Why Generic Prompts Fail
One of the most common frustrations educators experience with AI is getting results that do not quite meet the request, or are too generic. The reason is not that AI is ineffective; it’s that humans and AI communicate very differently. When humans communicate, we rely on shared experiences, implied meaning, tone, and assumptions about what should be obvious. AI does not have access to any of that. Instead, AI treats every prompt as a standalone instruction.Because AI lacks the ability to deduce meaning that hasn't been explicitly provided, it does not draw conclusions about intentions. It does not know what matters most unless explicitly told.
This is where many prompts break down. Educators naturally assume context: grade level, student needs, curriculum alignment, but AI requires that context to be explicitly stated. As a result, poor prompts result in generic, low-value results. AI responds to clarity, context, and constraints. In other words, prompting isn’t about asking. It’s about communicating in a clear, explicit, and structured way of giving instructions that align to how AI processes language.
Deep Dive into RCCOI: A Better Way to Prompt
To move from generic outputs to meaningful instructional support, educators need a repeatable framework. The R.C.C.O.I. model provides exactly that:
Role – Who should the AI act as?
Context – What is the instructional situation?
Constraints – What boundaries or requirements matter?
Output – What do you want the AI to produce?
Iteration – How should it refine or follow up?
Let’s apply this to a classroom example. Imagine you’re teaching The House on Mango Street and want a differentiated lesson plan. A weak prompt might look like “Create a lesson plan for The House on Mango Street.” A stronger RCCOI-aligned prompt would look more like:
“Act as a high school English instructional coach. I am teaching The House on Mango Street to 9th-grade students with a mix of reading levels, including English learners. Create a 50-minute lesson plan focused on theme and identity, including scaffolds for struggling readers and an extension task for advanced students. Provide discussion questions and a formative assessment. Before finalizing, ask clarifying questions and then provide two variations: one more scaffolded and one more rigorous.”
This works because the role prevents generic responses by shaping tone and expertise. The context anchors the task in a specific text, grade level, and objective. Constraints ensure differentiation and appropriate rigor. Output defines exactly what should be produced. Iteration turns the prompt into a conversation, reduces guessing, and improves quality through refinement. Instead of hoping AI understands your needs, you design the conditions for quality output.
Troubleshooting: When Results Miss the Mark
Even strong prompts don’t always produce perfect results, and that’s normal. The key is not to start over, but to diagnose. When an AI prompt doesn’t produce the results you expected, the solution is almost never ‘the AI doesn’t work’; it’s usually a signal to adjust how you’re communicating. Instead of rewriting from scratch, begin by pausing to analyze what went wrong. Ask yourself whether the response is too generic, off-grade, off-topic, or missing important context or constraints. Often, the issue isn’t the tool; it’s that the prompt didn’t fully communicate what mattered most.
From there, revisit your prompt using the RCCOI framework as a diagnostic tool. Did you clearly define the role the AI should take? Did you fully explain the instructional context, including grade level, content, or student needs? Were there meaningful constraints to guide the level, tone, or rigor? Did you explicitly state the desired output? And did you allow for iteration? Gaps in any of these areas can lead to vague or misaligned results. In many cases, small refinements, such as adding grade-level expectations, clarifying the format, or including a time frame, can significantly improve the output without requiring a full restart.
It’s also important to remember that you can build on what the AI gives you rather than discarding it. Iteration is where much of the value happens. You might ask the AI to revise its response to be more rigorous, simplify it for younger learners, or better support English learners. If the response still feels off, prompt the AI to ask clarifying questions before revising. This turns the interaction into a feedback loop rather than a one-time request. Ultimately, improving AI output is less about starting over and more about refining your communication and applying professional judgment throughout the process.
The Ethical Guardrail: AI Is a Drafting Tool, Not a Decision-Maker
As powerful as AI can be, it’s critical to stay grounded in its limitations. The AI produces content that appears correct, but may still be incorrect. That’s because it operates on probability, not understanding, judgment, or professional responsibility. This leads to one essential principle: AI can accelerate your work, but it cannot replace your expertise.
It is essential to always remember that AI outputs are drafts. In addition to inaccuracy, bias and appropriateness must always be checked. Educators remain responsible for final decisions. AI is most useful for high-impact use: tasks that are effective for automation and amplification. It is using AI to measurably improve teaching and learning conditions by saving educator time, strengthening instructional quality, and increasing human capacity without replacing professional judgment.
Call to Action
When educators shift from implicit communication to explicit design using frameworks like RCCOI, AI becomes far more than a novelty. It becomes a practical, powerful partner in planning, differentiation, and instructional improvement. Instructure’s Professional Learning & Strategy team provides strategy on effective prompting using the RCCOI model, as well as the AI Toolkit for Teachers training session.
The one-hour Get the Most Out of AI through Effective Prompting strategy session explores how Generative AI processes a prompt, the core components of an effective prompt, and practices diagnosing and iterating on results to get the most out of AI. The AI Toolkit for Teachers training session is a 90 minute workshop that focuses on transforming AI into educators’ most efficient professional assistant, and crafting precise prompts with the RCCOI framework. For more information, reach out to your Customer Success Manager.
