And How to Do AI Better.

Artificial Intelligence has the potential to transform how organizations work, innovate, and serve customers. Yet many leaders find themselves underwhelmed by the results of their AI initiatives.

In fact, new data from MIT show that 95% of AI pilots are delivering zero measurable return!

Why? Because the problem usually isn’t the technology. It’s how you’re using it.

Here are the five biggest mistakes:

1. Your data isn’t ready.

The old adage applies: garbage in, garbage out. If your data is messy, inconsistent, or incomplete, AI can only amplify those flaws.

• Is your data structured in a usable way?
• Are you using a data lake or other scalable storage strategy?
• Did you flood your LLM with everything you had, without cleaning, tagging, or segmenting it by relevance?

Raw data without curation leads to weak insights. AI’s output is only as strong as the input it’s given.

2. You implemented without a strategy.

Too often, companies pursue AI simply because everyone else is doing it. That approach almost guarantees disappointment. Successful AI adoption requires:

• A strategy aligned with business outcomes.
• Clear metrics for value (beyond “hours saved”). Think faster time to market, improved customer experience, or stronger employee engagement.
• A focus on use cases, not hype. AI should solve real business problems, not be a vanity project.

If you deployed AI without knowing the problem you wanted it to solve, you’re wasting your investment.

3. You didn’t prepare your people.

AI changes how people work. It requires different ways of thinking, new workflows, and new communication patterns. If your workforce hasn’t been:

• Trained in how to interact with AI,
• Guided on how their roles will evolve, and
• Supported in building new skills

…then you’re in for employee resistance and underperformance.

4. Your prompts (and people’s AI literacy) are weak.

Large language models don’t just “know” what you want. They need to be guided. That’s where prompt engineering comes in. It’s an art as much as a science. Weak, vague, or overly broad prompts generate disappointing results. Strong prompts, by contrast, can unlock nuanced, actionable insights.

But prompts don’t live in a vacuum. Without AI literacy, your workforce won’t understand how AI works, its limits, or its ethical considerations. Teaching people how to think critically, ask better questions, and apply responsible practices is the foundation of writing better prompts.

5. You underestimated change management.

AI adoption isn’t just about technology; it’s a change initiative. Like any organizational change, it requires leadership sponsorship, communication, and reinforcement.

And here’s the critical piece: storytelling. You must tell AI success stories in a way that makes employees feel part of the journey.