Writing The Best Possible Prompts: The Evolve Method

A sticky note caught my eye as it fluttered to the floor of my home office. "Ask AI about project rubrics?" was scrawled in rushed handwriting, one of countless prompts I'd jotted down during my early experiments with artificial intelligence. Picking it up, I smiled at my past self's tentative question mark. Those early days of random chat interactions with AI felt like learning to swim by being thrown into the deep end. Every conversation was an adventure, but also a reminder that I needed a better system.

Today, as I research AI integration within Project-Based Learning frameworks through my venture AIxPBL, that journey from chaos to clarity shapes how I guide others in leveraging AI for education. The transformation wasn't immediate. It evolved through careful observation, systematic testing, and a willingness to start small and build incrementally.

Let me take you through that evolution, using the EVOLVE Method I've developed to help others master AI interaction. This systematic approach has transformed how educators and learners engage with AI, making it an invaluable partner in the educational journey.

The EVOLVE Method: A Framework for Mastery

The EVOLVE Method emerged from my work with hundreds of educators across K-12 and higher education settings. It offers a structured path to mastering any aspect of AI interaction, but it's particularly powerful for developing effective prompts.

Let's explore each component through the lens of prompt engineering:

Evaluate: Understanding Where You Start

My initial prompt attempts resembled scattered breadcrumbs across various AI platforms. Some worked brilliantly; others fell flat. The breakthrough came when I started treating each interaction as data.

I created a simple Notion database to track:

• Prompt text
• Educational objective
• Response quality
• Student engagement level
• Necessary revisions

This systematic evaluation revealed patterns I hadn't noticed in my random experimentation. Successful prompts shared certain characteristics: clear context, specific objectives, and structured output requirements.

Vision: Crafting Better Possibilities

With baseline data in hand, I began envisioning improved approaches. Using AI itself as a brainstorming partner, I explored questions like:

• How might this prompt better serve diverse learners?
• What structure would make the output more actionable?
• Where could this prompt fit within larger learning sequences?

This visioning phase transformed vague ideas into concrete possibilities. For example, a simple prompt requesting "project ideas" evolved into a structured template that considered grade level, subject integration points, and assessment criteria.

Optimise: Selecting Your Focus

Improvement opportunities often feel overwhelming. The key is choosing one small, high-impact change. In my practice, I developed a simple optimization matrix:

• Impact on learning outcomes
• Ease of implementation
• Scalability across subjects
• Time investment required

This helped prioritise changes that offered the best return on effort. Often, the most powerful improvements were surprisingly simple, like adding clear context statements to every prompt.

Launch: From Theory to Practice

The transition from optimised prompt design to actual implementation taught me valuable lessons about change management in educational settings. Through AIxPBL research, we discovered that successful launches share common elements:

Take this real example from a high school English classroom. The original prompt was simple:

Generate discussion questions for Chapter 1 of Lord of the Flies.        

After applying the EVOLVE Method, it transformed into:

Context: Advanced Placement English Literature class analyzing Lord of the Flies, Chapter 1. Objective: Generate discussion questions that promote critical thinking and textual analysis. Requirements: - Create 5 questions that progress through Bloom's Taxonomy - Include relevant textual references - Focus on symbolism, character development, and thematic elements - Add follow-up prompts for each question to deepen discussion Format: Present questions in order of complexity, with corresponding textual evidence.        

The difference in AI responses was striking. While the first prompt generated basic comprehension questions, the evolved version produced rich, layered discussion materials that supported deeper learning.

Verify: Measuring Impact and Understanding Results

Verification requires both quantitative and qualitative measures. Through AIxPBL, we developed a simple but effective framework for prompt assessment:

Quality Indicators:

1. Response Relevance (1-5 scale)
2. Educational Value (1-5 scale)
3. Implementation Ease (1-5 scale)
4. Student Engagement Level (1-5 scale)
5. Learning Outcome Achievement (1-5 scale)

We track these metrics using a simple spreadsheet that calculates an overall effectiveness score. This data-driven approach reveals patterns and insights that guide future improvements.

For instance, we found that prompts incorporating specific student skill levels consistently scored higher in engagement and learning outcomes. A middle school science teacher reported:

"Before using the EVOLVE Method, my AI prompts were hit-or-miss. Now, by including clear skill level indicators and learning objectives, I get consistently useful content that actually matches my students' needs."

Extend: Building on Success

Extension isn't about making things more complex—it's about thoughtful expansion of what works. Through AIxPBL research, we've identified three reliable extension patterns:

1. Vertical Extension (Depth) Start with a successful prompt template and add layers of sophistication. For example, this progression in mathematics instruction:

Basic Prompt:

Generate word problems for 7th-grade algebra equations.        

Extended Prompt:

Context: 7th-grade algebra class working on single-variable equations Objective: Create engaging word problems that connect to real-world scenarios Requirements: - Generate 3 problems at each difficulty level (basic/intermediate/advanced) - Include problems from different real-world contexts (sports, shopping, science) - Provide step-by-step solution guides - Add extension questions for advanced students - Include visualization suggestions Format: Present problems in increasing complexity with teaching notes.        

1. Horizontal Extension (Breadth) Apply successful prompt patterns across different subjects and contexts. A winning prompt structure for literature analysis can be adapted for history, science, or art appreciation.
2. System Extension (Integration) Connect successful prompts into larger learning sequences. We call these "prompt chains"—series of interconnected prompts that guide students through complex learning journeys.

Practical Application: Building Your Prompt Library

As you begin applying the EVOLVE Method to your own prompt engineering, consider creating these essential templates:

1. Lesson Planning Template

Context: [Grade level + Subject area] Learning Objectives: [Specific outcomes] Student Background: [Prior knowledge + Skills] Special Considerations: [Accommodations + Extensions] Output Requirements: - Learning activities - Assessment strategies - Resource needs - Differentiation options        

1. Assessment Design Template

Context: [Learning unit + Grade level] Assessment Type: [Formative/Summative] Learning Targets: [Specific skills/knowledge] Output Requirements: - Assessment items - Rubric criteria - Sample responses - Feedback guidelines        

Student Support Template

Context: [Learning challenge + Student level] Goal: [Desired outcome] Current Status: [Present understanding] Output Requirements: - Scaffolding strategies - Practice activities - Progress indicators - Success criteria        

The Path Forward

The EVOLVE Method transforms prompt engineering from an art into a science, while maintaining the creativity and flexibility needed in educational settings. Through AIxPBL, I've seen educators move from random experimentation to systematic success, building confidence and competence along the way.

Remember:

• Start with careful evaluation
• Envision clear improvements
• Choose small, impactful changes
• Launch with purpose
• Verify through data
• Extend thoughtfully

Your journey with AI prompts is unique, but the principles of EVOLVE provide a reliable framework for growth. Begin with one template, one class, one subject. Watch what works. Build on your successes. The path to mastery is built one prompt at a time.

Phil

Please Support my Project

Hi, I'm Phillip Alcock, an ex-teacher, and current AI in Education researcher.

I’m passionate about exploring how AI can transform the learning experience to better serve students and educators.

Each cup of coffee supports my time researching, writing, and sharing new ideas, so I can continue creating practical resources and insightful articles.

If you’d like to help fuel this work, consider buying me a coffee at buymeacoffee.com/aixpbl , or via paid subscription to my newsletter on Substack:

Subscribe now at aiforbeginners.substack.com or aixpbl.substack.com

PBL Future Labs: Empowering Schools with AI and Project-Based Learning

Does your school wish to explore the exciting possibilities of AI and Project-Based Learning?

PBL Future Labs offers a range of support options tailored to your needs.

Whether your school is just beginning your PBL journey or aiming for a comprehensive integration of AI and PBL, we can help.

From introductory workshops to personalised coaching and curriculum development and implementation support, we're here to guide you every step of the way.

Connect with me directly at [email protected] to discover how we can collaborate to transform learning at your school.

Thank you!