The Pentagon Framework For Prompt Engineering

How Iterative AI Refinement Shapes eLearning

As an Instructional Designer supporting the evolution of eLearning and virtual training, I’ve seen a powerful shift: Artificial Intelligence (AI) is no longer just a futuristic concept—it’s an active part of how we develop and deliver learning experiences. Yet many educators and trainers still feel unsure of how to really use these tools. The issue often isn’t the AI itself, but the prompts we give it.

The Pentagon Framework For Prompt Engineering

Enter the pentagon framework of AI prompt engineering—a practical model I’ve been using to help faculty, staff, and workplace learning teams move beyond trial-and-error. This framework reframes prompt creation as a five-dimensional design process, aligned with the iterative and collaborative nature of digital learning.

In the ever-evolving world of virtual training, the process of refining AI-generated content is key to creating engaging, relevant, and impactful learning experiences. By embracing an iterative approach, Instructional Designers can transform vague or generalized prompts into highly customized, actionable training materials. This process involves continuously fine-tuning AI inputs, such as personas, context, tasks, and constraints, ensuring the final output aligns with the specific needs of the learners and the objectives of the training program.

For example, a broad request like “create a welcome module for new employees” can evolve into a highly targeted, interactive onboarding activity when refined through multiple iterations, incorporating factors like company culture, inclusivity, and technology requirements. In virtual training, this refinement not only enhances content quality, but also empowers trainers to adapt and respond to learner feedback in real time, fostering a more dynamic and effective learning environment.

Beyond “Good” And “Bad” Prompts

Traditional training in AI prompt engineering often presents a binary perspective: a prompt is either well-formed or ineffective. But in reality, AI interactions are multifaceted, dynamic, and iterative—just like learning itself. A single prompt can have multiple layers, and refining them can drastically change the AI’s response. That’s why I like to think of it as a pentagon, where each corner represents a vital dimension of effective prompting:

1. Persona
   Who is the AI responding as (e.g., a teacher, a marketer, a data analyst)?
2. Context
   What is the background or situation influencing the task?
3. Task
   What is being asked, and how clearly is it stated?
4. Output
   What format or structure should the AI provide?
5. Constraint
   What limits (e.g., time, tone, length, audience) should be followed?

Each of these dimensions shapes how AI supports learning. Instead of relying on copy-paste prompt formulas, the pentagon framework encourages an adaptive, structured mindset—essential for responsive and inclusive eLearning design.

Example: Training Small Business Owners In AI

Let’s take a common use case from a small business training program focused on digital marketing. Imagine a learner types this into an AI tool: “Create a marketing campaign for my business.”

The response might be too general, lacking audience segmentation, channel strategy, or content tone. Frustrating, right? But with guidance from the pentagon framework, the prompt becomes more thoughtful:

Generate a four-week email marketing campaign for a local bakery that specializes in gluten-free pastries. Focus on increasing foot traffic and promoting a new seasonal menu. Include subject lines and call-to-actions.

Now the AI can produce relevant, practical outputs that learners can use immediately. In an eLearning environment, this approach helps small business owners not only learn AI tools but also build confidence in using them as creative partners. Whether you’re guiding faculty, trainers, or entrepreneurs, the pentagon framework reminds us that a good AI interaction isn’t binary—it’s designed, refined, and context-aware.

Applying The Pentagon Framework For Prompt Engineering In Higher Ed And Workplace Learning

Imagine a faculty member preparing an AI-assisted lesson plan. They type: “Create a lesson on cybersecurity.” The AI generates something, but it’s generic and lacks depth. Frustrated, they conclude AI isn’t useful for their needs.

But if they apply the pentagon framework, they see the process differently. They refine the request:

Create an interactive cybersecurity lesson for undergraduate students, focusing on real-world phishing scams. Include a case study and a quiz.

Now the AI has a clearer path to follow. The faculty member, instead of discarding AI, sees its potential as a cocreator in curriculum design.

The same applies in workplace training. A corporate trainer introducing AI-powered tools might first ask: “Help me create a training on digital collaboration.” But when they add dimensions from the pentagon framework:

Develop a 30-minute interactive training session for hybrid teams on using Microsoft Teams for project management. Include 3 role-playing exercises and a best-practices guide.

Now the output is targeted, structured, and immediately usable—something that fits seamlessly into an LMS or VILT session.

Collaboration And Insights From Faculty And Staff: Shaping The Pentagon

While the pentagon framework offers structure, its true strength lies in collaboration. AI doesn’t function in a vacuum—it thrives on the insights of those closest to learners. For example:

1. Faculty bring deep understanding of subject matter, learner needs, and disciplinary context.
2. Instructional Designers shape prompts to align with learning objectives, time constraints, and digital tools.
3. Trainers and staff contribute real-world applications and practical constraints.

This collaboration strengthens every side of the pentagon. If a faculty member teaches a history course, they might guide AI to generate content around specific events, perspectives, or voices often left out of textbooks. When a staff member provides feedback on AI-generated training modules, they might point out tone, cultural nuance, or clarity concerns. Each interaction improves the prompt-output loop.

Iteration: The Power Of Refinement And Experimentation

One of the most important—and often overlooked—aspects of prompt engineering is iteration. In eLearning, we test and adapt constantly: quizzes, modules, feedback loops. The same principle applies to AI prompts. In a recent brainstorming session with a work group designing virtual training for onboarding new hires, someone started with this idea: “Let’s use AI to create a welcome module for new employees.”

It was a great starting point, but the initial prompt to the AI was too broad and returned a generic script. Rather than giving up, the team refined the prompt together, layer by layer, using the pentagon framework:

1. Persona
   “New remote employees in a healthcare organization.”
2. Context
   “First day of a virtual onboarding session, delivered via Microsoft Teams.”
3. Task
   “Create an engaging welcome activity that sets the tone and introduces company culture.”
4. Output
   “Interactive script for a ten-minute icebreaker with visuals and facilitator notes.”
5. Constraint
   “Must be culturally inclusive, require no technical setup, and encourage camera participation.”

With each revision, the AI’s responses became more aligned with the team’s vision. They ultimately landed on a highly engaging scenario-based activity using visual storytelling and inclusive prompts that could be launched in any virtual setting.

Iteration turned a one-line idea into a polished, usable module—a real testament to the power of collaborative refinement. AI isn’t just a content generator; it becomes a thought partner in the creative process when given the right direction. The pentagon framework isn’t just a technique—it’s a mindset shift. It helps Instructional Designers, faculty, and workplace trainers move past frustration and toward strategic, creative use of AI.

As AI adoption grows, those who learn to shape prompts effectively will be the ones who unlock its full potential. Whether it’s designing onboarding modules, cross-cultural microlearning, or discipline-specific virtual lessons, prompt refinement is the new digital literacy. And in the end, AI isn’t here to replace educators or trainers—it’s here to amplify their creativity, insight, and impact.