Getting Better Results from AI Starts with Better Human Skills
Why do some people get thoughtful, nuanced, genuinely useful results from AI, while others get generic slop?
It is easy to assume the difference comes down to the tool, or to who is more technical. But research suggests something else is going on.
MIT Sloan recently highlighted a study showing that only about half of the performance gains from moving to a better model came from the model itself. The other half came from how users adapted their prompts. In other words, better results were not just about better AI – they were also about better communication.
That rings true in practice.
Many people approach AI with very little clarity and a lot of expectation. Their request is vague, the context is thin, the goal is fuzzy, and then disappointment follows when the response comes back broad, flat, or off target. The assumption is that AI should somehow know what was meant, even when the ask was rushed or unclear.
A much more useful frame is to think of AI as a very smart colleague, one who is fast, capable, and ready to help, but completely dependent on the quality of the briefing. It still does not know the audience, the backstory, the tone, the hidden assumptions, or the real definition of success unless those things are made visible. That shift alone changes the interaction.
Instead of saying, “Make this better,” and hoping for magic, the conversation becomes more thoughtful: Here is what this is for. Here is who it is for. Here is what is working. Here is what is off. Here is what I want more of. Here is what I want less of. That level of detailed prompting is often when the quality of the response starts to change.
What is so interesting is that this is not mostly about technical tricks. It is about human skills: Clarity. Context. Judgment. Feedback. The ability to make expectations visible. The patience to refine instead of expecting perfection on the first try. OpenAI’s own guidance reinforces that same pattern: strong output usually follows strong instruction.
There is also some research behind the tone of the interaction. A 2024 cross-lingual study found that prompt politeness can affect model performance, and that rude prompts often performed worse, though extreme politeness was not automatically best either. That is an important nuance. The lesson is not simply “be nice to AI.” It is that the quality of the interaction matters. Clarity is respectful. Specificity is respectful. Helpful context is respectful. Thoughtful correction is respectful.
And maybe that is one of the most important lessons AI is teaching us right now.
For all the talk about technical disruption, AI is quietly reminding us that human skills matter more, not less. The better someone is at communicating, framing, guiding, and refining, the better the output tends to be. AI is not removing the need for those skills. It is exposing how essential they have always been.
This is true in leadership, too. Strong leaders do not assume people can read their minds. They provide direction. They share context. They define what good looks like. They course correct without shaming. They stay engaged in the process. The same habits that make someone effective with people also tend to make them more effective with AI.
Of course, AI is not human, and it is important not to overhumanize it. But there is a long line of research showing that people naturally respond to computers in social ways. Used wisely, that instinct can actually help here, because it nudges us toward clearer, more collaborative communication.
Maybe that is the hidden gift.
When AI gives us slop, sometimes it reveals a limitation in the tool. But sometimes it also reveals a gap in how we asked. And when it gives us something sharp, useful, and unexpectedly strong, it is often because we brought the very skills that matter most in any meaningful collaboration.
Not technical tricks. But human ones.
The people getting remarkable results from AI are often not doing anything mysterious. They are:
communicating clearly,
thinking carefully, and
staying in the conversation.
What have you learned about communication from working with AI? Where have you seen the biggest difference between slop and really strong output?
Ready to Get More Human Value from AI Tools?
The organizations seeing the biggest returns from AI are not necessarily the ones with the most advanced tools. They are the ones helping their people develop the communication, critical thinking, and leadership skills that make those tools effective.
At LAeRRICO & Partners, we help leaders and teams build the human capabilities that drive stronger decision-making, clearer communication, and more effective use of emerging technologies. Because AI performance is often a reflection of human performance.
If you're exploring how to help your team work smarter with AI while strengthening the human skills that matter most, let's start a conversation.
***
Jennifer Budd, ACTC, PCC, is a consultant, coach, leadership advisor, and executive educator who works with senior leaders and leadership teams navigating complexity, organizational change, and the evolving demands of leadership. Drawing on more than 20 years of experience across organizations including NASA, Goodyear, and Babcock & Wilcox, she helps leaders strengthen effectiveness, team performance, and organizational capabilities.
