Prompt Engineering Essentials

Master the art and science of communicating effectively with AI systems to unlock their full
potential and achieve consistent, high-quality results.
What We Will Cover Today
01 02 03

Understanding Prompts Successful vs Unsuccessful Prompts Google Prompting Essentials

Learn what makes a prompt effective and how AI Explore industry-standard best practices for crafting
systems interpret your instructions Compare examples to identify key differences in professional prompts
prompt quality and outcomes

04 05 06

T-C-R-E-I Framework Writing Effective Prompts Optimizing & Refining Outputs

Master a proven five-step methodology for Apply practical techniques and strategies to create Learn iterative improvement methods to enhance
systematic prompt construction clear, actionable instructions AI-generated results

07

Hands-On Practice
Apply your skills across ChatGPT, Gemini, and Perplexity with guided exercises
What is a Prompt?
A prompt is your communication bridge to AI systems. It combines
clear instructions with essential context, relevant references, and
desired format specifications. Think of it as a recipe: the more
precise your ingredients and instructions, the more predictable Clear Instruction
and satisfying your results.
Specific task definition

Effective prompts transform AI from a guessing machine into a

reliable assistant. They establish expectations, provide
boundaries, and guide the AI toward outputs that meet your
specific needs. The difference between frustration and success
often lies in how well you craft your initial prompt.
Context

Background and purpose

References

Examples and guidelines

Format

Structure and style

Unsuccessful Prompt Example
"Write about branding."

Vague Task
What aspect of branding? Strategy, design, messaging, or implementation? The instruction lacks specificity and
direction.

No Context
Is this for a blog post, presentation, academic paper, or social media? Without context, the AI can't calibrate
appropriately.

Missing Audience
Should this be written for beginners, experts, executives, or students? Audience determines tone and depth.

Undefined Format
How long should it be? What structure? Bullet points, paragraphs, or sections? Format shapes usability.

This prompt will likely produce generic, unfocused content that requires extensive revision. It wastes time and undermines AI's
potential to deliver value.
Successful Prompt Example

"Act as a marketing educator. Explain branding in 120 words for beginners. Include 2 examples + 1 analogy. Format: bullet points."

1 2

Role Definition Clear Scope

"Act as a marketing educator" — establishes expertise level and perspective for the response "Explain branding" — defines the exact topic without ambiguity or room for misinterpretation

3 4

Length Constraint Target Audience

"120 words" — sets precise boundaries to ensure concise, focused output "For beginners" — calibrates complexity, vocabulary, and depth appropriately

5 6

Content Requirements Format Specification

"2 examples + 1 analogy" — specifies concrete elements to include for clarity "Bullet points" — defines structure for easy scanning and comprehension
Google's Prompting Essentials
Google's AI research team has identified five fundamental principles that consistently produce superior results across different AI models and use cases. These
essentials represent distilled wisdom from thousands of prompt experiments.

1. Clear, Specific Instructions 2. Provide Examples 3. Ask for Structured Output

Remove ambiguity. Use precise verbs and Show, don't just tell. One-shot prompting Request specific formats like bullet points,
concrete nouns. Specify exactly what you want (single example) or few-shot prompting tables, or numbered lists. Structure makes
the AI to do, avoiding vague language that (multiple examples) helps AI understand outputs immediately usable and easier to
leaves room for misinterpretation. patterns and expected output quality. evaluate.

4. Ground in Context/Data 5. Iterate for Improvement

Provide relevant background information, constraints, and reference View prompting as an iterative process. Analyze outputs, identify gaps, and
materials. Context helps AI generate responses aligned with your specific refine your instructions progressively to achieve optimal results.
situation.
Introducing the T-C-R-E-I Framework
A systematic approach to crafting professional-grade prompts that consistently deliver high-quality results. Each component builds upon the previous one to create comprehensive, effective
instructions.

T — Task
Define the specific action you want the AI to perform

C — Context
Provide background, purpose, and situational information

R — Reference
Supply examples, style guides, or sample outputs

E — Evaluate
Review output quality, accuracy, and alignment with goals

I — Iterate
Refine your prompt based on evaluation to improve results

This framework transforms prompt writing from guesswork into a repeatable, professional process that scales across any AI application or use case.
Memory Hook
Thoughtfully
Approach each prompt with intention and careful consideration of your desired outcome

Create
Build your prompts systematically, incorporating all necessary elements for success

Really
Commit to excellence by investing effort into reference materials and examples

Excellent
Aim for outstanding results by critically evaluating outputs against high standards

Inputs
Remember that quality inputs directly determine the quality of your AI-generated outputs

Use this mnemonic device to recall the framework whenever you're crafting prompts. The more you practice, the more natural this approach becomes.
T-C-R-E-I Framework in Action
Here's how the framework transforms a vague request into a structured, effective prompt that delivers consistent, actionable results.

Task
1
Summarize customer reviews to identify key themes and sentiment patterns

Context
2
Analyzing 50 product reviews from the past month to inform product development decisions and marketing messaging

Reference
3 Follow this sample summary style: "Customers appreciate [positive themes]. Main concerns include [negative themes]. Suggestion: [action item]"

Evaluate
4
Check for clarity of themes, balanced representation of sentiment, and actionable structure that leadership can use immediately

Iterate
5
If output is too long, request "Shorten to 3 bullet points." If tone is off, ask to "Simplify language for non-technical stakeholders"

Notice how each element builds specificity, reducing ambiguity and increasing the likelihood of getting exactly what you need on the first attempt.
Effective Prompt Template
Use this reusable template as your starting point for any prompt. Fill in each section systematically to ensure comprehensive, well-structured instructions.

Task Context Reference

What specific action should the AI perform? Why is this needed? Who is the audience? What's the What examples, style guides, or standards should be
background? followed?
Example: "Create a product description""Analyze
survey data""Generate social media posts" Example: "For e-commerce website""Target Example: "Follow our brand voice guide""Similar to
audience: busy professionals""Launching new competitor X's approach""Use attached sample as
sustainable product line" template"

Constraints Format Evaluate & Iterate

What limitations apply? Length, tone, style, technical How should the output be structured? After generation, assess and refine
requirements?
Example: "Bullet points with headers""3 paragraphs Check: Accuracy, clarity, completenessRefine: Adjust
Example: "Maximum 150 words""Professional but with subheadings""Table with 4 columns" tone, length, structureRegenerate: With improved
friendly tone""Avoid technical jargon" prompt
Writing Effective Prompts
Master these six essential techniques to dramatically improve your prompt quality and AI output consistency. Each technique addresses a common weakness in amateur prompting.

Use Precise Verbs

Replace "write" with "summarize," "analyze," "compare," or "explain." Specific verbs set clear expectations for the type
of output you need.

Add Audience & Tone

Specify who will read this and what tone is appropriate. "For executives" differs dramatically from "for high school
students."

Add Constraints

Define length limits (word count, character count), style requirements (formal, casual), and technical level (beginner,
advanced).

Provide Examples

Show the AI what "good" looks like. Even a single example dramatically improves output quality and alignment with
expectations.

Break Tasks into Steps

For complex requests, decompose them into sequential steps. "First analyze, then summarize, finally recommend"
creates structure.

Request Structured Output

Ask for specific formats like numbered lists, tables with headers, or sections with headings. Structure makes outputs
immediately actionable.
Structured Output Examples
Structure transforms AI responses from walls of text into scannable, actionable formats. Choose the structure that best serves your purpose and audience.

Bullets Tables Steps

Perfect for quick scanning of key points, action items, or feature Ideal for comparing options, displaying data, or organizing Essential for processes, tutorials, or any sequential workflow
lists. Use when order doesn't matter. information across multiple dimensions. where order matters critically.

• First key point• Second insight• Final recommendation Feature | Plan A | Plan BPrice | $10 | $20Storage | 5GB 1. Prepare materials2. Execute procedure3. Verify results
| 50GB

Headings JSON
Best for longer content requiring clear sections, topics, or When you need machine-readable output for technical
thematic organization. integration, data processing, or API consumption.

## OverviewContent here...## DetailsMore content... {"name": "value", "status": "active", "count": 42}
Refining Outputs
Even the best first attempts can be improved. Use these refinement prompts as follow-ups to transform good outputs into excellent ones. Think of refinement as a conversation where you guide the AI toward perfection.

Length Adjustment

"Shorten to 50 words" or "Expand to 200 words with more examples"

Clarity Enhancement

"Improve clarity by simplifying complex sentences and defining technical terms"

Audience Recalibration

"Rewrite for beginners with no prior knowledge" or "Make more technical for experts"

Tone Modification

"Make tone professional and formal" or "Make more conversational and friendly"

Format Conversion

Why Refinement Matters "Convert to bullet points" or "Reformat as a comparison table with 3 columns"

Your initial prompt might produce 80% of what you need. Refinement closes that gap efficiently, teaching you what
adjustments produce which improvements. Over time, you'll need fewer iterations as you learn to anticipate requirements.
The Iteration Loop
Quality outputs rarely happen on the first try. Professional prompt engineers embrace iteration as the path to excellence. Each cycle teaches you something new about effective
prompting.

1. Draft 2. Generate
Create your initial prompt using the T-C-R-E-I framework Submit your prompt to the AI and receive the initial output
and best practices

6. Finalize 3. Evaluate
Once the output meets all requirements, save your Assess output quality, accuracy, completeness, tone, and
successful prompt for future use format against your goals

5. Regenerate 4. Improve
Apply your refinements and generate an improved version Identify specific gaps and craft targeted refinement
of the output prompts to address them

Pro Tip: Save your successful prompts in a personal library. Over time, you'll build a collection of proven templates that accelerate future work.
Prompt Troubleshooting
When AI outputs miss the mark, the problem usually lies in the prompt, not the AI. Use this diagnostic guide to identify and fix common issues quickly and systematically.

Add Missing Context

If output feels generic or misaligned, you probably omitted crucial background information. Add details about purpose, audience, or situation.

Increase Specificity
Vague verbs like "write about" produce vague outputs. Replace with precise verbs: "compare," "analyze," "summarize in 3 bullets."

Add Constraints
Without boundaries, AI makes assumptions. Specify length, tone, complexity level, format, and any content to avoid.

Include References/Examples
If output style misses expectations, show the AI what "good" looks like. A single example can dramatically improve alignment.

Break into Smaller Tasks

Complex requests overwhelm AI. Decompose into sequential steps: "First analyze X, then summarize Y, finally recommend Z."

Most prompt problems trace back to one of these five issues. Develop the habit of checking each element when outputs disappoint.
Hands-On Exercise 1: ChatGPT
Objective: Distinguish Good Prompts from Bad Prompts

Access ChatGPT
Open ChatGPT Free version at [Link]. No account required for basic access, though creating one enables saving history.

Enter This Prompt

"Give me 2 examples of unclear prompts and 2 examples of improved versions for any common daily task."

Analyze Results
Review the four prompt examples ChatGPT generates. Identify specific elements that make the improved versions more effective.

Discussion Points
• What specific words or phrases distinguish good from bad prompts?
• Which elements of T-C-R-E-I appear in the improved versions?
• How do constraints and format specifications improve clarity?

Time allocation: 5-7 minutes. If working in groups, compare results and discuss patterns you notice across different examples.
Hands-On Exercise 2: Gemini
Objective: Apply the T-C-R-E-I Framework

The Challenge Required Components

Vague prompt: "Write about online branding." Task

This prompt lacks every element of effective prompting. Your job is to Define the exact action (explain, compare, create, analyze)
transform it using T-C-R-E-I principles into a comprehensive, professional-
grade instruction. Context

Add purpose, audience, and background details

Reference

Include examples, style preferences, or format samples

Constraints

Specify length, tone, complexity level, format

Format

Define structure (bullets, paragraphs, sections, table)

Access Gemini Free at [Link]. Test your improved prompt and evaluate whether the output meets professional standards. Iterate if needed.
Hands-On Exercise 3: Perplexity
Objective: Experience Grounded, Source-Cited Output

Access Perplexity
1
Navigate to [Link] and access the free version. Perplexity's unique capability is real-time web search and source citation.

Enter This Prompt

2
"Search and summarize the top 3 challenges professionals face in adopting AI tools. Output as bullet points with brief explanations."

Observe Grounding
3
Notice how Perplexity provides numbered citations linking to source websites. Click several citations to verify information accuracy.

Discussion: Why Grounding Matters

• How does source citation increase trust and credibility?
4 • When is grounded output essential vs. nice-to-have?
• How might this capability influence business decisions?
• What types of prompts benefit most from real-time search?

Pro Insight: Grounded outputs are crucial for research, fact-checking, competitive analysis, and any scenario where accuracy and verifiability matter.
Hands-On Exercise 4: Refinement Practice
Objective: Master Output Refinement Across Any AI Tool

Choose any AI tool (ChatGPT, Gemini, or Perplexity) and select any output from the previous three exercises. Now practice the art of refinement through targeted follow-up
prompts.

Shorten Simplify Change Tone

Prompt: "Condense this to 50 words maximum while Prompt: "Rewrite using simpler language suitable for Prompt: "Adjust tone to be more
preserving key points." someone with no background in this topic." [formal/casual/enthusiastic/neutral]."

Notice what gets eliminated and what remains See how word choice shapes perception.
essential. Observe how complexity affects accessibility.

Restructure Professionalize
Prompt: "Convert this to [bullet points/a comparison table/numbered steps]." Prompt: "Make this more professional and suitable for executive presentation."

Experience how format changes usability. Learn what "professional" means in practice.

Reflection: Which refinement technique had the biggest impact on output quality? How will you apply these refinement strategies in your daily work?