AI Prompt Writing for Marketing Teams — Training Guide
<!-- dual-compat:start -->
Use when
- Produces a practical training guide for client teams on prompt engineering for marketing tasks — covering the Alpha-Beta-Gamma-Delta-Epsilon prompt structure, 10 prompt components, 5 prompting approaches, and 7 copywriting frameworks with worked East African examples. Invoke when the user says "create a prompt writing training guide", "teach my team how to use AI for marketing", "write a prompt engineering workshop", "AI copywriting training for staff", or needs a structured training document for client employees who use AI tools to produce marketing content.
- Use this skill when it is the closest match to the requested deliverable or workflow.
Do not use when
- Do not use this skill for graphic design, video production, software development, or legal advice beyond the repository's stated scope.
- Do not use it when another skill in this repository is clearly more specific to the requested deliverable.
Workflow
- Collect the required inputs or source material before drafting, unless this skill explicitly generates the intake itself.
- Follow the section order and decision rules in this ; do not skip mandatory steps or required fields.
- Review the draft against the quality criteria, then deliver the final output in markdown unless the skill specifies another format.
Anti-Patterns
- Do not invent client facts, performance data, budgets, or approvals that were not provided or clearly inferred from evidence.
- Do not skip required inputs, mandatory sections, or quality checks just to make the output shorter.
- Do not drift into out-of-scope work such as code implementation, design production, or unsupported legal conclusions.
Outputs
- A structured markdown document, plan, playbook, or strategy ready for client-facing or internal use.
References
- Use the inline instructions in this skill now. If a directory is added later, treat its files as the deeper source material and keep this execution-focused.
<!-- dual-compat:end -->
How to Use This Skill
Collect the Required Input below. Then generate the full training guide in four modules, substituting all bracketed placeholders with the client's specific details. Output is a complete, facilitator-ready training document — not a slide deck. For a slide deck version, use the
prefix convention and build slides separately.
Required Input
Ask for the following before generating the training guide:
- Client business name and industry — trading name and sector (e.g. telecoms, microfinance, FMCG, hospitality)
- Country / city — default Uganda / East Africa
- Primary goal — what the client wants the team to achieve after training
- Team size and prior AI experience level — number of participants and experience (none / basic / intermediate)
- Primary content types produced — captions, emails, blogs, WhatsApp broadcasts, SMS, ad copy, etc.
- Preferred AI tools — ChatGPT, Gemini, Claude, or other (specify versions where known)
- Training format — in-person half-day / virtual session / self-guided handout
Output: Complete Training Guide
Generate the following four modules in full. Use the client's name, industry, preferred AI tools, and primary content types throughout. Write in plain English — no jargon. Tone: practical, encouraging, professional.
Training Overview
Programme: AI Prompt Writing for Marketing Teams
Total Duration: Approximately 2.5 hours (150 minutes)
Audience: Marketing, communications, and content staff with any level of AI experience
Format: [Insert training format]
Prepared for: [Client Business Name]
Industry: [Industry]
Primary Sources: Upadhyay, M.A. (2024) Generative AI for Marketing (Packt); Anderson, D. (2022) AI in Digital Marketing Training Guide (Self-published)
Module 1: Why Prompt Quality Matters (30 minutes)
Learning Objective
Explain why the quality of AI output is directly determined by the quality of the prompt — and establish the principle that good prompts are a learnable, professional skill.
1.1 What AI Can and Cannot Do
Explain the following to participants:
AI tools such as ChatGPT, Gemini, and Claude are language prediction engines. They produce statistically probable text based on instructions. They do not:
- Know your brand unless you tell them
- Understand your audience unless you describe them
- Know what to avoid unless you specify constraints
- Default to quality — they default to plausible
AI tools excel at:
- Drafting and editing at speed
- Generating multiple variations quickly
- Applying a named copywriting framework consistently
- Adapting tone, length, and format on instruction
1.2 The GIGO Principle
Apply the GIGO principle (Garbage In, Garbage Out — Anderson, 2022) as the session's opening argument: vague instructions produce vague output. Demonstrate this directly.
Live demonstration — run this exercise:
Ask the AI tool: "Write me a caption."
Show the output to participants. Ask: "Would you post this?"
Then ask: "You are a social media copywriter for a Ugandan mobile money brand targeting market traders aged 30–50 in Kampala. Write one Instagram caption using the PAS framework (Problem → Agitate → Solution) for a new zero-fee transfer promotion. Under 150 characters. End with a call to action. No exclamation marks."
Show the second output. Discuss the quality gap.
1.3 The Quality Gap
Define the quality gap as the difference between raw AI output and human-quality output ready for publishing. The gap is closed by:
- Providing complete context (Beta element)
- Assigning an expert role (Alpha element)
- Specifying the exact task and format (Gamma and Epsilon elements)
- Restricting what to avoid (Delta element)
1.4 Discussion Prompt
Ask participants: "What is the worst AI output you have seen or produced? What do you think was missing from the prompt?"
Module 2: The Alpha-Beta-Gamma-Delta-Epsilon Framework (45 minutes)
Based on Upadhyay (2024), Generative AI for Marketing, Chapter 3.
Learning Objective
Teach participants to construct a complete, high-quality marketing prompt using the five-element Alpha-Beta-Gamma-Delta-Epsilon structure.
2.1 The Five Elements
Every high-quality marketing prompt contains five components. Walk through each one.
Alpha — Role Assignment
Assign an expert persona to the AI. The role anchors the AI's output register, expertise level, and perspective.
Formula: "You are a [role] who specialises in [specialism] for [market/industry]."
Example: "You are a senior copywriter who specialises in social media content for Ugandan financial services brands."
Why it matters: Without a role, the AI defaults to a generic, cautious register. With a strong role, output is confident, specific, and appropriate to the industry.
Beta — Context Block
Provide all relevant background. Include brand voice, target audience, platform, campaign background, and any specific product or offer details.
Formula: Brand name + audience demographics + platform + tone/voice + product or offer details.
Example: "The brand is [X], targeting urban professionals aged 25–35 in Kampala. Brand voice: confident, friendly, slightly witty. Product: 5G data bundle, UGX 15,000 for 5GB/7 days."
Why it matters: Thin context produces generic output. The more specific the Beta, the more tailored the result.
Gamma — Task Instruction
State exactly what to produce. Include format, length, quantity, perspective, and any framework to apply.
Formula: "Write [quantity] [format] using [framework]. Each must be [length constraint] and include [required elements]."
Example: "Write 3 Instagram captions using the PAS framework (Problem → Agitate → Solution). Each caption must be under 150 characters, end with a clear CTA, and include 2–3 relevant hashtags."
Why it matters: Without a clear task instruction, the AI interprets ambiguity in its own favour — often producing the easiest, most generic interpretation.
Delta — Constraints
State explicitly what to avoid: banned words, competitor mentions, topics, tone characteristics, or content types.
Formula: "Do not [X]. Do not use the word/phrase [Y]. Avoid [Z]."
Example: "Do not mention competitor brands. Do not use the words 'affordable' or 'seamless'. Avoid exclamation marks."
Why it matters: AI tools default to common marketing clichés. Constraints remove the most predictable and weakest outputs.
Epsilon — Output Format
Specify exactly how to present the result: numbered list, table, plain paragraph, JSON, labelled sections, etc.
Formula: "Present as [format]. Label each [element] with [label]."
Example: "Present as a numbered list with the PAS label for each element."
Why it matters: Formatting instructions make output paste-ready and prevent the AI from burying the useful content inside paragraphs of explanation.
2.2 The 10 Prompt Components
Expand the framework into the full 10 components that can appear in a complete marketing prompt (Upadhyay, 2024):
| # | Component | Description |
|---|
| 1 | Role assignment | Expert persona for the AI (Alpha) |
| 2 | Audience description | Demographics, psychographics, location |
| 3 | Platform / channel | Instagram, WhatsApp, email, LinkedIn, etc. |
| 4 | Objective | Awareness / engagement / conversion / retention |
| 5 | Tone instruction | Formal, friendly, witty, urgent, warm, etc. |
| 6 | Length / format | Character count, word count, list vs. paragraph |
| 7 | Constraints | What to avoid — words, topics, brands |
| 8 | Context block | Brand, product, campaign background (Beta) |
| 9 | Few-shot examples | 2–3 examples of the desired output style |
| 10 | Output structure | How to present the final result (Epsilon) |
Explain that not every prompt requires all 10 components. Simple tasks may need only 4–5. Complex, high-stakes content (campaign launches, CEO communications) should use all 10.
2.3 Common Mistakes
Walk through the four most frequent prompt errors:
- Role missing — output lacks expertise and authority; reads like generic text
- Context too thin — output is generic, uses placeholder language, misses brand voice
- No constraints — output is full of clichés: "seamless", "affordable", "game-changing"
- No format instruction — output is buried in paragraphs; hard to use without reformatting
2.4 Hands-On Activity (15 minutes)
Ask participants to build a complete Alpha-Beta-Gamma-Delta-Epsilon prompt for their own brand. Provide a blank template:
Alpha (Role): You are a ___________________________________
Beta (Context): The brand is _____, targeting _____ aged _____ in _____. Brand voice: _____. Product/offer: _____.
Gamma (Task): Write _____ using the _____ framework. Each must be _____ and include _____.
Delta (Constraints): Do not _____. Do not use the word "_____". Avoid _____.
Epsilon (Output format): Present as _____.
Participants submit their prompt to the AI tool and share their output with the group.
Module 3: Copywriting Frameworks in Prompts (45 minutes)
Learning Objective
Teach participants to embed copywriting frameworks in the Gamma element of a prompt, producing structured, persuasive marketing copy with consistent logic.
3.1 Why Frameworks Matter in Prompts
Explain that copywriting frameworks give the AI a logical structure to follow. Without a framework instruction, AI copy tends to be flat and unconvincing. With a framework, output follows a proven persuasion sequence.
3.2 The Seven Frameworks
Explain each framework with a Uganda/East Africa brand example. Show participants how to embed the framework name in the Gamma element.
1. PAS — Problem / Agitate / Solution
Structure: Identify the pain point → make it emotionally vivid → present the answer.
Prompt instruction (Gamma): "Write using the PAS framework: first name the problem the audience faces, then agitate it (make it feel urgent or frustrating), then present the product as the solution."
EA example: For a Kampala fuel delivery brand — Problem: running out of petrol at the wrong time. Agitate: stuck in Ntinda traffic with an empty tank, no fuel station in sight. Solution: [Brand] delivers fuel to your location in under 30 minutes.
2. AIDA — Attention / Interest / Desire / Action
Structure: Grab attention → build interest → create desire → prompt action.
Prompt instruction (Gamma): "Write using the AIDA framework: open with a bold attention hook, develop interest with relevant detail, build desire with a benefit or outcome, close with a direct call to action."
EA example: For a Ugandan savings app — Attention: "Every week you delay saving costs you more than you think." Interest: [App] automatically saves UGX 5,000 from every mobile money transaction. Desire: Thousands of Ugandans have saved their first UGX 1,000,000 without thinking about it. Action: Download and start saving today.
3. BAB — Before / After / Bridge
Structure: Show the painful before state → paint the desirable after state → bridge with the product.
Prompt instruction (Gamma): "Write using the BAB framework: describe the before state (audience's current frustration), describe the after state (what life looks like with the solution), then bridge the two with the product."
EA example: For a Nairobi laundry service — Before: spending your Sunday washing instead of resting. After: clean, folded, delivered by Saturday evening. Bridge: [Brand] handles your laundry so your Sunday is yours again.
4. FAB — Features / Advantages / Benefits
Structure: State the product feature → explain its advantage → describe the real benefit to the customer.
Prompt instruction (Gamma): "Write using the FAB framework: name the feature, state the functional advantage it provides, then articulate the emotional or practical benefit the customer experiences."
EA example: For a Ugandan solar home system — Feature: 200W solar panel with 10-hour battery backup. Advantage: powers lights, phone charging, and a small TV without grid electricity. Benefit: your household stays connected and your children can study after dark.
5. SSS — Star / Story / Solution
Structure: Introduce a relatable character (Star) → tell their struggle (Story) → resolve it with the product (Solution).
Prompt instruction (Gamma): "Write using the SSS framework: introduce a relatable character facing a problem, tell a short story of their struggle, then resolve it with the product or service."
EA example: For a Tanzanian health insurance brand — Star: Mama Zawadi, a market vendor in Dar es Salaam. Story: when her son fell ill, she had to close her stall for three days and borrow money she couldn't afford to repay. Solution: [Brand] covered the hospital bill. She never closed her stall again.
6. PPPP — Picture / Promise / Prove / Push
Structure: Create a vivid mental image → make a clear promise → provide evidence → push to action.
Prompt instruction (Gamma): "Write using the PPPP framework: paint a vivid picture of the desired outcome, make a specific promise, provide proof or social evidence, then push the reader to act now."
EA example: For a Ugandan real estate brand — Picture: imagine driving through your gate, not your landlord's. Promise: we help first-time buyers own a home in Wakiso from UGX 120M. Prove: over 400 families have moved in since 2021. Push: book your site visit this weekend — plots are releasing fast.
7. AFOREST — Alliteration / Facts / Opinions / Repetition / Examples / Rhetorical questions / Statistics / Three-part lists
Structure: A toolkit of persuasive writing devices — use selectively rather than all at once.
Prompt instruction (Gamma): "Write using AFOREST persuasive devices. Include at least: one factual statement, one rhetorical question, one statistic, and a three-part list."
EA example instruction to AI: "You are a copywriter for a Kenyan mobile data brand. Write a Facebook post using AFOREST devices: include a rhetorical question, a statistic about mobile data usage in Kenya, and a three-part list of what customers can do with the data bundle."
3.3 Embedding Frameworks in the Gamma Element
Show participants the exact pattern for embedding any framework in a prompt:
"Write [quantity] [content type] using the [FRAMEWORK NAME] framework ([brief description of the framework sequence]). [Add length, CTA, and hashtag requirements here.]"
3.4 Hands-On Activity (20 minutes)
Participants complete two exercises:
Exercise A: Write one Instagram caption using the PAS framework for their own brand. Run it through the AI, review the output, and refine once.
Exercise B: Write one email subject line using the AIDA framework for their own brand. Run it through the AI, compare 3 variations.
Share and discuss outputs with the group.
Module 4: Practical Prompt Library and Iterative Refinement (30 minutes)
Learning Objective
Equip participants with a reusable set of prompt templates and the habit of iterative refinement — so they leave training with tools they can use the next day.
4.1 The Prompt Library
Walk through 7 ready-made prompt templates from the
prompt-engineering-library
skill. Cover one template from each common content type:
- Instagram / Facebook caption
- WhatsApp broadcast message
- Email subject line
- Email body (short-form)
- LinkedIn post (thought leadership)
- Blog post introduction
- SMS / short copy
For each template, show participants where to substitute their own brand details. Demonstrate live on the AI tool.
Reference: For the full library of templates, see the
prompt-engineering-library
skill.
4.2 Iterative Refinement
Explain the five prompting approaches (Upadhyay, 2024) and when to use each:
| Approach | Description | Best for |
|---|
| Zero-shot | Single instruction, no examples | Simple tasks: date, format, quick edits |
| One-shot | One example provided before the request | Tone matching when you have one good example |
| Few-shot | 2–5 examples provided before the request | Brand voice matching; consistent series of posts |
| Chain-of-thought | Ask AI to reason step by step before outputting | Strategy tasks: campaign planning, audience analysis |
| Iterative refinement | Start broad, narrow with follow-up prompts | Any high-stakes content requiring multiple drafts |
Demonstrate iterative refinement live using a three-prompt sequence:
- Prompt 1 (broad): Run the initial Alpha-Beta-Gamma-Delta-Epsilon prompt. Review output.
- Prompt 2 (narrow): "Good. Now shorten the first caption to under 100 characters and make the CTA more direct. Keep everything else the same."
- Prompt 3 (refine): "Replace the third hashtag with a more location-specific Ugandan tag. Do not change the rest."
Show that iterative refinement produces better results than trying to write the perfect prompt first time.
4.3 The AI Content Quality Checklist
Before any AI-generated content is published, apply the checklist from the
skill. Key checks include:
- Does this sound like a human wrote it, or an AI?
- Does it match the brand voice guide?
- Are all facts, statistics, and figures verified?
- Are there any AI clichés ("game-changing", "seamless", "dive into", "in today's fast-paced world")?
- Has a human editor reviewed and approved it?
Reference: For the full humanisation checklist and editing protocol, see the
skill.
4.4 East African Context Notes
Apply these additional considerations when prompting for EA-market content:
- WhatsApp broadcast prompts require different constraints than email or Instagram. Specify: conversational tone, short paragraphs, no HTML, no markdown formatting, clear opt-out line.
- Luganda / Swahili phrases can be embedded. Instruct the AI: "Include one Luganda phrase in the caption and provide an English translation in brackets."
- Low-data context means shorter, punchier copy performs better. Default to under 150 characters for captions; under 160 characters for SMS.
- Local cultural context is powerful. Instruct the AI to reference locally recognisable elements: boda-boda, rolex (the Ugandan street food), market day, supper culture, Kampala traffic, KCCA, etc.
- UGX pricing should always appear in prompts where price is relevant. Do not let the AI invent pricing — always provide it in the Beta element.
4.5 Prompt Type Taxonomy and Examples-First Methodology
Eight prompt types — use as a decision tool:
| Type | When to Use |
|---|
| Zero-shot | Simple tasks, quick edits |
| Few-shot | Brand voice matching — correct default for any voice task |
| Instructional | Complex multi-part deliverables |
| Conversational | Refinement, ideation, brainstorming |
| Contextual | Strategy documents, personas, reports |
| Creative | Storytelling, campaign concepts |
| Chain-of-Thought | Strategy, audience analysis, decision-making |
| Systematic | Repeatable tasks: caption series, email sequences |
Few-shot is the correct default for brand voice work. Supply 2–3 examples of actual human writing before issuing a content task. Master structure:
Here is an example of [content type] in the voice and style I want:
###
[Example 1]
###
[Example 2]
###
Now write a new [content type] on [topic] in the same voice and style.
Adjectives are imprecise; examples are exact. Three or more examples enable reliable voice matching; fewer produce inconsistency. Training activity: participants paste three real brand posts, generate a new post, compare against tone-adjective-only output.
Worked Prompt Examples
Example 1: Instagram Caption — PAS Framework (Telecoms Brand)
Alpha: You are a senior social media copywriter specialising in Ugandan telecoms brands.
Beta: The brand is [X], targeting urban professionals aged 25–35 in Kampala. Brand voice: confident, friendly, slightly witty. Product: 5G data bundle, UGX 15,000 for 5GB/7 days.
Gamma: Write 3 Instagram captions using the PAS framework (Problem → Agitate → Solution). Each caption must be under 150 characters, end with a clear CTA, and include 2–3 relevant hashtags.
Delta: Do not mention competitor brands. Do not use the words "affordable" or "seamless". Avoid exclamation marks.
Epsilon: Present as a numbered list with the PAS label for each element.
Example 2: Email Subject Lines — AIDA Framework (Microfinance Institution)
Alpha: You are an email copywriter for an East African microfinance institution.
Beta: We are sending a re-engagement email to customers who have not transacted in 90 days. Brand tone: warm, encouraging, professional.
Gamma: Write 5 subject line options using the AIDA framework (Attention → Interest → Desire → Action), each under 50 characters.
Delta: Do not mention overdue balances. Do not use the word "urgent".
Epsilon: Numbered list, subject line only, no explanation.
Example 3: WhatsApp Broadcast — BAB Framework (Delivery Service)
Alpha: You are a WhatsApp copywriter for a Kampala-based same-day delivery service.
Beta: Brand: [X]. Audience: small business owners and traders aged 28–45 in Kampala. Tone: direct, reliable, no-nonsense. Promotion: free delivery on orders above UGX 50,000 this Friday only.
Gamma: Write one WhatsApp broadcast message using the BAB framework (Before → After → Bridge). Maximum 3 short paragraphs. End with a WhatsApp reply CTA ("Reply YES to book your first delivery").
Delta: No emojis. No markdown. No bullet points — plain text only. Do not use the phrase "limited time".
Epsilon: Single plain-text message, no formatting, ready to paste into WhatsApp Business.
Quality Criteria
The completed training guide meets the standard if:
- All 4 modules are included with accurate time allocations totalling approximately 2.5 hours
- The Alpha-Beta-Gamma-Delta-Epsilon framework is explained in full with at least one complete worked example per element
- All 7 copywriting frameworks are named, defined, and illustrated with a Uganda/East Africa brand example
- Hands-on activities are specified in Modules 2, 3, and 4 — not lecture content only
- All worked examples use Ugandan/EA brands, UGX pricing, and local cultural references where relevant
- Output is structured so a non-technical facilitator can deliver it without additional preparation
- The
prompt-engineering-library
- The guide is written in British English with imperative language throughout