
Guide to ai Prompting
be C.L.E.A.R.
Quality prompting is critical for quality output. Use the C.L.E.A.R. framework to improve your interactions with any LLM.
What it Means | Questions to Ask Yourself | |
---|---|---|
C | Context & Role | Who or what should the AI “be”? What background does it need? |
L | Language & Tone | What style or voice? (e.g., formal, friendly, technical) |
E | Exact Task & Format | What do you want? A list? A table? A summary? |
A | Add Constraints & Details | Word limit? Specific data points? References? |
R | Review & Refine | How will you check accuracy? What follow-up prompts might you need? |
For example
Bad Prompt | Good Prompt | |
---|---|---|
Strategy | Help me with strategy. | You are a retail consultant. Our US same-store sales fell 5% in Q2. Propose 3 marketing strategies to boost foot traffic—list KPIs for each. |
Summary | Summarize this report. | Summarize the attached 12-page market report in 4 bullets, focusing on customer trends and risks. |
Email Draft | Write an email. | Draft a friendly 3-paragraph email to staff announcing our new remote-work policy; address key FAQs. |
Meeting Prep | Prepare meeting agenda. | Generate a 1-hour meeting agenda for a Q3 planning session with 5 bullets: topics, time allocation, and desired outcomes. |
Checklist
Before you hit “Send,” run through:
Role & Context: Have I defined the AI’s persona and given key data?
Desired Format: Did I ask for the right structure (bullets, table, memo)?
Clarity & Constraints: Are length, style, and any necessary metrics specified?
Next Steps: Have I built in a review or follow-up (“If you need more info, ask…”)?