Anthropic released their AI Fluency Index last month, surveying 1,000 US adults about their AI usage and proficiency. One finding stood out. It wasn't about age, education, or job title. It was about iteration.

The people getting the most value from AI aren't the ones asking questions once and accepting the first answer. They're the people asking follow-ups. Refining. Pushing back. Using the output as a starting point, not a destination.

I've been teaching AI leadership for two years. That finding didn't surprise me. It confirmed something I've seen in every cohort.

The Finding That Didn't Surprise Me

The index ranked people by fluency level, from non-users to power users. The differences were interesting. Power users weren't necessarily technical. They weren't all in tech jobs. What they shared was a mindset.

The non-users asked AI one question and trusted the result. The casual users asked multiple questions, but didn't pressure-test. The power users were relentless. They asked, they read, they questioned, they asked again. They were having conversations, not requesting outputs.

Every founder I've worked with who's getting outsized value from AI shares this trait. They use iteration as a fundamental tool. They write a prompt, get a response, write another prompt that builds on it. They're not trying to get it perfect on the first try. They're trying to explore the possibility space with the AI as a thinking partner.

The ones who aren't getting value? They ask, they get an answer, they use it or discard it. They're not iterating. They're not pushing back. They're not treating it like a conversation.

The Finding That Should Worry Every Business Leader

Here's the concerning part. The index also showed that most people don't know what they don't know. Most non-users think they know what AI can't do. Most casual users think they know the boundaries of what AI is useful for.

They're wrong. The boundaries of what AI can do are expanding every month. And the people who are winning aren't the ones who defined AI's capabilities once and stopped. They're the ones who keep experimenting. Who keep iterating. Who stay curious.

For business leaders, this is critical. If you're building a company in the AI era, you need people on your team who iterate. Who experiment. Who push on the boundaries. Because the companies winning are the ones discovering new use cases, not copying existing ones.

Most companies are buying the same tools. The ones winning are using those tools in new ways. And they're doing that through iteration.

The 30% Problem

One stat from the index struck me. Around 30% of people surveyed said they use AI regularly. That means 70% either don't use it, or use it so infrequently that they didn't pick "regularly" as their answer.

In five years, that 30% will be the norm. The 70% will be at a structural disadvantage. Not because AI will replace them. Because they'll be slower at their jobs. Because they won't have the thinking tools their colleagues have access to.

The scary part? That 70% probably doesn't feel behind. They're probably comfortable doing things the way they've always done them. By the time they realize they're behind, the gap will be huge.

This is why teaching iteration matters. It's not just about getting better answers from AI. It's about developing a mindset, a way of working, that's fundamentally different from the way previous generations worked.

What This Means If You're Teaching AI Fluency

If you're building an AI fluency program, whether inside a company or for a broader audience, here's what matters: teach iteration first. Not prompting. Not theory. Not what AI is.

Teach people how to ask, listen, refine, ask again. Teach them that the first output isn't the answer. It's the beginning of a conversation.

Teach them to push back. "That's interesting, but I'm thinking about it differently. Here's what I'm trying to accomplish." Watch how the AI adapts. Watch how a seemingly useless response becomes useful because they reframed the question.

In the Leadership in the AI Era program, we structure everything around this. We don't teach people to be prompt engineers. We teach them to be iterative thinkers. To use AI as a tool for thinking, not just a tool for output.

The people who graduate from that program don't just get better answers. They think differently. And that difference compounds over time.

The Takeaway

Anthropic's AI Fluency Index revealed something simple but profound. The secret to getting value from AI isn't intelligence. It isn't technical skill. It's iteration.

The people and organizations winning with AI share a bias toward action, experimentation, and refinement. They don't ask once. They ask, they listen, they refine, they ask again.

If you're using AI and not iterating, you're leaving value on the table. If you're leading people who aren't iterating, you're leaving performance on the table. And if you're training people to use AI, teach them iteration first. Everything else flows from there.