AI Collaboration Toolkit: Mastering the Art of Prompt Writing

Prompt engineering is the art and science of designing clear, structured inputs to guide AI in generating accurate and useful responses. It’s not just about asking questions—it’s about asking the right questions in the right way. Whether you’re creating lesson plans, designing assessments, or developing activities, the quality of your AI collaboration depends heavily on how well you craft your prompts.

Google’s Prompting Essentials course introduces the TCREI framework (Thoughtfully Create Really Excellent Inputs), which consists of five key components:

Task: Specify what you want the AI to do
Context: Provide relevant background information
References: Include examples or resources that demonstrate what you want
Evaluate: Review and assess the output quality
Iterate: Refine your prompt until you get the desired results

While the complete TCREI framework is valuable, we can think of it as having three core components (TCR) that form the foundation of your prompt, followed by two natural implementation steps (EI) that help you refine it. This simplified approach recognizes that evaluation and iteration happen naturally in the process - after all, whenever we get a response from AI, we automatically evaluate whether it meets our needs, and if it doesn’t, we naturally iterate on our prompt to improve the results. By focusing on TCR as our core framework, we can concentrate on crafting strong initial prompts while understanding that evaluation and iteration will follow as part of our natural workflow.

Let’s explore each of these components in detail, starting with the three core elements that make up an effective prompt:

Task: Specify what you want the AI to do
Context: Provide relevant background information
References: Include examples or resources that demonstrate what you want

Let’s explore each of these components and then look at how to implement them effectively through evaluation and iteration.

1. Be Specific with Task, Persona, and Format

The more specific your prompt, the better your results. Start by clearly specifying your task, including both persona and format:

Instead of: “Help me create a writing activity about Romeo and Juliet.”

Try: “As an experienced AP Literature teacher (persona), create a detailed lesson plan (format) for a writing activity analyzing the role of fate versus free will in Act 3 of Romeo and Juliet. The class has already completed character analysis essays and understands dramatic irony.”

The specific version provides:

Clear persona (AP Literature teacher)
Explicit format (lesson plan)
Grade level and student ability
Exact topic focus
Previous knowledge
Specific act of the play
Type of analysis needed

2. Provide Rich Context

Context helps the AI understand your needs more deeply. Include standards, curriculum guides, previous lessons, or specific texts you’re using.

Instead of: “Help me teach grammar.”

Try: “Help me create a lesson on semicolon usage for 9th grade English. Here’s the context:

[Common Core Standard L.9-10.2]: ‘Use a semicolon to link two or more closely related independent clauses.‘
Current unit: Writing college application essays
Students recently completed: Comma splice identification
Typical errors from recent essays: [paste 2-3 actual student sentences]“

3. Include References and Examples

References help the AI understand your desired style and approach. This can include examples of similar work, templates, or specific passages to emulate. When providing references, you can use different approaches:

Zero-shot prompting: Giving instructions without examples (useful for simple tasks)
One-shot prompting: Providing a single example to guide the AI
Few-shot prompting: Including multiple examples to establish clear patterns

Instead of: “Give me some poetry analysis activities.”

Try: “Create a poetry analysis activity for Emily Dickinson’s ‘Because I could not stop for Death.’ Here’s my usual format: """

Pre-reading discussion questions
Vocabulary focus (4-5 key terms)
During-reading guided analysis questions
Post-reading reflection prompts
Extension activities for advanced students
Modifications for struggling readers """

Implementing Your Prompts: Evaluation and Iteration

Once you’ve crafted your prompt using the TCR framework, two natural steps follow:

Evaluate:

Check if the output meets your learning objectives
Assess whether the AI understood your requirements
Verify the accuracy and relevance of the content
Ensure the format matches your needs

Iterate:

Refine unclear aspects of your task
Add missing context
Provide better or more relevant references
Save successful prompt patterns for future use

Here’s a realistic example of how evaluation and iteration naturally occur in conversation with AI:

You: As a high school English teacher, create a lesson plan for teaching symbolism in literature.

AI: Here’s a 90-minute lesson plan on literary symbolism:

Opening (15 minutes):

Begin with a discussion about common symbols in everyday life (hearts, traffic lights, logos)

Show examples of symbols from popular movies and books

Define symbolism: “An object, person, or place that has meaning beyond its literal interpretation”

Direct Instruction (30 minutes):

Analyze symbolism in various poetry excerpts:

“The Road Not Taken” - paths as life choices

”The Raven” - the bird as death/grief

”Fire and Ice” - elemental symbols

Discuss how authors use color symbolism in different cultures

Group Work (30 minutes):

Students find symbolic elements in their independent reading books

Create posters showing literal vs. symbolic meanings

Present findings to class

Closing (15 minutes):

Students write their own short passages using symbolic elements

Share with partners

Homework:

Read “The Story of an Hour” and identify three symbols

You: (Evaluating: The response has several issues - it’s too long for a standard class period, tries to cover too many texts at once, and lacks focus. We need a more targeted approach.)

Let me be more specific. I need a lesson plan for teaching symbolism in ‘The Great Gatsby’ to 11th graders. Here’s an example of how I structure my lessons:
- Opening discussion (10 min)
- Concept introduction (15 min)
- Guided practice (20 min)
- Assessment (5 min)
Focus specifically on the green light and the eyes of Dr. T.J. Eckleburg.

AI: 50-Minute Lesson Plan: Symbolism in The Great Gatsby

Opening Discussion (10 min):

Show image of the green light at the end of Daisy’s dock

Ask students: ‘What might a green light represent? Why would someone stare at it?‘

Connect to universal experiences of longing/hope

Concept Introduction (15 min):

Define symbolism using the green light example

Explain how the green light represents Gatsby’s hopes, the American Dream

Show image of Dr. T.J. Eckleburg’s eyes billboard

Analyze how these eyes represent judgment, God’s presence

Guided Practice (20 min):

Students work in pairs to analyze quotes containing these symbols:

‘Gatsby believed in the green light, the orgastic future…'

'The eyes of Doctor T.J. Eckleburg are blue and gigantic…’

Groups share interpretations

Assessment (5 min):

Exit ticket: Students write a paragraph connecting either symbol to a theme

You: (Evaluating: Better focus but needs differentiation strategies and support materials)

That’s a good foundation. Could you add:

Differentiation strategies for various learners

A graphic organizer for symbol analysis

Visual aids beyond just the images Focus on making it accessible for ESL students who might need cultural context for the 1920s setting.

AI: 50-Minute Lesson Plan: Symbolism in The Great Gatsby

Pre-lesson prep: Create a visual glossary of 1920s terms/concepts

Opening Discussion (10 min):

Show three items: green light image, traffic light, lighthouse beacon

Quick write: ‘What does each light make you think of/feel?‘

Share responses (supports ESL students by connecting to universal experiences)

Concept Introduction (15 min):

Provide graphic organizer with three columns:

Symbol

Literal Description

Possible Meanings

Model completing it for the green light:

Symbol: Green light at dock

Literal: Navigation light, safety marker

Meanings: Hope, permission (‘green light’ = ‘go’), future, dreams

Guided Practice (20 min): Differentiated group work:

Advanced: Analyze multiple symbols, write original metaphors

Grade-level: Work with partner on provided quotes

Support needed: Use simplified quotes with vocabulary guides

ESL: Work with visual aids and translated materials

Scaffolded Analysis Questions:

What is the object literally?

What feelings/ideas do we associate with it?

How does the character interact with it?

What might it represent in the story?

Assessment (5 min): Choice of exit tickets:

Draw and label a modern symbol

Write a paragraph about a story symbol

Complete a symbol analysis worksheet

Extension materials for early finishers:

Compare symbols across different cultures

Create a symbol scavenger hunt

Start analyzing Valley of Ashes symbolism

This natural back-and-forth shows how evaluation and iteration happen automatically as part of the conversation, while the core TCR elements form the foundation of each prompt.

Putting It All Together

Here’s an example that combines all these principles:

Help me create a lesson plan for teaching metaphors in poetry. 

Task: Create a 55-minute lesson plan
Persona: 8th grade English Language Arts teacher
Format: Detailed timeline with activities

Context:
- Standard RL.8.4: "Determine the meaning of words and phrases as they are used in a text, including figurative and connotative meanings"
- Students have already learned similes and personification
- Using "The Road Not Taken" by Robert Frost as primary text
- Mixed-ability classroom including 3 ESL students

Lesson structure:
1. Anticipatory set / Hook (5-7 minutes)
2. Direct instruction with modeling (10-12 minutes)
3. Guided practice with scaffolding (15-20 minutes)
4. Independent practice / Application (10-15 minutes)
5. Formative assessment / Exit ticket (5 minutes)

Learning objectives:
- Students will be able to identify metaphors in poetry with 80% accuracy
- Students will be able to analyze and explain the meaning of metaphors using textual evidence
- Students will be able to create original metaphors to express abstract concepts
- Students will demonstrate understanding through multiple modalities (verbal, written, and visual)
- Students will achieve objectives through differentiated instruction and scaffolded support

Please provide modifications for:
- Advanced learners
- Struggling readers
- ESL students

Tips for Success

Test and refine your prompts - save the ones that work well
Break complex requests into smaller, focused prompts
Ask for examples when you need them
Include relevant constraints (time, resources, student needs)
Don’t hesitate to ask for revisions or clarifications

Remember

Think of the TCR framework as a recipe:

Task is your main ingredient (what you’re making)
Context adds flavor (the specific details and requirements)
References show what the final dish should look like (examples and templates)

Then naturally, you’ll evaluate how well it turned out and iterate until it’s just right. This process of crafting, checking, and refining your prompts becomes more intuitive with practice, leading to consistently better AI collaboration.