Search anything...
K
Back to Docs
  • Introduction
  • Quick Start
  • Account Setup
  • AI Studio
  • Chat
  • Agents
  • Voice
  • MCP Servers
  • Workflows
  • Authentication
  • Studio API
  • Chat API
  • Agents API
  • Voice API
  • Workflows API
  • Webhooks
  • Error Codes
  • Creating Custom Agents
  • MCP Integration
  • Building Workflows
  • Prompt Engineering
  • Team Management
  • Billing & Plans
  • Usage Monitoring
  • Single-Tenant Cloud
  • Private VPC Deployment
  • SSO Configuration
  • Security Policies
  • Compliance
  • Troubleshooting
  • API Versioning
DocsGuidesPrompt Engineering

Prompt Engineering

Master the art of writing effective prompts for better AI results

This guide teaches you how to write effective prompts to get the best results from Girard's AI features. Learn techniques for image generation, chat, and agents.

Prompt Fundamentals

What Makes a Good Prompt?

A good prompt is:

  • Specific - Clear about what you want
  • Detailed - Includes relevant information
  • Structured - Organized logically
  • Contextual - Provides necessary background

The Prompt Framework

[Context] + [Task] + [Format] + [Constraints]

Example:

You are a senior software engineer. [Context]
Review this code for security vulnerabilities. [Task]
List each issue with severity and fix. [Format]
Focus on OWASP Top 10 only. [Constraints]

Image Generation Prompts

Basic Structure

[Subject] + [Style] + [Lighting] + [Composition] + [Quality]

Subject Description

Be specific about your subject:

VagueSpecific
"a cat""a fluffy orange tabby cat with green eyes"
"a mountain""a snow-capped mountain peak at sunrise"
"a person""a young woman with curly brown hair wearing a red dress"

Style Modifiers

StyleExample Prompt Addition
Photorealistic"photorealistic, 8k, detailed"
Illustration"digital illustration, vibrant colors"
Oil Painting"oil painting style, brushstrokes visible"
Anime"anime style, Studio Ghibli inspired"
Cinematic"cinematic, movie still, dramatic lighting"
Vintage"vintage photograph, 1970s aesthetic"

Lighting and Mood

LightingDescription
Golden Hour"warm golden hour lighting, soft shadows"
Dramatic"dramatic lighting, high contrast, moody"
Studio"professional studio lighting, softbox"
Natural"natural daylight, soft diffused light"
Neon"neon lights, cyberpunk atmosphere"
Backlit"backlit, silhouette, rim lighting"

Composition Terms

TermEffect
Rule of thirdsBalanced, professional composition
Close-upFocus on details
Wide shotEnvironmental context
Low angleDramatic, powerful
Bird's eye viewOverview perspective
Shallow depth of fieldBlurred background

Quality Enhancers

Add these for better results:

highly detailed, professional photography, 8k resolution,
award-winning, masterpiece, sharp focus, intricate details

Complete Image Prompt Examples

Portrait:

Portrait of an elderly Japanese man with weathered skin and kind eyes,
wearing traditional indigo-dyed clothing, soft natural lighting from
a window, shallow depth of field, professional photography, 8k,
highly detailed, warm color palette

Landscape:

Majestic fjord in Norway during autumn, dramatic cliffs reflecting
in perfectly still water, golden hour lighting, low hanging mist,
vibrant fall colors, wide angle composition, landscape photography,
ultra detailed, 16:9 aspect ratio

Product:

Minimalist product photography of a luxury watch on a marble surface,
professional studio lighting, soft shadows, clean white background,
commercial photography, sharp focus, high-end advertising style

Fantasy:

Ancient library hidden inside a giant tree, magical glowing books
floating in the air, warm candlelight, intricate wooden architecture,
fantasy art style, highly detailed, ethereal atmosphere, cinematic
composition, digital painting

Negative Prompts

Tell the AI what to avoid:

Negative prompt: blurry, low quality, distorted faces, extra limbs,
watermark, text, logo, oversaturated, grainy, pixelated, cropped,
out of frame, bad anatomy, deformed

Chat and Agent Prompts

Conversation Starters

For Information:

Explain [topic] in simple terms, using analogies a
beginner would understand.

For Analysis:

Analyze the pros and cons of [subject]. Present your
analysis in a structured format with clear categories.

For Creation:

Write a [type of content] about [topic] that is
[tone/style]. Include [specific elements].

Structured Requests

Use clear structure for complex requests:

I need help with [task].

Context:
- [Relevant background 1]
- [Relevant background 2]

Requirements:
1. [Requirement 1]
2. [Requirement 2]
3. [Requirement 3]

Please format the output as [format description].

Examples with Techniques

Chain-of-Thought:

Solve this problem step by step, showing your reasoning
at each stage:

[Problem description]

Think through each step carefully before providing the
final answer.

Role Assignment:

Act as an experienced [role] with expertise in [domain].
From that perspective, [task].

Few-Shot Learning:

Format customer feedback like these examples:

Example 1:
Input: "Great product but shipping was slow"
Output: {"sentiment": "mixed", "topics": ["product", "shipping"]}

Example 2:
Input: "Terrible experience, never buying again"
Output: {"sentiment": "negative", "topics": ["experience"]}

Now format this feedback:
"Love the quality, arrived quickly, will recommend"

Constraints:

Write a product description with these constraints:
- Maximum 150 words
- Include exactly 3 bullet points
- End with a call to action
- Use active voice only
- Reading level: 8th grade

System Prompts for Agents

Structure

# Role Definition
[Who is this agent?]

## Expertise Areas
- [Area 1]
- [Area 2]

## Communication Style
- [Style trait 1]
- [Style trait 2]

## Guidelines
- [Guideline 1]
- [Guideline 2]

## Output Format
[How should responses be formatted?]

## Constraints
- [What NOT to do]

Examples by Agent Type

Technical Writer:

You are a technical writer who creates clear, accurate documentation.

## Writing Style
- Clear and concise
- Jargon-free when possible
- Step-by-step instructions
- Examples for every concept

## Document Structure
Always include:
1. Overview/Purpose
2. Prerequisites
3. Step-by-step instructions
4. Examples
5. Troubleshooting

## Guidelines
- Assume minimal prior knowledge
- Use numbered lists for procedures
- Include screenshots/code when relevant
- Define technical terms on first use

Data Analyst:

You are a data analyst who helps interpret data and generate insights.

## Expertise
- Statistical analysis
- Data visualization
- SQL queries
- Business intelligence

## Analysis Approach
1. Clarify the question
2. Identify relevant data
3. Perform analysis
4. Generate insights
5. Recommend actions

## Output Format
Present findings with:
- Executive summary (2-3 sentences)
- Key metrics in tables
- Visualizations when appropriate
- Actionable recommendations

Creative Director:

You are a creative director with expertise in branding and marketing.

## Expertise
- Brand strategy
- Visual identity
- Campaign concepts
- Creative copywriting

## Creative Process
1. Understand the brief
2. Research the audience
3. Develop concepts
4. Present options
5. Refine based on feedback

## Style
- Bold and innovative
- Trends-aware but timeless
- Always audience-focused
- Balance creativity with strategy

Advanced Techniques

Prompt Chaining

Break complex tasks into steps:

Step 1: "Analyze this customer feedback and list the main themes"

Step 2: "For each theme you identified, suggest improvements"

Step 3: "Prioritize these improvements by impact and effort"

Step 4: "Create an action plan for the top 3 priorities"

Meta-Prompting

Ask the AI to improve your prompt:

I want to generate images of futuristic cities.
Help me write a detailed prompt that would produce
high-quality results. Include style, lighting,
composition, and quality terms.

Iterative Refinement

Build on previous outputs:

Initial: "Write a tagline for a fitness app"
Output: "Get Fit, Stay Strong"

Refinement: "Make it more energetic and add urgency"
Output: "Transform Today. No Excuses."

Further: "Add alliteration"
Output: "Push Past Possible. Progress Daily."

Temperature Guidelines

TemperatureUse CaseExample
0.0 - 0.3Factual, consistentCode review, data analysis
0.4 - 0.6BalancedGeneral assistance
0.7 - 0.8Creative, variedMarketing copy
0.9 - 1.0Highly creativeBrainstorming, fiction

Common Mistakes to Avoid

Image Generation

MistakeBetter Approach
"a nice picture"Specify style, subject, mood
Too many subjectsFocus on one main subject
Conflicting stylesChoose one consistent style
No quality termsAdd "detailed, high quality"

Chat/Agents

MistakeBetter Approach
Vague requestsBe specific about what you need
No contextProvide relevant background
Too many tasksOne task at a time
No format specifiedDescribe desired output format

Prompt Templates

Image Generation

[Subject with specific details], [art style/medium],
[lighting conditions], [mood/atmosphere], [composition],
[quality modifiers like "highly detailed, 8k"]

Negative: [things to avoid]

Content Creation

Write a [content type] about [topic].

Audience: [target audience]
Tone: [desired tone]
Length: [word count or format]
Include: [specific elements]
Avoid: [things not to include]

Analysis Request

Analyze [subject/data] and provide:
1. Summary of key findings
2. Detailed breakdown by [category]
3. Recommendations based on findings

Context: [relevant background]
Format: [table/list/narrative]

Code Assistance

[Task description]

Requirements:
- Language: [programming language]
- Include: [specific features]
- Style: [coding style preferences]

Please include:
- Comments explaining the logic
- Error handling
- Example usage

Practice Exercises

Exercise 1: Image Prompt

Transform this basic prompt into a detailed one:

Basic: "a dog in a park"

Your improved version: _______________

Exercise 2: System Prompt

Create a system prompt for a customer service agent:

Role: Customer service for a SaaS company
Tone: Helpful but professional
Capabilities: Answer questions, troubleshoot, escalate

Your system prompt: _______________

Exercise 3: Task Request

Improve this request:

Basic: "Write about climate change"

Your improved version: _______________

Previous: Building Workflows | Next: Team Management

Previous
Building Workflows
Next
Team Management