Brainstorming with generative AI through documentation use cases

As documentarians, we've all been there: staring at a blank page, wrestling with the right words to reach our readers. You turn to your AI writing assistant, hoping it'll transform your ideas into harmonious paragraphs.

Instead, you experience an exhausting cycle of back-and-forth prompting. But here's what can change everything: providing AI with clear context through what MIT calls problem formulation. With a structured use case, I got a clear focus and outline for my documentation.

Why prompting falls short

When I jump into AI chats without structure, I go nowhere fast. I encounter two issues: I fail to share my constraints upfront, and I rely too heavily on basic prompts.

Brainstorming happens in a context

As documentarians, we naturally brainstorm within a context. We draw on our experience and training, with our goals, audience, and constraints already in the back of our mind. But generative AI can't access this mental framework unless we explicitly share it.

Think of it like onboarding a new team member—you need to explain the business, culture, and job requirements. The same applies to AI, though it can process this context more quickly. Before brainstorming, we need to get AI on the same page with our documentation needs.

Moving beyond basic prompts

While nearly 80% of users are interested in prompting skills, the term means different things to different people. MIT Sloan Teaching & Learning Technologies offers a simple view: prompts are conversation starters. They can be as simple or complex as you like.

Without structure, these conversations can quickly drift off course. A simple question could spiral into irrelevant technical discussions about proper formatting—interesting perhaps, but not helpful for documentation goals. I needed a better way to guide these conversations by giving my AI assistant clear context.

How use cases enhance your brainstorming

Here's the thing about us documentarians: we love structure, but we also need room for creativity. Our writing has an end goal, but we want the freedom to explore and use new ideas and unexpected connections. Sounds like a tricky balance, right?

You might be wondering: isn't adding structure to brainstorming defeating the whole purpose?

That's where use cases come in. I like to think of them as "brainstorming containers"—they give you just enough structure to stay focused without boxing in your creativity.

Think of it as sharing your documentarian know-how with your AI partner—the main points that experienced writers keep in their heads. Here's what I include in mine:

• Audience: Let's start with who's actually going to read this. Think about their expectations and background. What do they need to know?
• Purpose: Getting clear on the "why." What problem is this documentation solving? What are we trying to achieve?
• Style: What tone and approach will resonate with our readers while meeting stakeholder needs?
• Business context: Now let's consider the company's perspective: what drives them, where are they headed, and how can our documentation help them get there?
• Reference materials: Here's where we pull together our research: from user stories and tickets to interview insights—everything that builds our understanding.

This structured approach aligns with emerging research on effective AI collaboration. MIT's work on problem formulation emphasizes defining the problem's focus, scope, and boundaries, rather than just crafting better prompts. And the best part? Both you and your AI partner stay on track, avoiding those frustrating rabbit holes that waste your time.

Building your basic use case

Remember that new team member onboarding we discussed? Let's create that orientation packet for your AI partner. Here's how I put the bundle together:

• Audience: I tell my AI partner who we're writing for: documentarians creating and maintaining documentation, as defined in KnowledgeOwl's content strategy.
• Purpose: KnowledgeOwl states that content must be meaningful and actionable for documentarians' daily work. Again, the content strategy guides these goals.
• Style: The KnowledgeOwl style is professional but approachable. KnowledgeOwl's content strategy gives additional guidelines to nail this tone.
• Business context: I help my AI partner understand KnowledgeOwl’s customer-driven values. I link it to their “About Us” page so it can align the larger business picture to my writing.
• Reference materials: I include any other research material that would help in understanding the broader context, such as a few successful sample pitches and blogs to demonstrate what works.
• AI instructions (text file): Finally, I create a clear list of instructions for the AI model. It explains how to use each resource, what to prioritize, and when to reference specific materials.

See the screenshot below for the final result, with all these pieces together:

Once you've created this basic use case, you can reuse it whenever you start a new writing project. It's like having a well-trained writing partner who can help at any time!

Converse with the AI model about your content

With our use case ready, it's time to start the actual brainstorming conversation. Think of it like sitting down with a well-prepared colleague who already understands your goals and constraints. Here's how I approach it:

First, I did my homework. I combined my documentation experience with some research using Perplexity (an AI research tool) to learn more about documentation prompting. Just like you'd prepare for a brainstorming session with a colleague, I wanted to bring something to the table.

I shared my initial pitch with my AI partner, and here's what came back:

These questions sparked some interesting connections. At an AIPortland meetup, I'd learned about English teachers using AI for brainstorming. Through further discussions, I realized my approach to problem formulation was similar to creating user stories, a familiar concept for many documentarians.

After exploring these ideas with my AI partner, we landed on our focus: "Brainstorming with generative AI through documentation use cases." My AI partner confirmed we'd found the right direction:

"Got it - focusing on brainstorming with AI-friendly documentation practices makes sense and fits well with KnowledgeOwl's target audience of documentarians."

Because we started with a solid use case, we didn't need endless rounds of prompt refinement. We were ready to dive into outlining and writing.

Conclusion

Just as a new hire needs context to contribute effectively, generative AI needs a well-structured use case. When you take the time to create that orientation packet we discussed, you transform AI from a basic writing tool into a true brainstorming partner.

As AI tools evolve and our documentation needs change, our use cases will grow with them. But the core principle stays the same: give your AI partner the context it needs, and you'll spend less time tweaking prompts and more time developing great content. Are you ready to get started brainstorming with your generative AI model?