Planning Prompt Frameworks: BROKE, COAST, SCOPE, SPARK and More

Study guide

Planning Prompt Frameworks for Managers and Planners

This course teaches you how to use planning prompt frameworks to turn a rough request into a structured plan with measurable outcomes, constraints and iteration. You will learn seven frameworks: BROKE, COAST, SCOPE, SPARK, RACEF, SPEAR and CHAIN. Each one gives you a different way to shape a prompt so the AI produces something you can actually use at work, not a generic essay. You will also learn two process ideas for teams: Agile prompt engineering and prompt readiness levels.

The course is for managers and planners who need to set goals, build business plans, run analysis or improve a draft over several rounds. You do not need any technical background. You will see worked examples from real jobs, such as a regional sales plan, a hiring plan under a fixed budget, or a root cause analysis. You will also learn the common mistakes that make plans unrealistic, like leaving out constraints or accepting the first draft.

What is a prompt framework?

A prompt is the request you type into an AI tool such as ChatGPT, Claude or Gemini. A prompt framework is a reusable structure for that request: a short checklist of the parts a good prompt contains, usually named with an acronym so it is easy to remember. Why it helps: the AI fills in whatever you leave out with generic guesses. A framework makes sure you include the parts that change the answer, such as who it is for, what you want to achieve and the format you need.

Frameworks make prompts repeatable: you fill the same parts each time, save the prompt as a template and share it with colleagues. They are not magic words: the value comes from the detail you put in each part. They help most when a request is complex or open to interpretation; a quick, simple question can stay short. There are well over sixty named frameworks, and most of them remix the same ingredients: role, task, context, audience, format, constraints and examples.

In this course you learn Planning frameworks, which is a fill-in-the-slots framework: an acronym where each letter stands for one part of your prompt. It is best for: managers and planners: business plans with measurable outcomes, iteration loops and structured analysis.

Frameworks for plans and goals

Planning frameworks are useful when the first answer is the start of a process, not the final result. They give you a structure to guide the AI through several rounds. The key habit is iteration: read the draft, spot gaps, and ask for changes. A measurable outcome is a result you can count or check, such as a revenue target or a number of hires. Without one, you cannot tell if the plan is working. Always include at least one number in your prompt.

Constraints are the limits you must respect, such as budget, time, staff or location. They stop the AI from inventing a plan you cannot deliver. State them clearly and ask the AI to respect them. The prompt shape in this lesson has six parts: role, goal, constraints, measurable targets, timeline and review point. You can reuse this shape for many planning tasks. The review point is what makes the plan improvable.

A common mistake is accepting the first draft. Planning work needs at least one revision round. Another mistake is leaving out constraints, which leads to unrealistic plans. Check both before you use the output. This course covers seven frameworks: BROKE, COAST, SCOPE, SPARK, RACEF, SPEAR and CHAIN. It also covers two process ideas for teams. Each framework gives you a different way to structure a plan.

BROKE: outcomes and evolution

BROKE stands for Background, Role, Objectives, Key Results, Evolve, and it became popular in the Chinese prompt community in 2023. It is a planning framework, so it suits work that unfolds over weeks or months rather than one off questions. Key Results borrow from OKRs, the goal setting method used across many companies. An objective says what direction you are heading. A key result puts a number and a deadline on it, which gives the model something concrete to plan against.

Evolve asks the model to propose how to improve the plan after feedback. This is the difference between a single answer and a working session, because the model suggests changes rather than only applying the ones you name. The Background section should hold only the facts that change the answer, such as market, team size, product and last period's result. Long background paragraphs dilute the prompt and push the model toward generic output.

The Role line sets vocabulary, priorities and level of detail. Telling the model to act as a regional sales director with ten years in B2B software produces a very different plan from leaving the role blank. BROKE is strongest when the four setup parts are tight and the fifth part is used at least once. A plan that is never evolved is a snapshot, and real quarters rarely match the first draft.

COAST and SCOPE: constraint driven plans

COAST stands for Context, Objective, Actions, Scenario, Task. A second version uses Challenge, Objective, Actions, Strategy, Tactics. Both keep the situation and the goal ahead of the steps. SCOPE stands for Situation, Constraints, Objectives, Preferences, Execution. The order matters because reality and limits come before what you want.

Both frameworks share one habit. They put constraints and context up front, which keeps plans realistic instead of theoretical. A constraint is anything you cannot change, such as budget, headcount, fixed shifts or a deadline. State it plainly so the AI does not plan around a limit that does not exist.

Preferences shape the format of the answer. Saying you want short weekly actions with an owner gets you a plan you can actually use in a meeting. The most common failure is adding constraints after the request. By then the AI has already built a plan you cannot run, and you have to start again.

SPARK and CHAIN: framing and analysis

SPARK stands for Situation, Problem, Aspiration, Result, Kismet. The aspiration and result give the model a target, while kismet invites a useful angle you did not plan for. CHAIN stands for Context, Hypothesis, Analysis, Inference, Narration. It is built for analytical reasoning and debugging, where the path from data to conclusion must be visible.

The hypothesis step is the hinge of CHAIN. It changes the model from a describer of data into a tester of an idea, and it gives you something that can be supported or rejected. A good hypothesis is narrow and falsifiable. If no data could disprove it, the model will simply agree with you, which is not analysis.

Narration is the final CHAIN step, where the finding is written in plain English. Skipping it leaves the analysis in jargon that non technical colleagues cannot use. SPARK and CHAIN work well together. SPARK frames the problem and the desired result, then CHAIN tests the most likely explanation inside that frame.

RACEF and SPEAR: iteration loops

RACEF stands for Rephrase, Append, Contextualize, Examples, Follow-up. It is a repair kit for a prompt that is close but not right, and each letter is a small, cheap edit rather than a full rewrite. SPEAR stands for Start, Provide, Explain, Ask, Rinse and repeat. It is the beginner friendly version, a five step rhythm that builds a rough prompt up through conversation.

The shared principle behind both is that the first answer is a draft. Quality comes from the second and third message, where you react to what the AI produced instead of guessing upfront. Follow-up is the most valuable move in RACEF. Reading the draft and asking for one specific change costs seconds and prevents a full rewrite later.

A common failure is starting over when the answer is wrong. Loops keep the context you already provided, so each pass builds on the last instead of resetting to zero. Choose SPEAR when you are new to prompting or the task is simple. Choose RACEF when your prompt is already detailed and just needs targeted repairs.

Planning frameworks for teams

Agile prompt engineering rolls prompts out in phases across use cases. You test and improve each one before moving to the next, so the team learns what works before it scales. Prompt readiness levels score how production ready a prompt asset is. This matters for teams that build on prompts, because it tells everyone which prompts are safe to use daily and which are still drafts.

Pick BROKE for plans with targets, because it forces you to name the outcome and the key results. It keeps a plan honest about what success looks like. Pick SCOPE for constrained planning, when time, budget or people are fixed. The constraints go into the prompt, so the AI plans inside your real limits. Pick CHAIN for analysis, when you need a step by step reasoning path. It is for questions where the answer depends on following a sequence, not jumping to a conclusion.

A prompt is an asset, like a template or a checklist. It has a version, an owner and a status. Treating it that way is what separates a team habit from a personal trick.

Frequently asked questions