AI skills assessment for you and your team. It is prioritized direction you can act on, with us or on your own. It is not a finished implementation plan.
Human judgment where it matters. AI agency where it works. That is how AI gets actualized.
