I've been in technology for over 30 years. The last several of those have been spent building AI systems, consulting on AI strategy, and watching organizations make the same mistakes with AI that I watched them make with every other major technology shift since the '90s.
The pattern is always the same. New technology arrives. Vendors promise transformation. Leadership buys in without understanding what they're buying. Six months later, there's an expensive tool that nobody uses and a team that's more confused than when they started.
With AI, the stakes are higher and the pattern is playing out at scale. The numbers tell the story.
The Numbers Don't Lie
of AI projects fail to deliver meaningful valueSource: RAND Corporation, 2024
cause of failure: misunderstanding the problem to solveSource: RAND Corporation, 2024
the failure rate of traditional IT projectsSource: RAND Corporation, 2024
data scientists and engineers interviewed by RAND to identify whyMinimum 5 years experience each
These aren't pessimistic projections. They're documented results from RAND Corporation's peer-reviewed research. The AI industry has a success problem, and it's not because the technology doesn't work.
The Real Problem: Implementation Before Understanding
Here's the pattern I see repeatedly, an organization feels pressure to "do something with AI." Leadership attends a conference, reads about competitors' AI initiatives, or gets cornered by a vendor with impressive demos. Within weeks, they've committed to a platform, hired consultants, and started a pilot project.
Six months later, they have an expensive tool that nobody uses, a confused team, and serious questions about whether AI was ever the right solution.
I've watched this movie before. In the early 2000s it was "we need a website." Then it was "we need to be in the cloud." Now it's "we need AI." The technology changes. The mistake doesn't.
The Five Root Causes (Per RAND)
RAND's research identified five leading root causes. Notice how many of them are people problems, not technology problems.
- Misunderstanding the Problem: Stakeholders miscommunicate or never clearly define what problem needs to be solved. This was the most common cause of failure.
- Inadequate Data: The organization lacks the data needed to train an effective model. Nobody checked before committing budget.
- Technology-First Thinking: Choosing the latest technology instead of focusing on solving real problems for real users. "We need AI" isn't a business requirement.
- Infrastructure Gaps: No adequate infrastructure to manage data and deploy models. The AI works in the lab but can't run in production.
- Wrong Problems for AI: The technology gets applied to problems that are genuinely too difficult for current AI to solve. Sometimes the honest answer is "not yet."
What This Actually Means
Look at that list again. Three of the five root causes (misunderstanding the problem, technology-first thinking, and wrong problems for AI) come down to the same thing, leadership making decisions about technology they don't understand.
The models work. The platforms function. The algorithms do what they're designed to do. Organizations fail because they never understood what they were building or why.
RAND's top recommendation is blunt, "Industry leaders should ensure that technical staff understand the project purpose and domain context." In other words, the people building the AI need to understand the business, and the people running the business need to understand the AI. When that connection breaks, projects fail.
In my experience at Aktion Associates, where I've spent over 25 years working with businesses on technology decisions, the organizations that succeed are always the ones where leadership takes the time to actually understand what they're implementing. Not at a technical level, necessarily. But at a level where they can explain the "why" to their own team without reading from a vendor's slide deck.
What Successful AI Adoption Looks Like
The organizations getting real results from AI share common characteristics. They didn't rush. They didn't chase hype. They built understanding before building systems.
Education First
Leadership developed genuine AI fluency before making technology decisions. They can explain, in their own words, what AI does and doesn't do well. This is the single biggest differentiator I've seen between projects that succeed and projects that don't.
Problem-Driven Selection
They started with specific, measurable business problems and worked backward to technology solutions. Not the other way around. "Our customer response time is 48 hours and needs to be 4" is a business requirement. "We need a chatbot" is not.
Governance by Design
Security, privacy, and compliance were foundational requirements, not accommodations made after deployment. If you're thinking about governance after your data is already flowing through a third-party system, you're already behind.
Internal Ownership
They built internal capabilities and understanding, reducing dependency on external vendors for ongoing operation. The goal isn't to eliminate consultants. It's to make sure your organization can stand on its own when the engagement ends.
Appropriate Scope
They started with focused pilots, demonstrated value, and expanded deliberately. RAND specifically recommends leaders "be prepared to commit each product team to solving a specific problem for at least a year." That's not a sprint. That's patience.
The Bottom Line
AI project failure isn't inevitable. It's the predictable result of a predictable pattern, implementing technology you don't understand to solve problems you haven't clearly defined, without the infrastructure or governance to support it.
The organizations succeeding with AI aren't necessarily the ones with the biggest budgets or the most sophisticated technology. They're the ones who took time to build genuine understanding before writing the first check.
That's what 30 years in technology has taught me. The tools change. The principles don't. Understand the problem. Understand the technology. Build from there.
"AI adoption shouldn't be something that happens to your organization. It should be something your leadership understands, owns, and drives with confidence."
- Daryl Lantz, MindXpansion
How We Help Organizations Get This Right
At MindXpansion, we start where most consultants skip, understanding. We invest time helping your leadership develop real AI fluency. Not jargon. Not buzzwords. Practical understanding that lets decision-makers confidently evaluate options and ask the right questions.
Only then do we move to technology. This approach takes more time upfront. It saves you months of wasted effort and budget later.
Our Approach
- Strategic AI Consulting and Roadmap Development
- AI Fluency Programs (tailored to your industry and challenges)
- Technology Evaluation and Selection (vendor-neutral, always)
- Governance and Risk Framework Development
- Implementation Support and Change Management
Sources
- Ryseff, James, Brandon F. De Bruhl, and Sydne J. Newberry. "The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed: Avoiding the Anti-Patterns of AI." RAND Corporation, 2024. rand.org/pubs/research_reports/RRA2680-1.html