In the rapidly evolving world of artificial intelligence and large language models (LLMs), the ability to structure an effective prompt is both a science and an art. Prompting frameworks are powerful tools that **bring structure, clarity, and intent to our interactions with AI.** Think of a prompt framework as a blueprint or scaffold: it helps you organize your thoughts and communicate them effectively to the model. Each framework is designed with a particular use case in mind - from creative writing and content generation to customer support and business reporting. This article explores a broad collection of the most effective and widely used prompting frameworks, arranged alphabetically. We break down their structure, explain when to use them, and provide real-world examples. My favorite frameworks: - C.R.E.A.T.E - Character, Request, Examples, Adjustment, Type, Extras - M.I.N.D.S - Map, Investigate, Navigate, Develop, Sustain - S.O.L.V.E - Situation, Objective, Limitations, Vision, Execution ## **A.P.E (Action, Purpose, Expectation)** **Best for**: Blog writing, educational content - **Action**: What the model should do - **Purpose**: Why you want it - **Expectation**: What success looks like **Example**: > Action: Create a household cleaning checklist tailored for a family with two kids and pets. > Purpose: Help streamline weekend cleaning routines and share responsibilities between family members. > Expectation: Include daily, weekly, and monthly tasks categorized by room. List who in the family can handle each based on age or time availability. ## **B.A.B (Before, After, Bridge)** **Best for**: Storytelling, persuasion, product marketing - **Before**: Describe the current state/problem - **After**: Show the improved future - **Bridge**: Explain how to get there **Example**: > Before: “Every evening, I’d feel overwhelmed trying to decide what to cook. I’d often give up and order unhealthy takeout.” > After: “Now I cook balanced meals in under 30 minutes, and I actually look forward to dinnertime.” > Bridge: “I used a weekly meal planning template that builds grocery lists and gives prep shortcuts tailored to my diet.” ## **C.A.R.E (Context, Action, Result, Example)** **Best for**: Support, instructional responses - **Context**: What’s happening? - **Action**: What should LLM do? - **Result**: What should the outcome look like? - **Example**: Provide a sample **Example**: > Context: Your child is struggling with organizing homework and often forgets assignments. > Action: Generate a daily homework planner for schoolchildren. > Result: Should be simple, color-coded, and printable. > Example: “Include sections for due dates, subject, task breakdown, and a checkbox for completion. Use bright visuals for motivation.” ## **C.O.A.S.T (Context, Objective, Action, Steps, Task)** **Best for**: How-to guides, onboarding, training content - **Context**: Background information - **Objective**: Goal of the task - **Action**: Instructions to follow - **Steps**: Sequence of tasks - **Task**: Deliverable to be returned **Example**: > Context: You're preparing to host your extended family for Thanksgiving dinner for the first time. > Objective: Plan a full Thanksgiving meal for 10 people. > Action: Provide a detailed grocery list, prep schedule, and recipe recommendations. > Steps: 1) Plan menu 2) Buy groceries 3) Prepare dishes in advance 4) Organize serving stations. > Task: Output a printable 3-day preparation checklist with estimated cooking times. ## **C.R.E.A.T.E (Character, Request, Examples, Adjustment, Type, Extras)** **Best for**: Advanced prompt structuring - **Character**: Role LLM should play - **Request**: What it should do - **Examples**: Add 1–2 examples - **Adjustment**: Request revisions or iterations - **Type**: Format of output (list, table, etc.) - **Extras**: Any final context or tone **Example**: > Character: You’re a fitness coach who designs home workouts for busy parents. > Request: Create three 20-minute full-body routines using minimal equipment. > Examples: Include warm-up, main exercises, and cool-down. > Adjustment: One routine should be quiet for nap times, another should involve kids. > Type: Return in table format with sets and reps. > Extras: Keep language encouraging and fun. ## **C.R.I.S.P.E (Capacity, Role, Insight, Statement, Personality, Experiment)** **Best for**: Creative marketing, ideation, A/B testing - **Capacity**: Define role - **Insight**: Provide data/context - **Statement**: What you want done - **Personality**: Desired tone - **Experiment**: Request variations **Example**: > Capacity: Personal budgeting coach > Insight: My monthly expenses often exceed my income due to impulse buys and food delivery. > Statement: Help me design a realistic monthly budget. > Personality: Friendly, non-judgmental, motivating. > Experiment: Offer three budget templates — minimalist, balanced, and aggressive savings — and briefly describe their pros and cons. ## **F.O.C.U.S (Frame, Outline, Conduct, Understand, Summarize)** **Best for**: Campaign summaries, retrospectives - **Frame**: Set campaign or project goals - **Outline**: Define the strategy - **Conduct**: Execute the plan - **Understand**: Reflect on outcomes - **Summarize**: Consolidate insights **Example**: > Frame: Plan a family road trip to three national parks over 10 days with a focus on nature and learning. > Outline: Route includes Yellowstone, Grand Teton, and Bryce Canyon. Budget: $2,000. > Conduct: Reserve lodging, rent a minivan, prepare packing list, and pre-cook meals. > Understand: Journal reflections and review photos each day. > Summarize: Compile a photo album with notes for future trip planning. ## **G.R.W.C (Goal, Return format, Warnings, Context)** **Best for**: Business briefs, legal/technical prompts - **Goal**: What outcome do you want? - **Return**: Output format (table, bullets, summary) - **Warnings**: Any limitations or things to avoid - **Context**: Supplementary background **Example**: > Goal: Create a comparison chart of three different local daycare centers. > Return: Return results in a table with categories for cost, location, hours, meal options, and parent reviews. > Warnings: Avoid including unverified or outdated information. > Context: My child is 2 years old and we need care on weekdays from 8am to 5pm. ## **M.I.N.D.S (Map, Investigate, Navigate, Develop, Sustain)** **Best for**: Market strategy, life planning, competitive analysis - **Map**: Define the landscape - **Investigate**: Explore possibilities, pain points, or options - **Navigate**: Highlight challenges, trade-offs, or constraints - **Develop**: Build a clear path forward - **Sustain**: Plan for long-term follow-up or iteration **Example**: > Map: I want to switch careers from sales to UX design. > Investigate: What skills are required? What certifications are respected? How long will a transition take? > Navigate: I have limited time and budget. I also need to maintain a part-time job during the switch. > Develop: Create a 6-month learning and networking roadmap, with study goals, community events, and sample projects. > Sustain: Recommend monthly checkpoints and habits to stay motivated long-term. ## **P.A.C.T (Problem, Approach, Compromise, Test)** **Best for**: Decision making, parenting, collaborative planning - **Problem**: Define the issue - **Approach**: Offer a potential solution - **Compromise**: Acknowledge trade-offs - **Test**: Suggest how to trial or validate the approach **Example**: > Problem: My teen wants to increase screen time beyond the current 1-hour weekday limit. > Approach: Propose an extended limit (e.g., 90 minutes) with an earned bonus system for chores or good grades. > Compromise: I’ll allow up to 2 hours on Fridays only if homework is done and we check usage logs together. > Test: Run this plan for two weeks and evaluate behavior, mood, and school performance. ## **P.A.R (Problem, Action, Result)** **Best for**: Quick retrospectives, self-reflection, case notes - **Problem**: What was wrong? - **Action**: What did you do? - **Result**: What happened? **Example**: > Problem: I was constantly late picking up my kids from school. > Action: Set alarms 30 minutes ahead and adjusted my work calendar to block that time. > Result: Haven’t been late for over a month and my stress has reduced significantly. ## **R.A.C.E (Role, Action, Context, Expectation)** **Best for**: Business tasks, event planning, formal communications - **Role**: Define the persona or function - **Action**: What they must do - **Context**: Background information - **Expectation**: Output format or quality **Example**: > Role: You are a family reunion planner. > Action: Design a 2-day itinerary for a 25-person family gathering in a state park. > Context: The family includes children and grandparents, and dietary needs vary. Some prefer hiking, others board games and quiet. > Expectation: Include indoor and outdoor options, mealtimes, and accessible accommodations. Format the output by time slots. ## **R.I.S.E (Request, Input, Scenario, Expectation)** **Best for**: Decision prompts, scenario planning, recommendations - **Request**: What do you want done? - **Input**: Relevant data or preferences - **Scenario**: What's the situation? - **Expectation**: Desired form of the result **Example**: > Request: Recommend a family dog breed. > Input: Two kids under 8, a large fenced yard, someone always home > Scenario: We've never owned a dog before, but want one that is playful and gentle. > Expectation: Return three suitable breeds with pros, cons, and care needs in a comparison table. ## **R.O.S.E.S (Role, Objective, Steps, Expected Solution, Scenario)** **Best for**: Problem-solving, diagnostics, tech support - **Role**: Who is LLM? - **Objective**: What to achieve - **Steps**: Logical process to follow - **Expected Solution**: What success looks like - **Scenario**: Context for the task **Example**: > Role: You are a personal IT assistant. > Objective: Fix recurring Wi-Fi dropouts in a home office. > Steps: 1) Check router location and interference 2) Update firmware 3) Replace old cables 4) Run speed tests. > Expected Solution: A stable connection that supports video calls without interruption. > Scenario: The issue worsens during storms and happens mostly in the upstairs room. ## **R.T.F (Role, Task, Finish)** **Best for**: Quick setup prompts for single-task outputs - **Role**: What is the model acting as? - **Task**: What to do - **Finish**: Desired final format or structure **Example**: > Role: You are a meal prep expert. > Task: Design a 5-day lunch prep plan for a vegetarian who works 9–5 and wants to save money. > Finish: Return a grocery list and a chart with prep steps for each day. Meals should be under $5 and last 3–4 days refrigerated. ## **S.O.L.V.E (Situation, Objective, Limitations, Vision, Execution)** **Best for**: Long-term planning, relationship or lifestyle coaching - **Situation**: What’s going on? - **Objective**: Desired transformation - **Limitations**: What stands in the way? - **Vision**: Ideal future state - **Execution**: How to start moving toward that vision **Example**: > Situation: I'm feeling burned out from work and disconnected from my hobbies. > Objective: Reclaim time for self-care and creativity over the next 3 months. > Limitations: I have young kids, unpredictable work hours, and limited free time. > Vision: Spend at least 3 evenings a week doing something fulfilling — guitar, painting, or reading. > Execution: Block one hour daily as “me time,” reduce social media, delegate one weekly task, and join a local hobby group. ## **S.T.A.R (Situation, Task, Action, Result)** **Best for**: Job interviews, personal achievement logs - **Situation**: The context - **Task**: The problem to address - **Action**: What was done - **Result**: What happened **Example**: > Situation: Our local parent-teacher group struggled with fundraising. > Task: Improve participation and donation levels. > Action: Launched a “family movie night” series with low-cost tickets and a raffle. > Result: Attendance doubled, and we raised 3x our usual amount within 2 events. ## **T.A.G (Task, Action, Goal)** **Best for**: Fast plans, business or family decisions - **Task**: What to do - **Action**: How to do it - **Goal**: The outcome desired **Example**: > Task: Organize a surprise birthday party for my partner. > Action: Find a dog-friendly venue, send invites to 12 friends, and order a custom cake. > Goal: Make it a casual, joyful afternoon where everyone brings a memory card or letter to share. ## **T.R.A.C.E (Task, Role, Action, Context, Example)** **Best for**: Content creation, persuasive writing, campaign design - **Task**: Core job - **Role**: Who LLM should be? - **Action**: Specific action to perform - **Context**: Supporting information - **Example**: Template, tone, or structure sample **Example**: > Task: Write a school newsletter section on healthy snacks. > Role: You are a nutritionist writing to parents. > Action: Provide a top-5 list of after-school snacks with explanations. > Context: The school promotes healthy habits, and many kids have nut allergies. > Example: “Include one no-prep option, one warm item, and one allergen-free pick. Tone should be friendly and informative.” With the wide variety of frameworks at your disposal, your prompts can become consistently effective, clear, and aligned with your unique goals. Each framework offers a blueprint tailored to a particular kind of thinking - analytical, persuasive, operational, strategic. Learning to choose and combine frameworks based on the task at hand is one of the key skills of a modern AI-powered professional. Think of these tools not just as formulas, but as creative springboards that unlock the full value of language models.