Coaching Through Design: Scaffolding the Writing Process with AI Tools

How I use worksheets, podcasts, and bots to guide student thinking in asynchronous environments

We’re all prompting AI these days, aren’t we? Feeding it instructions, refining outputs. But what if the real innovation isn’t in us prompting the machine, but in the machine prompting us and our students?

That question shaped much of my sabbatical this spring as I rethought my fully online, asynchronous college writing course for first-year students at a California community college. These are students logging in between shifts, caring for family, or stealing time late at night to study. In this setting, traditional strategies—like in-class writing or synchronous debates—simply don’t apply.

I’ve thought of myself as a coach for years, but AI has opened up new ways to embed that coaching role into the design of the course itself. While teaching four fully online classes this summer, I developed a three-part system to support and extend student thinking on each major writing project. Each tool offers a different form of presence, structure, and encouragement, especially for students who may never meet me face-to-face.

This isn’t about preventing cheating. It’s about trust, clarity, and giving students tools that help them stay engaged and go deeper. Here’s what that coaching system looks like right now.

📄 The Worksheet: Surfacing the Hidden Curriculum

First, I create a detailed, step-by-step worksheet. On the surface, it breaks the project into clear phases. But it also does more: it surfaces the often-hidden expectations of college writing, especially for first-generation students. It spells out not just what to do, but why, and how each move contributes to the larger writing task.

AI helped me draft many of these worksheets, filling in places where even my best-written prompts left questions unanswered. It also let me chunk each stage—one of the most effective coaching moves—so students can work through the project in manageable, bite-sized pieces. It doesn’t hand them a script; instead, it makes expectations and processes explicit while also giving students room to experiment, reflect, and revise.

🎧 The Podcast: A Coach’s Voice in Their Ear

Next, I use Google’s NotebookLM to generate a short podcast that walks students through the writing assignment. I load the prompt, worksheet, and related readings into the tool and ask it to create a custom episode that speaks directly to students—not abstractly, but personally, with guidance like: “Here’s what matters in this project. Here’s where to begin. Here’s how to use your voice.”

This is especially useful in a project like writing a ‘Dear ChatGPT’ letter, which invites students to write a letter to an AI tool, reflecting on their interactions and how AI is showing up in their lives, communities, or schools. It’s part critique, part storytelling, and part reflection on the future of writing.

Here’s the podcast I created for that assignment, titled “What Makes Us Human in the Age of AI?” Students can listen while commuting or cooking dinner. It offers not just explanation, but tone, presence, and encouragement—things that often get lost in text on Canvas, our learning management system.

A young player trains with the help of an AI tennis coach, visualizing each stroke in real time. Image generated with ChatGPT (June 2025, GPT-4o).

🤖 The PlayLab Bot: Productive Friction and Prompted Insight

Finally, I build a PlayLab bot tailored to one key stage of each assignment. In their I-Search paper, students are asked not only to find and cite credible sources, but to engage meaningfully with them—explaining why a particular quote matters, what it reveals, and how it connects to their evolving inquiry. One of the most useful tools in this project is QuoteWeaver, a Socratic-style bot that helps students unpack and reflect on a single quote from one of their sources. The bot prompts exactly that kind of thinking. It asks, “Why did you choose this quote?” “What’s the deeper idea behind it?” “What does it add to your understanding of the topic?” The goal isn’t just to plug in a citation, but to make the thinking visible. In a project that’s all about process, not just product, QuoteWeaver becomes a reflective checkpoint—helping students move from finding a source to making it speak.

This is what cognitive scientists call productive friction—a small, deliberate obstacle that slows the student down just enough to think more clearly. The bot isn’t a shortcut. It’s a kind of tennis ball machine that feeds questions, allowing the student to practice responding. The aim isn’t perfect answers, but clearer thinking.

Other bots in my PlayLab system, like SongSpark and Generate Guiding Questions For Your Writing Project, do similar work in other assignments. I wrote more about my bot experiments in Playing Seriously and Student Voice, Socratic AI, and the Future of Feedback.

Some students copy what they’ve written in the bot straight into their draft. Others just use it to clarify their thinking. Either way, it’s a tool that encourages voice and agency, not passivity.

---

I’ve written before about how design is the real challenge of AI-era education—how we need to shape the flow of learning rather than just react to its outcomes. These tools let me coach not just with comments after a draft is turned in, but also through the very structure of the course itself.

The tools don’t replace my voice as a teacher. They extend it.