Why Prompts Fail Without Structure
Large Language Models (like GPT-4o, Claude 3.5, and Gemini 1.5) are probabilistic reasoners. When you feed them a brief, ambiguous request like "write a marketing plan", the model has to guess your business model, target audience, budget, and desired tone. It invariably picks the most generic average response possible.
| Letter | Element | Core Question It Answers |
|---|---|---|
| C | Context (السياق) | What is the background situation & setting? |
| O | Objective (الهدف) | What exact task or goal must be achieved? |
| S | Style (الأسلوب) | How should information be structured? (Table, bullets, report) |
| T | Tone (النبرة) | What voice and attitude? (Professional, witty, warm) |
| A | Audience (الجمهور) | Who will consume this output? |
| R | Response (المخرج النهائي) | What constraints, word limits, or rules apply? |
Practical Master Example
Marketing Campaign Launch Prompt
Context: We are a family-style Mediterranean restaurant in Cairo. We launched our second branch last month and want to increase weekend dinner reservations. Objective: Create 3 seasonal social media campaign concepts designed to increase advance reservations by 25%. Style: Organized comparison table containing: (Campaign Name, Target Channel, Core Hook, and Estimated Ad Budget). Tone: Warm, hospitable, authentic, and family-oriented. Audience: Families and couples aged 28-50 residing in the metropolitan area. Response: A structured table followed by a 4-step weekly launch checklist. Avoid jargon.
Compatible with: ChatGPT, Claude, Gemini, GroqOpen in Fahhim Workshop
How Fahhim Automates This for You
Typing out every framework attribute manually becomes tedious. Fahhim is built specifically to streamline this process: you choose your framework (COSTAR, ICDF, MICRO, RCR-EOC), fill in the guided prompts in English, Classical Arabic, or Egyptian slang, and Fahhim compiles a production-grade prompt ready to paste or share in 1 second.