I was at a conference last month when a woman pulled me aside during the coffee break. She’d just posted what she called her “best LinkedIn article ever” with the help of AI, and she wanted to show me.
I read it. It was…fine. Grammatically perfect. Professionally structured. And so generically AI-generated that I could have written the same thing by typing “write me a thought leadership post about innovation” and hitting enter.
It felt a bit like she’d spent two minutes on the prompt and zero minutes making it her own. There we so many em dashes plus a bunch of short. choppy. sentences. for. no. good. reason.
This same woman then told me, completely unprompted, that she’s worried AI is going to replace her. The irony is almost too much. She’s worried about being replaced…while actively replacing herself with copy-paste content that sounds like everyone else.
We have more powerful tools than ever, and far too many people are using them to produce more mediocre work, faster.
The problem, in my experience, isn’t the technology. The problem is that we’re skipping over the part where humans still need to bring something to the table. A lot of somethings, in fact.
We're skipping over the part where humans still need to bring something to the table. A lot of somethings, in fact.
Everyone's Buying the Ferrari
According to McKinsey’s latest State of AI report, 85% of workers have no value-driving AI use cases. Eighty-five percent! (I’ll wait while you let that sink in.)
They’re using AI as a glorified Google search because they genuinely don’t know where else it fits. Meanwhile, only 6% of companies qualify as “AI high performers” seeing meaningful business impact, even though 92% plan to increase their AI investments over the next three year.
In other words, organizations are absolutely getting value from AI…they just can’t always see where or how it’s actually showing up. That disconnect isn’t a technology problem, it’s a skills problem.
It’s like buying a Ferrari and then only using it to drive to the mailbox. The car isn’t the issue.
What Leaders Actually Want
When researchers asked business leaders what “AI skills” they’d pay a premium for, the answers were surprising. According to LinkedIn’s 2024 Workplace Learning Report, 76% said they’d pay up to 10% more for candidates with strong AI skills. But when pushed to define what that actually means, they didn’t say prompt engineering or tool expertise.
What they said was adaptability and continuous learning plus some critical thinking and problem-solving. In other words, people who can think, decide, and act. How novel!
It turns out that the people getting hired and promoted aren’t the ones who know the most keyboard shortcuts, they’re the ones who can think clearly, communicate precisely, and exercise judgment when the AI spits out something that looks right but isn’t. (I knew my Master’s degree in communication would come in handy!)
So how do you actually build these capabilities into how you work?
The 20-60-20 Reality
I talk a lot about what I call the 20-60-20 model of working with AI. (Okay, maybe more than “a lot.” My husband has started timing how long I can go without mentioning it. Current record: 47 minutes.)
- The first 20% is you: setting direction, providing context, knowing what you actually need.
- The middle 60% is where AI does the heavy lifting.
- The final 20% is you again: reviewing, refining, adding judgment, making it actually good.
Most people want to skip straight to the 60%. They want to type a super short, oversimplified direction (aka prompt) like Marie Antoinette shouting, “Let them eat cake!” and get a finished product. But that’s not how it works. The quality of the 60% depends entirely on what you brought to the first 20%. And the final output depends on whether you did the work in that last 20% or just hit “copy” and moved on.
Every skill I’m about to describe shows up in that first and last 20%. They’re the parts AI can’t do for you. And they’re the parts that determine whether you get mediocre results or exceptional ones.
The future doesn't belong to the people with the fastest AI; it belongs to the people who are AI-empowered. Those who know that the most important part of the conversation happens before they start typing and after the AI stops.
The Skills That Actually Matter
I keep coming back to five capabilities that separate people who thrive with AI from people who just…use it.
1. Critical Thinking
AI will confidently hand you ten options. Your job is to know which three are worth considering and which seven are nonsense dressed up in professional language. Without that filter, you’re just a human rubber stamp.
In practice, this shows up every time you review AI output.
- Did you actually read it, or did you skim it because it looked professional?
- Did you catch that the statistic it cited doesn’t exist?
- Did you notice that the argument sounds persuasive but falls apart if you think about it for more than thirty seconds?
Most people don’t catch these things because they’ve already decided the AI is smarter than they are. It isn’t. It’s confident. That’s not the same thing. (My young adult kids are also very confident. Doesn’t mean I’d trust them to run the business.)
To build this skill, start treating AI output the way you’d treat a first draft from a new employee who’s eager to impress but doesn’t know your business yet. Question the claims. Verify the facts. Ask yourself whether the logic actually holds together or just sounds like it does.
The more you practice skepticism, the faster you’ll spot the gaps.
2. Communication
The quality of what you get out of AI is directly tied to the quality of what you put in.
Vague instructions produce vague output. The people getting exceptional results are the ones who can articulate exactly what they want, provide the right context and explain why it matters.
This is something I’ve thought about for a long time, probably because of where I came from. I have two degrees in communication. A BS in Speech Communications with a minor in Journalism, and an MA in Communications. I’ve spent decades working in tech, and for a long time, I wondered if those degrees were actually useful in this industry. Turns out they were the perfect preparation for exactly this moment. Because what communicators do, what we’ve always done, is serve as the translation layer and the filter between complex systems and the humans who need to use them. That’s exactly what working with AI requires.
Day to day, this looks like the difference between typing “write me a marketing email” and actually thinking through what you need first.
- Who is this for?
- What do I want them to feel?
- What action should they take?
- What does my brand sound like?
- What constraints am I working with?
People who communicate well don’t just write better prompts; they think more clearly before they start typing, which means they waste less time fixing bad output on the back end.
If you want to get better at this, practice articulating what you want before you open the AI tool. Write a brief for yourself first, even if it’s just three sentences. The discipline of clarifying your own thinking will improve everything downstream, including the conversations you have with actual humans.
3. Creativity
Research from Wharton found that only 6% of AI-generated ideas were truly unique, compared to 100% in human-only groups. AI is brilliant at variations on a theme. It’s terrible at the original theme. If you’re not bringing your own ideas to the collaboration, you’re just remixing what already exists.
This shows up when people use AI to brainstorm and then wonder why all the ideas feel generic. The AI is pulling from patterns it’s seen before. It can’t make the unexpected connection between your weird hobby and your client’s business problem. It doesn’t know about that conversation you overheard at the coffee shop that sparked something. Those leaps are yours to make.
To strengthen your creativity, try brainstorming before you prompt. Spend ten minutes with your own brain first. Write down your half-baked ideas, your weird angles, your “this probably won’t work but what if” thoughts. Then bring those to the AI and ask it to build on them. You’ll get far more interesting output than if you’d asked the AI to start from scratch. Yes, this requires you to actually think before you type. I know. We were ALL hoping we could give that up, but alas, no.
4. Judgment
AI doesn’t know your company culture. It doesn’t know your client’s sensitivities or the political dynamics of your team. It doesn’t know that your CEO hates exclamation points or that your biggest customer is going through a merger and needs a softer touch right now. It doesn’t know when something can be technically right but absolutely the wrong thing to say.
This plays out constantly in small ways. The AI suggests a joke that would land badly with your audience. It recommends an approach that’s efficient but would alienate your team. It produces content that’s fine for a general audience but completely wrong for the specific humans who will actually read it. Judgment is the skill that catches all of this before it goes out the door.
Building better judgment requires staying close to your context.
- Know your audience deeply.
- Understand the unwritten rules of your organization.
- Pay attention to what works and what doesn’t in your specific environment.
AI can give you general best practices. Only you can know what actually works here.
5. Grit
Great AI output rarely comes from the first attempt. It comes from the third or fourth round, after you’ve pushed back, refined the direction and refused to accept “close enough.” Quite simply, most people stop too early (let’s say they do the last 5 instead of the last 20).
You can see this in the quality gap between casual AI users and people who’ve learned to collaborate with it effectively. The casual user types a prompt, gets a response and either uses it as-is or gives up because it wasn’t what they wanted. The skilled user treats that first response as a starting point. They give feedback. They ask for revisions. They try a different angle. They keep going until it’s actually good, not just good enough.
If you want to build this muscle, make a personal rule: never accept the first output.
Even if it looks fine, push for one more round. Ask the AI what’s weak about its own response and how it could be stronger. You’ll be surprised how often the second or third version is dramatically better than what you would have settled for. (And if you’re thinking “but that sounds like work,” well…yes. That’s the reality of being BETTER than a bot.)
The Real Investment
Chess grandmaster Garry Kasparov lost to IBM’s Deep Blue in 1997. But instead of walking away bitter, he spent years studying what happens when humans and machines work together. What he found surprised people.
In freestyle chess tournaments where humans could partner with computers, the winners weren’t the grandmasters with the most powerful machines.
As Garry put it: “Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.”
Your process matters. Your skills matter. The AI is just the amplifier.
So if you’re looking for the best AI investment you can make this year, it might not be another subscription (do NOT tell my husband this!). It’s developing the thinking, communication, and judgment that make every tool you touch more effective.
Pick one of those five skills. Whichever one you’re weakest at. That’s where your AI output is suffering. And that’s where the work starts.
If you want help building these capabilities across your team, that’s what my work is all about. Let’s talk.
About Julie: A Hall of Fame AI keynote speaker, tech founder, and innovation strategist, Julie works with associations, real estate professionals, and corporate sales teams to help them lead smarter, sell more, serve better, and save time with AI. She delivers highly actionable and engaging keynotes on becoming AI-empowered, leading in an AI-driven world, and transforming work and customer relationships.
FAQ: The Human Skills AI Depends On
AI is powerful, but it amplifies whatever you bring to it. If you bring clear thinking, good communication, and sound judgment, you'll get great results. If you bring vague prompts and zero oversight, you'll get mediocre output that sounds like everyone else. The tools aren't the differentiator anymore; the human skills are.
Critical thinking (knowing what's worth keeping and what's nonsense), communication (articulating what you actually need), creativity (bringing original ideas AI can build on), judgment (knowing what's right for your specific context), and grit (pushing past "good enough" to get genuinely great results).
AI is trained to produce fluent, professional-sounding text. It's optimized for coherence, not accuracy. So it can generate something that reads beautifully while being factually wrong or completely generic. Confidence and correctness are two different things.
Skipping the human bookends. They don't invest in the first 20% (clear direction) or the last 20% (thoughtful review). They just type a vague prompt, copy the output, and wonder why it doesn't sound like them or serve their actual goals.
AI will state incorrect information with the same assured tone as correct information. It doesn't flag uncertainty the way a human expert would. So you have to bring your own skepticism and verification, because the AI certainly won't tell you when it's wrong.
Surprisingly, they don't mean prompt engineering. According to research, leaders value adaptability, critical thinking, and problem-solving. They want people who can work effectively with AI as a tool, not people who've memorized keyboard shortcuts.
You invest the first 20% setting direction (defining goals, audience, and context). AI handles the middle 60% (drafting, generating options, doing the heavy lifting). You return for the final 20% to edit, validate, personalize, and apply judgment. This ensures AI amplifies your thinking rather than replacing it.
Bring your own ideas to the conversation. Brainstorm before you prompt. Provide specific context about your audience, your brand, and your goals. And always review output through the lens of "does this sound like me, or does it sound like everyone else?"
Never accept the first response. Even if it looks fine, ask for a revision. Ask the AI to identify weaknesses in its own output and suggest improvements. This simple habit consistently produces dramatically better results.
Developing your own thinking, communication, and judgment skills. These make every AI tool you use more effective. Another subscription won't help if you don't have the human capabilities to direct and evaluate what AI produces.
