I noticed a pattern in 2025, most of my conversations with executives at growing companies had a similar theme. There is a need and urgency to do something with AI. The board is asking questions. Competitors are making announcements. Vendors are flooding inbox with promises.
But in reality nothing was being done with AI. Or if there was anything it was a small project, but in pockets with confusion and not much clarity on how to scale.
AI feels like hype, but most growing companies feel stuck.
After two decades leading transformations at Fortune 100 organizations, and working with growing medium size organizations throughout 2025, I’ve seen the same patterns repeat.
The problem isn’t the technology. It’s how we’re approaching it.
Here are four things we probably got wrong about AI and what we can learn from 2025.
1. Getting paralyzed by vendor noise
I’ve sat across from more than a few executives this year who admitted they haven’t started using AI in their business yet. Not because they don’t see the value. Because they’re overwhelmed.
Every platform now has AI features. Every vendor claims their solution will transform operations. The inbox is full of demos, whitepapers, and case studies.
But there’s not much clarity. No proof if these features wil work.
So, nothing happened. These growing mid size organizations are waiting for dust to settle.
However, the companies that made progress in 2025 didn’t wait for perfect information. They started with a real business problem and worked backward. What’s broken? What’s slow? What’s costing us money we shouldn’t be spending?
Then they looked for tools that fit the problem. Not the other way around.
AI isn’t a strategy. It’s a tool that supports one.
And if the right tool doesn’t exist yet? It’s okay to wait. Just don’t wait forever.
Start an experiment and learn from it. That brings me to my next observation.
2. Confusing a prototype with a production solution
I watched this happen a lot, not just in 2025 but since 2024.
A team launches an AI experiment. A quick prototype is built, something vibe-coded, scrappy, just enough to test the hypothesis. It works. Leadership is excited. The business case looks real.
Then try to move it to production.
That’s where it stalls.
Because a prototype that proves a concept is not the same as a production-grade solution. Production means reliability. Security. Scale. Integration with existing systems. Support and maintenance.
You can’t vibe-code your way to production. You need proven frameworks. Clear guardrails. Proper project management. The discipline to build something that won’t break when real users touch it.
The gap between “this works in a demo” and “this works in our business” is bigger than most teams expect.
And that gap doesn’t close with more AI. It closes with better execution.
So vibe-code your heart’s desire. But when it’s time to productionize, you need proven frameworks and guardrails. AI can’t solve that. Which leads to another gap I saw everywhere.
3. Experimenting individually but never deploying at enterprise scale
AI is now embedded in almost every enterprise platform. Your CRM has it. Your ERP has it. Your productivity tools have it.
Most people tried it at least once. Drafted an email. Summarized a document. Generated a report.
But what about about enterprise-level use cases, the actual business problems being solved with AI at scale… the answer is almost always no luck.
Why? Because individual experimentation doesn’t translate to enterprise deployment.
One person using AI to write emails is fine. But it doesn’t move the business forward. It doesn’t reduce costs. It doesn’t improve customer outcomes. It’s just… happening.
What enterprise deployment actually looks like: your sales team uses AI in the CRM to automatically qualify leads based on your specific criteria, route them to the right rep, and generate personalized follow-up sequences. That’s a workflow. That’s repeatable. That’s measurable.
The gap is leadership. Someone needs to identify the business use case, define the problem, set the guardrails, and deploy AI in a way that creates measurable value.
That’s not happening in most companies yet, at least in mid size growing companies who are strecthed for resources. And, that’s my next observation from 2025.
4. Skipping the hard work of understanding the process first
I worked with a services company this year on process optimization. We didn’t talk about AI in the kickoff. We talked about their workflows.
What are you doing today? Why are you doing it that way? Where does it break down? What takes too long?
We mapped it. We questioned it. We streamlined it.
Only then did we look at where AI-powered systems could automate parts of the process.
The project worked. Not because the AI was sophisticated. Because we understood the work first.
I’ve seen the opposite happen too many times. Teams buy an AI tool, point it at a messy process, and expect it to figure things out. It doesn’t. It automates the mess. Sometimes it makes it worse.
You can’t optimize what you don’t understand. And AI won’t do the understanding for you.
The companies that made progress didn’t treat AI as a transformation initiative. They treated it as a tool to improve specific parts of their business.
They started with the work, not the technology. They understood their processes before automating them. They set clear business outcomes before experimenting. They built proper frameworks to move prototypes to production. They learned by doing, not by reading case studies.
The ones that struggled were waiting. For the right platform. For clearer ROI. For vendors to make it easier.
AI won’t get easier to deploy. But the work of understanding your business and defining clear problems will always pay off.
Heading into 2026
The noise will get louder. More vendors. More features. More pressure to move fast.
Your advantage isn’t speed. It’s clarity.
Understand the work. Define what success looks like. Build it properly. Then bring in AI to support it.
AI is a powerful tool. But only if you know what you’re trying to build.
Most companies I see struggling with AI have the same foundational gaps. If you want to talk through where you’re stuck, feel free to reach out. I’d be glad to help.
I’m curious how 2025 played out for you. What did you try with AI? What worked? What didn’t?
If any of this resonates, I’d love to hear your perspective. You can reply to this newsletter or connect with me on LinkedIn and continue the conversation.
