I asked AI to analyze my questions: this is what came out
AI is becoming more mature. But how do I actually use AI myself? I regularly work with ChatGPT, Copilot, Mirage, and NanoBanana (for images). I sometimes have mixed feelings about that. Does it really add value? Am I wasting energy? And am I still sufficiently in control? That’s why I asked one of these systems—ChatGPT itself—to analyze how I’ve been using this tool over the past period.
It turns out I use ChatGPT for:
- Strategic change and communication questions
- Structuring and organizing information
- Translating ideas into work formats
- (Re)writing texts
- Consistency checks
- Reflective and testing questions
- Practical help questions
In the analysis below, the “I” refers to ChatGPT and the “you” refers to me as the subject of analysis. I add my personal reflections as “Reflection.”
ChatGPT’s analysis
Based on your questions over a longer period of time, a fairly clear pattern emerges. Below, I (ChatGPT) have analyzed and clustered your questions, and for each cluster I describe:
- the type of questions they are
- what they say about your way of working
- typical examples (paraphrased)
1. Strategic change and communication questions
“Is this approach substantively and strategically sound?”
This is the largest cluster. You use ChatGPT as a sparring partner at a strategic level.
Characteristics
- You test whether an approach is logical, complete, and consistent
- You explicitly connect communication to change, behavior, and organizational context
- You are not looking for standard models, but for tailored solutions
Typical questions
- Does this communication approach fit the change challenge?
- Which change color is dominant—and what does that mean?
- Am I missing an important track (leadership, channels, monitoring)?
- Where are the risks or blind spots?
What this says
You think like a designer of change, not like an executor of tools. Communication is never a standalone product for you.
Reflection – Huib
I do indeed use ChatGPT as a kind of sounding board. In my personal profile, I’ve entered a lot of information from my two most recent books, so the system knows how I think. That usually works well, although you do need to stay sharp. Sometimes it adds a bit too much of its own interpretation. But as a check, it’s really helpful. It keeps me on my toes.
2. Structuring, organizing, and clustering
“Help me see patterns in complexity.”
A very clear second cluster: you let ChatGPT organize what is diffuse or overwhelming.
Characteristics
- Intake forms, feedback, loose ideas, contradictory answers
- You explicitly ask for clustering and for naming differences
Typical questions
- Can you cluster these intake responses?
- Where are the contradictions or tensions?
- Can you summarize this per question?
- Do you see dominant themes or fragmentation instead?
What this says
You work a lot with groups, diversity, and multiple perspectives. You look for overview without flattening nuance.
Reflection – Huib
For me, AI is a real time-saver here. I like to use online intakes, and AI is incredibly fast at clustering those. As a result, I’m using online intakes more and more. And just to reassure you, I also have regular conversations with participants.
3. Translating ideas into work formats, cases, and interventions
“How do we make this concrete and usable?”
You are strikingly often engaged in the step from thinking to doing.
Characteristics
- Workshops, training sessions, off-sites, learning situations
- You want participants to experience and apply things themselves
Typical questions
- What is a good case for this group?
- How can they really try out the canvas for their own situation?
- Which work format fits this change color?
- How do I make this interactive instead of just broadcasting?
What this says
You have a strong didactic and facilitative orientation. Not “I explain it,” but “they need to work with it.”
Reflection – Huib
I love how ChatGPT helps me offer tailored solutions, for example in in-company training sessions. Sometimes during a training, it turns out the group wants to further develop a specific case—and I can then prepare that case on the spot.
4. Writing, rewriting, and sharpening tone
“Am I saying this right—and for whom?”
A large part of your questions revolves around language, style, and readability.
Characteristics
- From strategic documents to LinkedIn posts
- You actively steer tone: empathetic, clear, not too bullet-heavy
Typical questions
- Can you write this out as continuous text?
- Fewer bullets, more story
- Is this empathetic enough?
- Is this understandable for employees / students / managers?
What this says
You are language-aware and audience-aware. For you, form is not cosmetic, but part of the effect.
Reflection – Huib
This is one of the most frequently used applications of AI for me. It’s especially enjoyable to move from change-related questions to a first draft of a change narrative.
5. Consistency with your own models and vision
“Does this fit my way of looking at things?”
You explicitly use ChatGPT to ensure everything aligns with your own framework.
Characteristics
- Regular references to your canvas, change colors, guiding principles
- You correct it when something feels “too generic” or “not your approach”
Typical questions
- Can you link this to my Internal Communication Canvas?
- Where do you see my change communication approach reflected?
- Which color dominates here?
- Does this fit with empathetic change narratives?
What this says
You don’t see ChatGPT as an author, but as a co-thinker within your frame. You retain ownership.
Reflection – Huib
A very practical way to see which “change color” is dominant in a piece of text. Extremely useful.
6. Reflective and testing questions
“Where could this go wrong?”
You ask critical questions about assumptions and risks remarkably often.
Characteristics
- You look for friction, not confirmation
- You explicitly ask what’s missing or what doesn’t sit right
Typical questions
- What fits less well here?
- Which expectations are unrealistic?
- Where are the tensions between vision and practice?
- Does this feeling make sense—that this case doesn’t fit?
What this says
You are reflective and professionally critical. You use doubt productively.
Reflection – Huib
It’s great to use AI as a check to see whether you might be wearing blinders in your analysis. AI can really help you look at a situation from multiple angles.
7. Practical, instrumental questions (supportive)
“How does this actually work?”
A smaller but recurring cluster.
Characteristics
- Tools, technology, formats, small practical issues
- Always functional, never just for fun
Typical questions
- How do I set this up properly in MailerLite?
- Which communication formats are emerging?
- How do I make this suitable for mobile?
What this says
You want strategy to work in practice, all the way down to execution.
Reflection – Huib
Over the past few months, ChatGPT has been an amazing tool for implementing a new website, newsletter, and other technical systems. Whenever I got stuck, I shared a screenshot of the situation, and in almost all cases AI helped me move forward. It saved me many sleepless nights. Truly fantastic!
Overall profile (summary)
When ChatGPT puts all my questions together, it seems I mostly ask questions like:
“How do I design change and internal communication that is substantively sound, feels human, is strategically coherent, and works in practice?”
According to ChatGPT, I primarily use it as:
- a framework enhancer
- a structuring tool
- a critical mirror
- an editorial partner
And not as a quick-answer machine. It also turned out that my questions can be divided into functional questions, innovative questions, and uncomfortable questions.
- About 60–65% of the questions are functional
→ necessary to maintain pace, quality, and consistency - 25–30% are clearly innovative
→ this is where new language and sharper insights emerge - 5–10% touch on truly uncomfortable questions
→ that’s where the greatest development potential lies
Of course, AI often formulates answers in a positive tone. I want to take a closer look at that, to use AI in an even more targeted way. How do you use AI? Have you ever asked the systems themselves—and do you recognize this? I’m curious to hear your responses.
Huib Koeleman
