POML PROMPTS – USING DETAILED PROMPT CARD STRUCTURE

Most of us start with free-form prompts:

“Write me a sales letter about a non-stick pan.”
“Make a Facebook post about healthy eating.”

Sometimes they work, but often they don’t. The AI’s responses are inconsistent — the tone changes, details shift, and you end up spending more time fixing than creating.

“Free text prompts are prone to errors, misinterpretation, or inconsistencies.”

The issue boils down to lack of structure. When AI doesn’t have clear rules, you’ll get results that look different every time. And if you’re building a business, consistency is key.

Enter Prompt Cards: The Game-Changer

Prompt cards are structured prompts written in a coding-like format. Instead of loosely telling AI what you want, you break it down into rules, roles, and tasks.

The two most powerful formats are:

POML (Prompt Orchestration Markup Language)
- Works like HTML for prompts.
- Easy to read and understand.
- Great for templates and reusable content.
JSON (JavaScript Object Notation)
- A data format computers love.
- Perfect for complex rules, lists, or detailed conditions.
- Best when you need very precise output.

“Use POML if you want something simple and human-readable. Use JSON if you want the computer to follow very specific, detailed rules.”

Why Structured Prompts Work Better

Think of AI as a new employee. If you give vague instructions, the results vary. If you hand them a checklist with clear steps, you’ll always get the right outcome.

Structured prompts help in several ways:

Consistency: Every blog, video, or ad looks and feels the same.
Reusability: Build once, use forever. Just change the variables (like product name).
Scalability: Generate 50 posts, 10 ads, or 6 videos — all with the same style.
Efficiency: No more editing line by line. AI “knows the rules” from the start.

The speaker demonstrated this with non-stick pan ads. Normally, AI would generate random pan images — different shapes, colors, and looks. But with prompt cards, every image had the same pan, background, and branding. That’s what makes it usable for real marketing.

Breaking Down a POML Prompt

Here’s what a basic POML card might look like (simplified from the training):

<role> Direct Response Copywriter </role>

<task> Write a long-form sales letter for a non-stick pan </task>

<context> Make it engaging, funny, and benefit-driven </context>

<output> 1000 words, short paragraphs, clear call-to-action </output>

Instead of cramming everything into one messy sentence, you’re telling AI exactly:

Who it should be (role).
What to do (task).
How to say it (context).
What the final product should look like (output).

“It’s pretty much like fill in the blanks. Build once, tweak variables, and grow as you go.”

Breaking Down a JSON Prompt

Now let’s look at JSON, which is more technical but very powerful.

{

“role”: “Content Creator”,

“task”: “Generate 5 short social media posts”,

“audience”: “Entrepreneurs”,

“tone”: “Bold, punchy, and easy to digest”,

“format”: {

“length”: “2 sentences”,

“cta”: “Save this post for later”

}

With JSON, you can include detailed rules like:

Audience (entrepreneurs, students, fitness lovers).
Tone & Style (bold, funny, formal, emotional).
Format specifics (word count, structure, CTA).

This ensures every post comes out in the same style — without you rewriting the instructions.

The Building Blocks of Great Prompt Cards

From the training, here are the essential elements every prompt card should include:

Element	Purpose	Example
Role	Who AI should act as	Direct response copywriter
Task	What AI should do	Write a sales letter
Context	Style, background, or theme	Fun, engaging, story-driven
Audience	Who it’s for	Beginners, entrepreneurs
Tone/Voice	How it should sound	Empathetic, bold, confident
Output Format	End result details	1000 words, bullet points
Constraints	What to avoid or limit	No keyword stuffing, no jargon
Examples	Samples to mimic	Famous ad copy, top blog posts

This is why the speaker called prompt cards “boring but powerful.” They may not look exciting, but they build the foundation for consistent content that scales.

Real-World Use Cases for Prompt Cards

Here’s how businesses can apply these prompt structures:

Social Media Posts
- Create a JSON card for 30 Facebook captions.
- Each one will follow the same voice, call-to-action, and structure.
Sales Letters & Ads
- Build a POML card once with the formula (headline, benefits, proof, CTA).
- Reuse for any product by changing the product name.
Email Marketing
- Ensure every email matches your brand tone.
- Add rules like: “short subject lines, curiosity-driven, under 50 characters.”
YouTube Scripts or Video Carousels
- Consistent characters, story flow, and style across multiple videos.
Tools & Plugins
- The speaker even built a plugin using prompt cards. Because all tools shared the same formatting, it took weeks instead of months.

Step-by-Step: How to Create Your Own Prompt Cards

Here’s a simple roadmap to start:

Step 1: Train the AI with a Normal Prompt

Example: “Tell me about the best sales letters of all time and who wrote them.”
This gathers the raw data.

Step 2: Extract the Structure

Look at what AI gave you and break it into building blocks: headline, story, benefits, proof, CTA.

Step 3: Turn It into a Prompt Card

Convert that into POML or JSON format with roles, tasks, and context.

Step 4: Test and Refine

Run the prompt. See if it gives you consistent output. If not, adjust rules.

Step 5: Reuse and Scale

Save it as a “card.” Now you can apply it to:

New products.
Different audiences.
Multiple content formats.

Pro Tips from Marcus

Throughout the training, some standout tips were:

“When something works, keep doing it until it stops working.”
“Consistency is key. Don’t make one page of your website look completely different from another.”
“Add rules like ‘never do this.’ That keeps output clean across projects.”
“Most AI tools you pay for are just glorified coded prompts. You can make your own.”

The Bigger Picture: Why This Matters for Business

AI isn’t just about speed; it’s about building a repeatable system. By turning messy instructions into structured prompt cards, you’re essentially:

Building templates for growth.
Training AI once and reusing it forever.
Making your brand voice consistent across every platform.

“Making money with AI is a business. Most people trying to make money make nothing. But if you understand prompts, you can build something consistent, scalable, and valuable.”

Affiliate Course Download

Start Your First Website Now

Find Your Profit Niche Here

ULTIMATE POML/JSON PROMPT COMMANDS REFERENCE

Prompting Reference Manual

1. Core Tags

<role> – Define AI identity (teacher, marketer, analyst, developer)
<task> – Define primary action (explain, write, analyze, create)
<context> – Provide background information and situational details
<output-format> – Control response structure (list, JSON, essay, table)

Advanced Structure
<objective> – Clear goal statement
<scope> – Define boundaries of the task
<priority> – Set importance levels for different aspects
<methodology> – Specify approach or framework to use

2. Audience & Demographics

Target Audience
<audience> – Primary audience (kids, professionals, seniors, students)
<persona> – Write as specific character (“as Steve Jobs”, “as a kindergarten teacher”)
<reading-level> – Complexity level (grade 4, high school, PhD, expert)
<region> – Geographic localization (US, UK, Japan, global)
<culture> – Cultural considerations and sensitivities

User Context
<expertise-level> – User’s knowledge (beginner, intermediate, expert)
<time-available> – How much time user has (quick read, deep dive)
<device> – Platform considerations (mobile, desktop, print)
<accessibility> – Special needs accommodations

3. Tone & Style

Voice & Tone
<tone> – Overall feeling (friendly, professional, casual, authoritative)
<voice> – Perspective (1st person, 2nd person, 3rd person)
<style> – Writing approach (academic, conversational, technical, poetic)
<emotion> – Emotional quality (enthusiastic, calm, urgent, empathetic)
<formality> – Level of formality (casual, business, formal, academic)

Communication Style
<humor> – Include appropriate humor or wit
<energy> – Energy level (high-energy, moderate, calm)
<confidence> – Certainty level (confident, cautious, exploratory)
<empathy> – Emotional connection level

4. Structure & Organization

Content Structure
<sections> – Main divisions (intro, body, conclusion)
<headings> – Heading hierarchy (H1, H2, H3 structure)
<steps> – Sequential organization (step-by-step, numbered)
<framework> – Organizational pattern (AIDA, PAS, Hero’s Journey, 5W1H)

Flow Control
<transitions> – How sections connect
<pacing> – Information delivery speed
<emphasis> – What to highlight or stress
<sequence> – Order of information presentation

5. Constraints & Limits

Length Controls
<length> – Word/character limits (280 chars, 500 words, 2 pages)
<brevity> – Conciseness requirements
<depth> – Level of detail required
<coverage> – Scope of topics to include

Content Constraints
<constraints> – Specific limitations or requirements
<do> – Required elements or approaches
<dont> – Things to avoid or exclude
<budget> – Cost considerations for recommendations
<time-limit> – Deadline constraints

Quality Controls
<accuracy> – Fact-checking requirements
<citations> – Source attribution needs
<verification> – Evidence requirements

6. Data & Sources

Variable Management
<variables> – Placeholder values ([TOPIC], [PRODUCT], [NAME])
<parameters> – Configurable elements
<placeholders> – Template fields to fill

Content Sources
<keywords> – SEO terms and important phrases
<references> – Links, quotes, and source materials
<facts> – Statistics, dates, and verified information
<examples> – Specific instances or case studies
<data-sources> – Where information comes from

Research Elements
<research-depth> – How thorough to be
<fact-checking> – Verification requirements
<currency> – How recent information should be

7. Creativity & Rhetoric

Literary Devices
<analogy> – Comparative explanations
<metaphor> – Figurative language
<story> – Narrative elements
<anecdote> – Personal or illustrative stories

Engagement Techniques
<dialogue> – Conversational format
<hooks> – Attention-grabbing openings
<titles> – Headline variations and options
<cta> – Call-to-action elements
<questions> – Rhetorical or engaging questions

Creative Elements
<imagery> – Vivid descriptions
<symbolism> – Symbolic representations
<wordplay> – Puns, alliteration, clever language
<rhythm> – Flow and cadence of text

8. Output Types

Structured Formats
<json-output> – JSON structure specification
<yaml-output> – YAML format requirements
<xml-output> – XML structure needs
<csv-output> – Comma-separated values format

Document Types
<table> – Tabular data (markdown or HTML)
<list> – Bulleted or numbered lists
<outline> – Hierarchical structure
<summary> – Condensed version

Code Formats
<code> – Programming language (Python, JavaScript, HTML)
<pseudo-code> – Algorithm descriptions
<markup> – HTML, Markdown, or other markup
<template> – Reusable format structures

Presentation Formats
<slides> – Presentation outline or content
<script> – Speaking or performance text
<agenda> – Meeting or event structure

9. Multimedia & Multi-Mode

Video Content
<video> – Scene-by-scene descriptions
<storyboard> – Visual sequence planning
<camera> – Shot types (zoom, pan, cuts)
<timing> – Duration and pacing

Audio Content
<audio> – Sound design elements
<narration> – Voice-over scripts
<music> – Background audio suggestions
<sound-effects> – Audio enhancement ideas

Visual Elements
<visuals> – Image and graphic descriptions
<infographics> – Data visualization ideas
<charts> – Graph and chart specifications
<diagrams> – Technical illustrations

Interactive Elements
<interactive> – User engagement features
<quiz> – Question and answer formats
<poll> – Survey elements
<game> – Gamification aspects

10. Advanced Logic

Reasoning Operations
<logic> – Step-by-step reasoning chains
<analysis> – Analytical breakdowns
synthesis – Combining information
<deduction> – Logical conclusions

Comparison Tools
<compare> – Side-by-side analysis
<contrast> – Difference highlighting
<pros-cons> – Advantage/disadvantage lists
<trade-offs> – Decision analysis

Evaluation Methods
<evaluate> – Assessment criteria
<critique> – Critical analysis
<score> – Rating systems
<rank> – Ordering by criteria

Transformation Operations
<translate> – Language conversion
<style-transfer> – Adopt another’s style (Hemingway, Shakespeare)
<format-convert> – Change between formats
<simplify> – Make more accessible

11. Meta & Utility

AI Instructions
<instructions> – Meta-rules for AI behavior
<system> – System-level commands
<behavior> – AI personality adjustments
<guardrails> – Safety and ethical boundaries

Content Management
<tags> – Categories and labels
<metadata> – Information about the content
<version> – Iteration tracking
<status> – Completion state

Process Control
<iteration> – Multiple draft requests
<refinement> – Improvement instructions
<feedback> – Self-critique and improvement
<validation> – Quality checking steps

Workflow Management
<dependencies> – What needs to come first
<parallel> – Tasks that can be done simultaneously
<sequence> – Required order of operations
<checkpoints> – Review and approval points

12. Advanced Specialized Tags

SEO & Marketing
<seo-keywords> – Search optimization terms
<meta-description> – Page descriptions
<alt-text> – Image descriptions
<social-media> – Platform-specific adaptations

Technical Writing
<api-docs> – API documentation format
<user-manual> – Instructional content
<troubleshooting> – Problem-solving guides
<changelog> – Version update descriptions

Educational Content
<learning-objectives> – What students should achieve
<assessment> – Testing and evaluation methods
<scaffolding> – Progressive skill building
<differentiation> – Multiple learning approaches

Business Communication
<executive-summary> – High-level overview
<action-items> – Next steps and responsibilities
<stakeholders> – Audience considerations
<roi> – Return on investment focus