How to Write AI Prompts That Actually Work
The CLEAR Framework: 5 steps that turn vague requests into on-brand, usable output.
The problem usually isn't the model. Vague input produces vague output, and most people write prompts the way they'd write a quick note to a colleague who already knows the context: except the model has none.
The CLEAR Framework covers it: Context establishes the role and situation, Limits set the constraints that keep output usable, Examples show what success looks like, Action states the specific task, and Refinement treats the first response as a draft rather than a verdict.
Role-based prompting gets its own section. Assigning a specific expert role changes vocabulary, priorities, and depth. "Write about pricing" and "You're a pricing consultant for B2B service firms: audit this offer" produce entirely different work from the same model.
Chain of thought is the accuracy tool. Asking the model to show its reasoning before its answer surfaces bad assumptions where you can catch them, instead of burying them inside a confident paragraph.
The advanced section deals with hallucination and generic voice: how to constrain claims to supplied material, and how to feed voice examples so output stops sounding like the internet average. The session closes on the mistakes that cause most bad results, all of which come down to asking for something without defining it.
Questions people ask about this
Answers pulled from the session itself. Where a number or an outside claim shows up, the reference is footnoted to the source list on this page. Last reviewed August 20, 2026.
- What does this masterclass teach about writing AI prompts?
- This session teaches the 5 step CLEAR Framework: Context, Limits, Examples, Action, and Refinement. You learn role-based prompting to adopt expert perspectives, chain of thought techniques to audit reasoning, and practical methods to eliminate hallucinations. Every step turns vague inputs into accurate, on-brand responses.
- How do I start writing better AI prompts today?
- You start by shifting from a search engine mindset to a collaborator mindset that provides clear context, boundaries, and formatting instructions. Use the CLEAR Framework to structure your initial request before sending it to the model. Saving your successful prompt structures creates a reusable prompt library for your team.[1]
- How much does this training cost in time and money?
- The masterclass session is free to watch. You only invest the time required to complete the training and apply the steps to your workflow. Optional paid work products are available through Sell More Resources if you want ready-made prompt templates.
- What is the most common mistake in AI prompt writing?
- The most common mistake is providing vague input and asking for something without defining it, which forces the model to fill gaps with assumptions. Insufficient context leads to AI hallucinations, contradictory advice, and generic responses. You prevent this by setting clear limits, specifying required formats, and providing 3 to 5 examples.[2][3]
- How do I know if my AI prompts are working?
- You know your prompts are working when the model generates usable, structured outputs with minimal editing required. Chain of thought prompting also displays reasoning steps before giving an answer so you can catch bad assumptions immediately. Consistent output quality across daily tasks proves your prompts are correctly constructed.[1]
- Why should I assign a specific expert role to the AI?
- Assigning a specific professional role gives the model a clear point of view, which changes its vocabulary, priorities, and depth. Telling the model it is a pricing consultant for a service business operator produces a vastly different analysis than asking it to write about pricing broadly. This simple step hires the expert inside the model to improve performance.[4]
The class, mapped
Original diagrams built from this session: the order the work runs in, what each stage owes the next, and the list to work against once the video ends.
The CLEAR Prompting Process
- 1Prompt Author
Establish Context
Define the specific background, scenario, and operational environment.
- 2Prompt Author
Set Limits
Define precise boundaries, negative constraints, and length rules.
- 3Prompt Author
Provide Examples
Supply sample outputs that show expected tone, structure, and quality.
- 4Prompt Author
State Action
Specify the direct task using clear command verbs.
- 5User & Model
Execute Refinement
Treat initial output as a draft and prompt again to polish details.
Prompt Quality Control Checklist
Prompt Architecture
- Assign an expert role to adjust vocabulary, priorities, and depth.
- Provide full context before stating the primary task.
- Set strict limits to keep output within clear boundaries.
- Include at least 1 benchmark example showing desired output format.
Accuracy and Voice Guardrails
- Request chain of thought reasoning before the final response.
- Constrain factual claims exclusively to supplied context materials.
- Provide authentic voice samples to prevent generic writing.
Prompting Strategy Selection Matrix
↑ Complex Reasoning
↓ Standard Output
- 1Basic Direct Request Standard Output, Low Output Constraint
- 2Role-Based Persona Prompt Complex Reasoning, Low Output Constraint
- 3Constrained Source Audit Standard Output, High Output Constraint
- 4Chain of Thought Framework Complex Reasoning, High Output Constraint
Keep going
Related training
Same pillar, same problem, different angle.
Share the library
Send it to the one person on your team who needs it.
The training is free and open to all. Sharing is how the operator beside you stops guessing too.
Share the site
No login, no email gate. Watch, use it, pass it on.