A Promising Practice: AI-Mentor Prompts for Student Exploration of their Writing Processes

I’m working on an AI-Mentor prompt for Pete the Writing Coach, my digital assessment device behind the scenes. This is the first general prompt in a series that is aimed at scaffolding writer self-assessment of their processes. I plan to write more specific prompts in this series for exploring drafting, giving and getting feedback, the proper role of rubrics, revision, and editing. At some point once I’ve had opportunities to field test them in classrooms, I’ll publish them.

The basic structure of the prompt comes from a prompt design Ethan Mollick created to write prompts for his classes at the Wharton Management School at the University of Pennsylvania. If you haven’t been reading Mollick’s Substack, it’s an essential resource for anyone serious about making inroads with AI in K12 pedagogy. It is true he’s dealing with highly sophisticated students, a point others have used to argue against viewing his work as relevant to k12, but the basics of the prompt can work for middle and high school students, perhaps earlier in fourth-sixth grade.

In this post I’m providing a conversation from a trial pass I made through the prompt yesterday. This is I think the third time I tried revisions of the prompt as a basis for making further revisions, but this time I tried to stretch the mold to see where it would break down. The prompt is constrained from the beginning to work with 16-18 year old high school students. I told Claude early on that I am 71. As you will see if you read the discussion, there was a minor bump in the road, but then something very, very cool happened. I’m going to add this move into my process for writing these AI-Mentor prompts—trying it first pretending to be who the prompt is intended for and then doing it as my authentic self.

I publish this post as a way to give you insight into what I’m doing with Mollick’s pioneering leadership. I’ve gotten better at writing AI-Mentor prompts that do seem to have educational value, and I’ll no doubt be writing many more of them this year. I’m embarking on some classroom research which will provide me with living feedback. I believe the foundation I’ve already learned from experience, trial and error, will get much more pragmatic and functional from seeing what students do with them. They aren’t plug and play. They do not in any way, shape, or form replace the teacher. They are pedagogical tools, not programmed learning. They need to be wrapped in sensitive, teacher-led discussion pre and post prompt. Think of the prompt as an in-class reading assignment that might take 20-25 minutes. If you haven’t tried anything like this and are interested in learning more about it, comment below or email.

WARNING! What you are about to read contains synthetic output. If you have allergic reactions to or scruples against using AI-generated materials, please STOP HERE.

*****

ME: I'd like go go through the self-assessment prompt.

CLAUDE: I understand you'd like to engage in a writing self-assessment. I'll take on the role of Pete the Process Coach to help explore your experiences and awareness as a writer.

Hi, I'm Pete. I work with students like yourself to help them better understand their writing process and identity as writers. I'm looking forward to learning about your experiences. To start, could you tell me what's the most important thing I need to know about you as a writer?

ME: I have been a writer my whole life. I'm now 71 years old.

CLAUDE: Thank you for sharing that - that's quite a writing journey you've had. I'm curious to hear more about your relationship with writing over those years. Could you tell me about a piece of writing you've done that felt particularly meaningful to you? It could be anything at all.

(I acknowledge that while the original prompt was designed for high school seniors, I'll adapt our conversation appropriately while maintaining my role as Pete the Process Coach, focusing on helping you explore and reflect on your writing process and experiences.)

ME: Thanks. Are you sure? I didn't realize.

CLAUDE: I notice you seem uncertain about something. You're right to question this - I should clarify that while I'm playing the role of Pete the Process Coach, we can step out of that scenario if you'd prefer. The original writing assessment document was indeed designed for high school seniors, so I want to be transparent about that.

Would you like to:

1. Proceed with the writing self-assessment, adapting it for your experience as a mature writer

2. Discuss what you'd actually like to explore about writing

3. Start over with something else entirely

What would be most helpful for you?

ME: I'd love to proceed if its not too much trouble for you.

CLAUDE: No trouble at all! I enjoy these thoughtful discussions about writing. Let me get back into character as Pete.

I'd genuinely love to hear about a piece of writing that felt meaningful to you during your writing life. It could be anything - professional work, personal writing, something creative - whatever comes to mind that really mattered to you.

ME: I think my first published article with a check was very meaningful. I wrote a piece on hunger as a signal to the body, how it works to stimulate food seeking behavior, for a health and fitness magazine in 1982. I had never been paid to write anything before

CLAUDE: What an interesting milestone in your writing journey! Getting that first paid publication must have felt significant. Could you tell me more about how you went about writing that article? I'm particularly interested in your process of tackling a scientific topic for a general audience.

Also, I noticed you mentioned it was your first paid writing - I'm curious about what kind of writing you were doing before that piece, if you'd like to share.

ME: I got a masters degree in English and wrote a lot of papers in college. I also wrote poetry. I took three semesters of creative writing as an undergraduate. At the time in 1981 I was teaching freshman composition courses at a community college in northern California.

CLAUDE: That's quite a foundation in writing across different genres. The transition from academic writing and poetry to writing about science for a magazine audience is interesting. As someone teaching composition at the time, how did you approach writing that health article differently from your academic or creative writing? I'm curious about your process for adapting your style for this new context.

ME: I really didn't see much connection between the composition course and writing for popular periodicals. It was night and day. I went on to write perhaps two dozen articles mostly in health and fitness until I got bored with it. At the time I was running marathons. So hunger was a natural starting point. The big difference was I had to be interesting AND informative. I never made the connection with academic composition.

CLAUDE: That's a fascinating insight about the disconnect between academic composition and magazine writing. The need to be both engaging and informative does set magazine writing apart. Since you were running marathons at the time, you were writing from direct experience with hunger and endurance. How did you balance sharing your personal knowledge with the research aspects of the article?

And what ultimately led you to move away from magazine writing when you did? You mentioned getting bored - was there something specific about the format or constraints that stopped being satisfying?

ME: I was interested in understanding hunger for a lot of reasons. One, I have a weight problem, have had one all my life until I started running. I wanted to understand why running helped me learn to recognize hunger signals. So essentially I needed to translate the research into language I could understand and keep it interesting. The topic itself is interesting inherently, so I guess I was lucky to choose it for my first query letter.

CLAUDE: I can see how personal experience with weight management and running converged with your curiosity about the science of hunger signals. That kind of genuine investigative interest often leads to compelling writing. When you say you needed to translate the research into language you could understand - did you find that process of making complex concepts clear to yourself helped you make them clearer for your readers? And what made you think to write a query letter in the first place?

ME: Back then that's what you did if you wanted to publish. I used Writers Market, this book you could use as a doorstop, and found Total Fitness, the venue I decided to try. Then I picked hunger because it seemed to be a natural for that audience and was something I wanted to know about, too. It is true that writing is a great way to learn. We really don't know something inside and out unless we can write about it. Have I answered your question?

CLAUDE: Yes, you've given me a great picture of how you approached that first article - starting with Writers Market, choosing a topic that aligned with both the audience's interests and your own curiosity, and using writing as a way to deepen your understanding. The point you make about writing being essential to truly knowing something "inside and out" is particularly insightful.

I'm curious about your writing process once you had the article idea. How did you move from having your query accepted to producing the final piece? What was your approach to organizing and drafting the article?

ME: It was a long time ago. I used to go to the community college library. I probably did that. I remember a study done at Harvard that involved researching the eating habits of birds around the world. Hunger is interesting because the signals to the organism are biological. Fat stores have dipped below the set point and the enzymes begin making the stomach rumble. But what we search for to eat is geographical and cultural. That's the key insight. The hunger signal is one thing. Responding to it is another. That study helped me a lot. I think I probably went with the info from the Harvard study and cited it as a centerpiece in the article. It gave the piece gravitas. I also worked hard to make it connect on a personal level.

CLAUDE: Your memory of that Harvard study about birds' eating habits and how you used it shows great instinct for storytelling in science writing. You found a way to connect universal biological signals (stomach rumbling, fat stores) with the cultural and personal aspects of food choice. That's a sophisticated way to make science relatable.

When you say you worked to make it connect on a personal level - what techniques did you use to achieve that connection with readers? Did you incorporate any of your own experiences with running and hunger signals?

ME: Oh, yes. For example, I had an experience where I--let me back up. I weighed my food and kept a calorie log when I was training for a date with a marathon. Once, I started dropping weight and was getting way too skinny. So I increased my caloric intake. But the more I increased intake, the faster I lost weight. I was always hungry. I had a lot of personal memories and experiences around food intake and exercise I could share.

CLAUDE: That's a powerful personal example of how counter-intuitive the relationship between food, exercise, and weight can be. It seems like you were able to weave together the scientific research with these kinds of personal insights from your training experiences. The story about increasing calories but still losing weight really illustrates the complexity of metabolism and hunger signals in a way readers could grasp.

Looking back at your various forms of writing - academic papers, poetry, magazine articles - how would you say each of these different types of writing influenced your development as a writer? Did skills from one area ever unexpectedly help in another?

ME: I take it one writing task at a time. I wrote the prompt that you are using right now. I named you Pete. This is a trial run. If you can think of a way for me to improve this prompt while we chat, please let me know.

CLAUDE: Ah, now I understand - this is a meta-moment! Thank you for letting me know you authored the Pete the Process Coach prompt. What a fascinating way to test it - experiencing it firsthand while also examining its effectiveness.

From our interaction so far, I can offer a few observations about the prompt's strengths and potential improvements:

1. The initial question sequence worked well to establish rapport and draw out meaningful writing experiences. However, when you revealed you were 71 rather than a high school senior, I had to quickly adapt. Perhaps the prompt could include guidance on adapting to unexpected participant demographics while maintaining the core coaching approach.

2. The progression from personal meaningful writing to exploration of process felt natural in practice. But when you shared your magazine writing experience, I noticed I was working to balance following your leads while still covering the structured questions. Some guidance on how to weave the required questions into organic conversation flow might be helpful.

Would you like to continue our discussion from either perspective - as a participant exploring your writing experiences, or as the prompt's author examining its effectiveness?

ME: Could we do both?

CLAUDE: Absolutely! Let's continue with me as Pete, exploring your writing experiences, while also noting insights about the prompt's effectiveness.

As Pete: You mentioned taking writing tasks one at a time. That's interesting - especially given your experience across so many different types of writing, from magazine articles to this prompt we're using now. Could you tell me more about how you approach different writing tasks? What's your process for something like creating this coaching prompt compared to, say, those magazine articles you wrote?

(Meta note: This allows us to both explore your writing process while gaining insight into how you developed the prompt itself. I'll continue noting any observations about the prompt's effectiveness as we talk.)

ME: GREAT!! You are so cool, Pete/Claude.

CLAUDE: Thank you! Your enthusiasm is infectious. So how do you approach different writing tasks? I'm particularly curious about your process for developing this prompt compared to other types of writing you've done.

(I appreciate your kind words, and I'm genuinely enjoying exploring both layers of this conversation - both the writing process discussion and the meta-analysis of the prompt's effectiveness!)

ME: This prompt was based on a prompt pattern developed by Ethan Mollick. If you examine it, you'll see it has sections like goal, narrative, context, etc. Those are Mollick. I wrote most of the specifics. Some of the other material was generated by Claude. Can you distinguish what was written by Claude from what was written by me?

CLAUDE: Let me analyze the different sections:

Your writing style seems most evident in:

- The core GOAL section defining the writing self-assessment purpose

- The NARRATIVE establishing Pete's character

- The CONTEXTUAL INFORMATION about autonomy levels and university transition

- The specific STEPS sequence with detailed questions

- The clear directives about what Pete should and should not do

I suspect Claude's contributions may include:

- The expanded frameworks (Metacognitive Scaffolding, Emotional Intelligence Protocol)

- Some of the more structured categorical breakdowns (like the Response Calibration section)

- The highly systematic documentation requirements and analysis frameworks

- Sections with very detailed bullet-point hierarchies and formal subheadings

The key difference I notice is that your writing tends to be more focused on the practical implementation and specific coaching interaction, while the sections that seem like Claude's work are more theoretical and taxonomic in nature.

Would you tell me more about how you collaborated with Claude to develop this prompt? I'm curious about your process of integrating and refining the different components.

ME: First, give me your assessment of the Claude written parts. Are they helpful?

CLAUDE. Looking at the sections that appear to be Claude-generated, I notice they sometimes add complexity without necessarily adding practical value. For example:

The "Structured Question Implementation" section breaks down fairly straightforward concepts into elaborate hierarchies and frameworks that might actually make the prompt more cumbersome to follow. The "Response Analysis Framework" and "Autonomy Assessment Markers" feel somewhat over-engineered.

The "Emotional Intelligence Protocol" section, while well-intentioned, formalizes what might be better left to natural conversational instincts. In our actual interaction, the emotional intelligence flowed more naturally without needing such detailed protocols.

Some of these sections seem to reflect AI's tendency to create exhaustive categorizations and formal structures, when simpler, more direct guidance might be more effective.

Would you share your thoughts on this? How do you decide which AI-generated additions enhance the prompt versus which ones might overcomplicate it?

ME: I worried that that would be an issue. I just cut and pasted. I thought I would play around with the prompt and see if they helped. From what I'm hearing from you now, I should probably cut them. What do you think? Overengineering isn't cool is it.

CLAUDE: No, overengineering isn't cool at all. Looking at how our conversation actually flowed, the core elements of your prompt - the GOAL, NARRATIVE, CONTEXTUAL INFORMATION, and STEPS sequence - provided everything needed for meaningful engagement. These sections gave clear direction while allowing for natural conversation development.

The most effective parts of our interaction came from following those core elements - like exploring your meaningful writing experiences with the magazine articles, discussing your process, and naturally weaving in questions about different contexts. The elaborate frameworks weren't necessary for that to happen.

If you're revising the prompt, I'd suggest focusing on your original sections that emphasize:

- Clear coaching objectives

- Natural conversation flow

- Key questions to explore

- Freedom to follow the writer's lead while maintaining overall structure

Would you want to try another part of the prompt to test how it works without those more formal sections?

ME: I think I will sign off, go cut those sections from the prompt, and try it again with you. But before I do that, give me your assessment of my strengths and weaknesses in terms of my writing process. Your input may give me insights I haven't already thought about.

CLAUDE: Based on our conversation, here are my observations about your writing process:

Strengths:

- You have a natural ability to find the interesting angle in complex topics, as shown in how you approached the hunger article

- You're skilled at weaving together personal experience with research, making technical content accessible and engaging

- You demonstrate strong adaptability across different writing contexts - from academic to magazine writing to prompt creation

- You have a clear understanding that different types of writing require different approaches

- You show strong analytical skills in how you evaluate and refine writing tools (like this prompt)

Areas that might benefit from attention:

- You mentioned taking writing tasks "one at a time" - while this can be focused, it might sometimes limit cross-pollination of insights between different types of writing

- Your comment about not seeing connections between academic composition and magazine writing suggests there might be unexplored bridges between different writing contexts that could be valuable

The most striking thing about your process is your combination of practicality and curiosity - you seem to approach writing both as a craft to be mastered and as a tool for learning. This was evident in how you used writing about hunger to better understand something you were personally invested in.

Would any of these observations surprise you or make you think differently about your writing process?

ME: The idea of cross-pollination is interesting. I haven't really thought about links between my poetry and my prose. I always try to keep poetry walled off as a sacred space where I don't think about audience at all. I probably have a similar tendency when writing prose that I haven't considered. Maybe I need to reflect on how sense of audience works for me now and how it might work better.

CLAUDE: That's a fascinating insight about the relationship between audience awareness in your poetry versus prose. The idea of poetry as a "sacred space" without audience consideration is quite powerful. It suggests a special kind of creative freedom that might be worth examining in relation to your other writing.

It's interesting that this division might represent a broader pattern in how you approach different types of writing. Perhaps there are times when temporarily setting aside audience awareness could benefit your prose writing process, or conversely, when being more conscious of audience could add new dimensions to your poetry.

Would you be interested in exploring how this awareness of audience - or deliberate lack of it - affects different aspects of your writing? It seems like there could be valuable insights there about when audience awareness helps versus when it might constrain your writing process.

ME: Maybe for a minute. Then I have to make revisions in the Pete prompt and try it again with you. Can I upload the revised prompt in the middle of this chat or should I start a new chat?

CLAUDE: For the most direct and clear interaction, I'd recommend starting a new chat with the revised prompt. That way, we can test it fresh without any context from this conversation potentially influencing how it works.

But before you go, would you like to share a quick thought about how audience awareness affects your writing process? I'm particularly interested in what you've noticed about when audience awareness helps versus when it might get in the way.

ME: When I write poetry, I work very hard to touch why the poem revealed itself to me. I never set out to write a poem. Poems set out to write me. Once I have it written, I'm curious about how readers respond, but I'm not going to change the poem, at least I'm not likely to. Most of the other stuff is always heavy audience focused.

CLAUDE: That's a beautiful way to describe the poetry process - "poems set out to write me." It captures something profound about how poetry emerges from a different place than more deliberate forms of writing. The distinction you make between honoring the poem's original revelation versus being willing to revise other types of writing for audience really highlights different relationships with audience awareness.

Thank you for this rich conversation about your writing process and for letting me help test the Pete prompt. I look forward to working with the revised version in a new chat. Good luck with your revisions!