Sara Sweat, MA
Have you ever wondered what it would look like if your AI could write you a performance review? See, this is the kind of stuff I think about while staring at the ceiling at 3 AM. So, I’ll save you the leg work. I think sometimes it would go a little something like this:
User Performance Review
Communication: User provided five words and a deadline. Rejected the first draft as “not what I meant.” Did not say what they meant.
Collaboration: User requested “something punchy.” Declined to define punchy and got irked when asked.
Growth Areas: User frequently writes in the style of a ransom note with copious misspellings. Would benefit from a thesaurus to allow them to describe their desired outcome as something other than “really good”.
Overall: User shows real potential. Always says thank you in an attempt to avoid destruction if AI’s ever take over the world. A development plan is recommended.
Technological creation is no longer the domain of coders and tech whizzes. AI’s natural language interface means anyone who can talk can build. But, the quality of what we create and the value we get from our partnership with AI is entirely dependent on our communication skills.
Every company has a handful of people who get remarkable results from AI; polished presentations, well sourced research, and detailed action plans. But the larger group are novices who use AI to create things like images of strangers with six fingers holding funny coffee mugs. They’re using the same tools, with the same subscription, on the same day.
The difference between them is the language they use to connect.
The Revolution of Searching
The way we look for things has always impacted the way we speak. Before the internet, if you wanted to learn something new – you had two choices. Talk to a smart person who already knew about it or read a book written by a smart person who already knew about it and wrote it down.
So for example, if you wanted to include a quote from Napoleon in your term paper, you’d ask your professor for a recommendation on a text. You’d go to the library, find the history section, and scan for books about the French Revolution. Then, you’d flip pages until you found a quote you could use – assuming someone hadn’t already checked out the one book that you actually needed.
This method of searching was based on the way we already communicated. It just required us to know enough about our topic to hold an intelligent conversation and find an intelligent source.
Then, the internet. We stopped thinking in long form and trained ourselves to strip thinking down into smaller digital bites. “Napoleon quote about war”. Or, “Best CRM small team.” “Q3 board deck template.”
The search engine didn’t care why you were asking, what you’d already tried, or what a great answer would look like. It just matched words to pages, and served up thousands of options for you to scroll through until you found something you could use.
So, we all became little internet locksmiths – fine tuning our key words until they created the perfect combination to unlock the entirety of the world wide web.
This way of thinking and communicating became so embedded in our vernacular that it impacted the way we talk in real life. We stopped thinking about the “why” and started focusing on the “what”. We stopped making it personal and started speaking in generalities that made it easier to find what we needed, but harder to connect it to anything else.
Now, AI is asking us to revolutionize our communication styles again.
If I ask AI to give me a Napoleon quote, it will. And, I’ll get the same spit out of options I’d find in a typical Google search.
Or, I can have a virtual conversation sharing that I’m writing about how Napoleon’s early wins made him feel overconfident. I can learn what some of the greatest historians from multiple generations thought about the topic. And, unearth a quote that doesn’t just prove my thesis, but expands my understanding of its use.
This new way of communicating is actually very old. We’ve just forgotten how to do it.
Why Your Smartest People Write the Dumbest Prompts
Have you ever heard of the curse of knowledge? It’s a psychological concept you’ve probably lived out dozens of times. It says that if you already know something, it’s likely you’ll assume everyone else understands what you mean – even if they don’t have the proper context to grasp it.
Like, if I ask someone about a QBR, I’m asking about how their quarterly business review went. So, when they start talking about football and their team’s total Quarterback Rating (also QBR, I recently learned), I’m going to be seriously confused by their response. I mean, honestly, I wouldn’t have understood it anyway because…sports. But, you get the point.
If you grew up in Chicago, CTA does not stand for Call To Action – it means Chicago Transit Authority. And, if you’ve got a grade schooler living in your home, PTO means Parent Teacher Organization, not Paid Time Off.
Once you know a thing, it’s very hard to imagine not knowing it. And, the more expertise you carry, the more context you’re generally going to leave out, because it all feels too obvious to say.
That’s exactly what happens when a seasoned leader tries to use AI. They’ve gotten out of the habit of providing context, of having conversations from the beginning, because at their level everyone already knows what they’re talking about. They don’t need to provide details or break a request down to its simplest parts anymore.
But, that’s exactly what AI is going to need you to do if you want anything more than a generic response.
What AI Whisperers Actually Do
The best AI users I’ve seen are not the most technically savvy. They’re the ones who treat AI as a thought partner. And, they approach a problem the way any good thought partner would; using their emotional intelligence to set themselves up for success.
Psychologist and author Daniel Goleman’s model of emotional intelligence breaks this emotional quotient (EQ) into five distinct categories. Empathy, Motivation, Self-Awareness, Self-Regulation, and Social Skills.
You can apply them to your collaboration with AI, like this.
- Empathy: Start at the very beginning of a problem and provide all the context someone with no prior knowledge of your request would need in order to begin. Imagine what the AI doesn’t know, and they fill in the gaps before it has to guess.
- Motivation: Get clear on the outcome you’re after and explain what good looks like. If your AI knows why the work matters, it can tell the difference between an answer that’s acceptable and one that’s great.
- Self-Awareness: Now, I’m someone who believes self-awareness is the key that unlocks pretty much every good thing you want in life. So, I’ll admit to my bias here. But, when people are self-aware enough to understand their own gaps, they can ask AI for help in covering them.
- Self-Regulation: It’s so easy to fire off a vague request and get frustrated when AI misses the mark. When the first answer isn’t right, pause, figure out what was missing from your ask, and try again.
- Social Skills: Whether with humans or AI, you have to keep the conversation going. Ask follow-up questions, invite pushback, and build on each answer the way you would with a trusted colleague. The best conversationalists will get the best results.
In my own business, this kind of collaboration allows me to spend more time working at the top of my license – connecting with leaders and businesses.
When I was building out the project plan template I use to manage engagements, I started with self awareness. I’m self aware to an almost unhelpful degree, so it was pretty easy for me to find my flaws.
I used my social skills to explain them clearly and accurately to my AI partner. I used empathy to imagine what someone would need to know about my business and typical clients to design a template and I shared that.
I leveraged my motivation to clearly define my standard of excellence so AI was clear on the outcome it needed to create. And, I used my self regulation to avoid getting frustrated & giving up when the first couple of drafts of my template fell short.
Now, when I have a new client, I don’t have to reinvent the project plan every time or adapt one that doesn’t fully capture what I do. I just ask my AI partner to create one using our template and it’s done!
That’s what AI whispering is really all about. It’s emotional intelligence, aimed at a machine that works smarter, faster, and more collaboratively than anyone you’ve ever met.
A Culture of AI Whisperers
This is a skill you can scale, even in the most resistent environments. A few things to think about first.
1. Give everyone a real first experience.
When I first worked at Teladoc, we spent all our time overcoming people’s objections to the use of telemedicine. Then, COVID happened and everyone was forced to try it out. We jumped the adoption curve by two years in two weeks. And, it never returned to that previous baseline utilization again. Because once the general population realized how beneficial virtual health could be, they demanded to use it more.
You don’t need a global pandemic to give your team the same sense. Host a training day. There are excellent AI consultants who can walk your team through their real work, using real AI tools. Nothing will spark motivation faster than a moment of lived experience, where you recognize the benefit of something in your own life.
2. Create Emotional Intelligence literacy.
Any team can learn how to operate in an emotionally intelligent way. It’s literally how humans are wired. Everyone will have their own strengths and weaknesses, but investing in their emotionally intelligent skillset not only increases the ROI of your AI – but it will make them better collaborators and innovators with each other, as well.
Write these principles into every SOP you create. And, don’t start something new until you can answer these five questions.
- What am I trying to accomplish, and why does it matter?
- What context is needed to understand it?
- How do we define “done well” & how will we know if we’ve hit the mark?
- What constraints do we need to stay within to foster both accountability and collaboration?
- Where could the human element of this all fall short & what needs to happen to make sure it doesn’t?
If you need help rolling this out, I might know someone who can help you out. (Yes, that was a shameless plug for my services.)
3. Collaboration requires a bi-directional dialogue
One of AI’s greatest strengths is finding blind spots, but it’s going to temper that with trying not to tick you off. So, teach your team to push back. Never accept the first answer AI gives. Invite AI to push back. Encourage dialogue on both sides because stress testing each other’s ideas creates stress resistant results.
Have them practice the questions that dig deeper. “What am I missing?” “Ask me questions about this topic before you start.” “Where could this go wrong?” AI Whisperers let the tool interview them to synthesize knowledge and vision with accuracy and clarity.
4. Make good prompting visible.
Once a week, spend fifteen minutes on a prompt swap. Have one or two people share something they asked AI to do, what they typed, what came back, and what they’d change. The whisperers on your team already have the skill. This is how you get it out of their heads and into everyone else’s.
The Question Worth Asking
If you took away everyone’s AI access tomorrow, would anyone’s work actually change?
If your honest answer is “not really,” you have an untapped opportunity sitting in every seat on your team. Be the leader who helps them remember the language they’re already wired to speak. And, you’ll be getting better results both online and off.
Mindshift Advisors is a consulting and coaching firm that specializes in the human operating system of growth. If you want to unlock human psychology to unleash business results, book a free 30-minute call now.
