Core AI skills that actually matter
When leaders ask me how to approach AI training, they’re usually thinking about technical things like prompt engineering, data security protocols, and when to use which tools. And yes, they are important. But if that’s where the AI training ends, it’s setting up teams to use Ferraris like golf carts. The real question is not, “How do I write better prompts?” The question is, “How can we fundamentally change the way our employees work?”
For 25 years, I’ve been helping organizations build learning ecosystems that drive real change, not just checking compliance boxes. And what I’m seeing now is a huge gap between what companies think they need to teach about AI and what actually makes teams more effective.
What about your technical skills? These are valuable, but you will learn them as your team uses the AI tools. What AI does not automatically develop, and what will determine whether it becomes a true multiplier or just an underutilized tool, are four human-centered AI skills that have nothing to do with technology and everything to do with how we think.
Overview of the four core AI skills
Delegation: Moving from transactional search to results-based management. Curiosity: Use repetition to turn tools into thought partners. Contextual intelligence: Make tacit knowledge explicit for better results. Discernment: The expertise needed to critically evaluate the output of AI.
Delegation: The Art of AI Management, or Beyond Improved Search Engines
Most people use AI like a very fast research assistant. They ask questions, get answers, and sometimes copy and paste the results. This approach is like hiring a talented analyst and only asking them to submit documents.
Effective delegation with AI means taking over substantive work with clear outcomes in mind. Instead of “Give me three bullet points about X,” you need to shift your mindset to “Here’s the problem I’m trying to solve. Here’s the context I need. Please help me come up with a solution.”
Why delegation is the most difficult AI skill
Consider what makes delegation difficult for real humans. To do this successfully you need to do the following:
You need to clarify your thinking by knowing what outcome you actually want. Understand what the person (or tool) you are delegating can and cannot do well. Provide enough context without micromanaging every detail. Critically review and iterate on your work.
Both of these challenges manifest themselves in the use of AI.
People on teams who have never learned how to delegate effectively—those who keep everything close together, those who throw work out without context and get frustrated with the results—will struggle with AI in exactly the same way. Team leaders who have developed strong delegation skills will quickly recognize this pattern.
When teams delegate well to AI, they do more than just produce faster output. Reduce cognitive workload and allow employees to focus on higher-level thinking. We are using AI not just as a production tool, but as a thinking partner.
Curiosity: What makes AI (and everything else) work
When people lack curiosity, the following things happen: Accept the first output provided by the AI tool and move on. They use it transactionally to get an answer, complete it, and do the next task.
But the true power of AI emerges through iteration and exploration. The first response is rarely the best response. The obvious answer is often not the most useful answer. And the question you start with may not actually be the question you should be answering.
From transactions to exploration
Curious people ask, “What if?” “Why not?” They push back against algorithms. They look for different approaches and explore points of contact that can lead in interesting directions. They tolerate some disruption in the process because they understand that this is where insight is born.
Think about how you work with really smart colleagues. You don’t just ask a question, get an answer and walk away. You’ll be able to have a conversation. You will follow up. You will end up imposing your ideas on each other. Together you will discover things that neither of you could have reached alone.
Curiosity turns tools into thinking partners. Organizations and teams that create space for exploration and foster curiosity rather than simply checking off tasks will gain exponentially more value from AI than those that treat AI as a faster typewriter.
Contextual intelligence: making tacit knowledge explicit
AI doesn’t know what you know. You don’t understand the culture of your organization, the unspoken needs of your customers, or the political landscape you’re navigating. I can’t read between the lines.
This is actually one of the most valuable aspects of working with AI. This creates a need to explicitly express what is usually implicit. But it only works if teams have situational intelligence, the ability to recognize and articulate important context information.
Contextual intelligence is key to a process called context engineering, which acts as a middle layer between external LLMs such as ChatGPT, Gemini, Claude, etc. and employees. This middle layer provides the necessary contextual intelligence to make LLM output relevant while ensuring the security of both your organization’s data and intellectual property.
“New Team Member” Test
When people with strong contextual intelligence delegate tasks to AI, they treat it like a new employee who needs onboarding and ongoing coaching. they ask:
What is the important background information here? What assumptions am I making that I need to state? What are the constraints that are not obvious?
Without this skill, employees will give the AI the bare minimum (often just a surface-level task description) and wonder why its results feel so generic.
Contextual intelligence is closely related to systems thinking, the ability to see how the parts are connected. These are the people who can explain not just what needs to happen, but why it’s important to the business.
Insight: A key people skill
Discernment, or knowing what is actually good, may be the most important skill in AI, and the one that worries me the most.
AI produces output that sounds confident, whether that output is insightful, accurate, or relevant. Teams need the insight to tell the difference.
Identification is not about finding obvious errors in AI output. It’s relatively easy. It is to evaluate the following:
Soundness: Is the logic flawed?Depth: Is the analysis deep enough?Relevance: Do the recommendations actually say what matters?Framework: Is the problem defined correctly?
Why subject matter expertise still matters
You can’t judge quality in a field you don’t have knowledge of. This is one reason I resist talk of AI replacing human experts. Evaluating the output produced by AI requires more expertise, not less.
Consider the L&D space. If your AI-generated training goals sound good but don’t match your actual business outcomes, you can quickly find out. Someone without an L&D background might think it looks and sounds professional…but an L&D professional would know better.
Without insight, your team will drown in plausible-sounding content that doesn’t actually move the job forward. Quantity increases, but quality stagnates or declines.
Long-term AI skills
If you’re building your team’s training on AI, be sure to cover the technical basics. We’ll teach you the basics right away. Address data security. Find out which tools serve which purposes. But don’t stop there.
Incorporate learning experiences that develop delegation skills through simulations that practice scoping work, providing context, and evaluating results. Make room for curiosity by encouraging experimentation and iteration rather than rushing to a finished product. Develop situational intelligence through exercises that formalize tacit knowledge. Strengthen discernment by having people evaluate AI output together and discuss what is really good and what is good enough.
These are not skills that can be learned in a two-hour workshop. It requires practice, reflection, and continuous development. It’s more like leadership development than technical training.
But importantly, these are also skills that improve employees’ work, regardless of AI. People who delegate well, are curious, provide the right context, and use insight are more effective in all aspects of their work.
AI will only make these skill gaps more apparent and increase costs.
Organizations that invest in developing these capabilities along with AI technical skills will find that AI can be a transformative tool. Their teams produce better thinking faster. They will solve problems more creatively. They will make better decisions.
Organizations that focus solely on technical skills will wonder why they aren’t seeing the productivity gains they expected, even though everyone knows how to write prompts.
Bottom line: AI tools are rapidly evolving. The technical skills you teach today may not be useful in six months. However, delegation, curiosity, situational intelligence, and insight are fundamental capabilities that apply to all AI tools your team uses, regardless of how the tools evolve. And these are features worth investing in and focusing on.
Want to learn more about innovative AI skills to future-proof your workforce and organization? Talk to one of our experts to find out how we can help you realize your vision.
sweet rush
Our job is to help you achieve your goals and succeed. From analytics to custom development (including e-learning, mobile, gamification, and ILT) to evaluation, we’re here to help.
