Use this guide like a playbook. Start with the quick start. Choose the right chatbot mode, give it real context, and work in short rounds. The goal is not a perfect prompt. It is a useful answer you know how to check.
You do not need a magic prompt. Give the chatbot a clear job, the material it should use, the constraints that matter, and the form you want back. Then improve the answer in a second round.
Task:
[What I want the chatbot to do]
Context:
[The facts, background, audience, examples, files, or links it should use]
Requirements:
- [What a useful answer must include]
- [Hard constraints or things to avoid]
Output:
- Format: [memo / checklist / table / plan / email / JSON / other]
- Length: [short / medium / detailed]
- Tone: [plainspoken / executive / friendly / direct / other]
Source rules:
- Use only [the attached material / these sources / current web sources / general knowledge].
- If using the web, cite the source beside each important claim.
- Separate confirmed facts from inference.
If critical information is missing, ask up to 3 questions. Otherwise proceed, state any important assumptions, and give me the answer.ChatGPT, Claude, Gemini, and Copilot can all draft, reason, search, and work from files or images. The important choice is no longer just which company to use. It is which mode to use, what context to provide, and how carefully to check the result.
The labels change, but the choices are increasingly similar. Match the mode to the difficulty and the cost of being wrong.
Use for questions, rewrites, summaries, brainstorming, formatting, and low-risk drafts.
Use for planning, comparisons, analysis, math, tradeoffs, and important recommendations.
Use when facts may have changed. Ask for direct sources and the date each fact was checked.
Attach the source and name the pages, sections, tables, or images the answer must use.
Use for stable preferences. Repeat task-specific rules in the current chat instead of trusting memory.
Start over when the goal, audience, or source material changes. Carry over only what still matters.
These are the rules for personal and consumer AI tools. Do not assume your personal chats are private by default. Settings differ by vendor and plan. If you would be upset to see it leaked, treat it carefully.
What if we use a paid version at work?
Use these guidelines only if your company plan keeps chats private, has the right admin settings, and says your data is not used to train or improve the model by default for that work plan.
See Vendor Differences with Data Privacy for a plain-English comparison.
The same lean prompt works across the major chatbots. The meaningful differences are reasoning controls, how they use long context, how explicitly they need format and style instructions, and what data their search or work mode can access. These labels were current in July 2026 and vary by plan.
| Chatbot | Current choices | Best prompt adjustment |
|---|---|---|
| ChatGPT | Instant for everyday work; Medium, High, Extra High, or Pro for harder work. | Keep prompts lean. State the task, hard constraints, sources, and success criteria once. Increase reasoning instead of asking it to show step-by-step thinking. |
| Claude | Use the default model for routine work and the strongest or extended-thinking choice for difficult work. | Be explicit and explain why constraints matter. Examples strongly steer style. For long documents, put the material first and the question last; use headings or XML tags when the prompt is complex. |
| Gemini | Use Flash-Lite or Flash for speed; Pro for complex reasoning; Search or Deep Research for current work. | Be concise and structured. Explicitly request the desired detail level. With long context, provide the material first and end with “Based on the material above…” followed by the task. |
| Copilot | Use Quick or Smart for everyday work, Think Deeper for analysis, and Search for current web answers. | Include goal, context, expectations, and source. Name the exact file, email, meeting, or work source it should use. Prefer positive “do this” instructions. |
Official reference points checked for this local draft: OpenAI ChatGPT model guide, OpenAI prompting guidance, Google Gemini model guide, Microsoft Copilot modes, and Anthropic prompting guidance.
You are the boss of the LLM. The better boss you are, the better responses you will get. These are practical prompt rules for beginners and improving users.
Task: [WHAT I NEED]
Context: [FACTS, AUDIENCE, EXAMPLES, OR ATTACHED MATERIAL]
Requirements: [WHAT A GOOD ANSWER MUST INCLUDE]
Output: [FORMAT, LENGTH, AND TONE]
Source boundary: [ATTACHMENTS / NAMED SOURCES / CURRENT WEB / GENERAL KNOWLEDGE]
If critical information is missing, ask up to 3 questions. Otherwise proceed and state any important assumptions.Relevant context:
- Situation: [WHAT IS HAPPENING]
- Audience: [WHO THIS IS FOR]
- Constraints: [WHAT LIMITS THE ANSWER]
- Good example: [OPTIONAL EXAMPLE]
- Source material: [PASTE OR ATTACH]
Use only the context that affects the task. If two details conflict, flag the conflict instead of silently choosing one.If missing information would materially change the answer, ask up to 3 focused questions.
Otherwise:
- proceed with the best answer
- state important assumptions
- label uncertainty
- do not invent missing factsThis is a high-consequence task. Use the deeper reasoning mode available.
Return:
1. Recommendation
2. Key assumptions
3. Important tradeoffs
4. Evidence or calculations used
5. What could make the recommendation wrong
6. What I should verify before acting
Do not provide private chain-of-thought. Give me the concise reasoning needed to evaluate the answer.First, give me a useful draft based on the information available.
After the draft, include:
- the weakest part
- the most important missing information
- 2 possible directions for the next revision
Do not revise again until I choose a direction.Draft the output.
Then critique it against:
- the goal
- the constraints
- what is missing
- what is unclear
Then produce an improved version.Give me options.
Generate 10 ideas.
Pick the best 3.
Explain why they win.
Then ask me which direction to take.Task: Write [THING] in my voice.
Use the writing samples below as evidence:
[SAMPLE 1]
[SAMPLE 2]
[SAMPLE 3]
First identify only the recurring traits that appear across the samples:
- sentence length and rhythm
- vocabulary and level of formality
- openings, transitions, and endings
- patterns to preserve
- patterns to avoid
Then write the draft. Match the recurring traits without copying phrases or inventing facts.Before you answer, do this:
1) List any contradictions in my instructions.
2) List what is missing.
3) Rewrite my prompt to be tighter and clearer.
4) Ask me the questions you still need.
Then wait.Pick a template. Fill in the brackets. Run it. Then improve it using the boss rules above.
Use this when you do not know where to start. It is the safest default template for most tasks.
Task:
[WHAT I NEED]
Context:
[RELEVANT FACTS, AUDIENCE, EXAMPLES, OR ATTACHED MATERIAL]
Requirements:
- [MUST INCLUDE]
- [CONSTRAINTS]
Source boundary:
[USE ONLY ATTACHMENTS / USE THESE SOURCES / SEARCH CURRENT SOURCES / GENERAL KNOWLEDGE IS FINE]
Output:
- Format: [FORMAT]
- Length: [LENGTH]
- Tone: [TONE]
If critical information is missing, ask up to 3 questions. Otherwise proceed, state important assumptions, and label anything you cannot verify.Before you answer, do this:
1. List any contradictions in my instructions.
2. List what is missing.
3. Rewrite my prompt to be tighter and clearer.
4. Ask me the questions you still need.
Then wait.Use this to set tone, depth, and perspective without pretending the model is magically authoritative.
Analyze this from the perspective of a [ROLE OR DISCIPLINE].
Task:
[SPECIFIC TASK]
Context:
[RELEVANT FACTS OR MATERIAL]
Focus on the tradeoffs and priorities this discipline normally considers:
- [PRIORITY 1]
- [PRIORITY 2]
Do not claim to be a real professional or invent expert consensus. If credible practitioners would disagree, explain the disagreement.
Output:
- Audience: [AUDIENCE]
- Format: [FORMAT]
- Detail level: [BRIEF / MODERATE / DEEP]Respond from the perspective of a clear, practical communication coach with expertise in writing firm but respectful emails.
What this role typically cares about:
- clarity
- respect
- getting to the point
Task: Write an email to my landlord about a leaking ceiling that needs repair.
Instructions:
- Use short sentences and plain English.
- Do not use hype.
- If anything important is missing, ask me questions first.
Output:
- Terminology level: accessible
- Tone: confident and respectful
- Format: subject line + email body + 3 alternate subject linesGive me 3 versions.
Explain what each version optimizes for.
Recommend one and tell me why.Use this when you want multiple perspectives, disagreement, and synthesis. Treat it as simulated viewpoints, not a substitute for real experts.
Analyze [PROBLEM] through these decision lenses:
- [LENS 1, such as finance]
- [LENS 2, such as operations]
- [LENS 3, such as customer experience]
For each lens, give:
- priorities
- likely recommendation
- main risk
- evidence needed
Then summarize:
- agreement
- conflicts
- uncertainty
- your recommended path and why
These are analytical lenses, not quotes from real experts. Do not invent facts or consensus.Work in rounds:
1. Clarify the problem.
2. Draft each viewpoint.
3. Summarize agreement, disagreement, and uncertainty.
4. Revise the recommendation.Use this when accuracy matters. It forces the model to verify instead of just sounding correct.
Turn on web search and fact-check the text below.
For each material factual claim, return:
- Exact claim
- Status: Supported / Unsupported / Contradicted / Outdated
- Best primary source
- Direct link
- Date checked
- Short explanation of what the source actually supports
Rules:
- Open the source; do not rely on a search-result snippet.
- Prefer primary or official sources.
- A citation must support the specific claim beside it.
- Separate facts from interpretation.
- If reliable sources disagree, show the disagreement.
- Do not guess.
Text:
[PASTE TEXT]Use sources I can click.
Prefer primary sources when possible.
If source quality matters, ask me what sources I trust.Use this to get a repeatable process, not just advice.
You are an expert in [domain] and an experienced teacher.
Create a step by step guide to accomplish [specific goal].
Instructions:
- If laws, policies, software, pricing, or product details may have changed, verify current information first.
- Ask questions if anything important is missing.
- Use plain language unless I ask for technical detail.
Structure your response as:
Requirements:
[what is needed]
Preparation:
[setup steps]
Instructions:
[step by step actions]
Troubleshooting:
[common issues and fixes]
Variations:
[alternate approaches]
Goal:
[specific objective]We are going to do this in rounds.
Round 1: Ask clarifying questions.
Round 2: Propose 2 approaches and recommend one.
Round 3: Draft the steps.
Round 4: Critique the steps for gaps and risk.
Round 5: Produce the improved version.Use this to turn bureaucratic confusion into practical steps and known pitfalls.
You are an expert at government processes and navigating bureaucracy.
I need help with: [what you are trying to do]
Use current official sources and link them.
Prefer official agency websites.
Tell me which parts vary by jurisdiction.
If anything important is missing, ask me questions first.
My details:
- Who I am: [details]
- Goal: [goal]
- Agency or system: [if known]
- Jurisdiction: [country / state / city / institution]
- Current status: [what I have done]
- Constraints: [deadlines, missing documents, budget, legal issues]
- Dates and deadlines: [known dates]
- Special factors: [anything unusual]
What I want:
1. A simple overview of how the system works
2. A numbered action plan for my situation
3. Required forms, documents, IDs, and evidence
4. Expected fees
5. Decision points and bottlenecks
6. Common mistakes and how to avoid them
7. A realistic timeline
8. Questions the agency is likely to ask me
9. A script for calling or emailing the agency
10. A backup plan if the main path stallsUse this when you want practice, feedback, and increasing difficulty.
You are an experienced teacher, coach, and interviewer in [field].
Help me prepare for [role / exam / conversation / topic].
Do the following:
1. Ask me for the relevant materials first.
2. Review the materials and identify focus areas.
3. Create a structured practice session with questions of varying difficulty.
4. After each answer, score me briefly on:
- clarity
- completeness
- credibility
5. Tell me whether I should expand, clarify, or move on.
6. End with a summary of strengths and the top improvements I should make.
Start by asking for my materials.Work in rounds:
1. Ask for the materials.
2. Build the practice plan.
3. Critique the plan for missing difficulty levels.
4. Improve it.Use this when you need output in a strict format every time.
You are excellent at following directions.
I am giving you a strict output template with [PLACEHOLDERS].
Rules:
- Output must match the template exactly.
- Do not add extra sections.
- Do not invent facts.
- If a field is missing, either ask me for it or mark it as [MISSING].
Template:
[PASTE TEMPLATE HERE]
Apply this template to:
[YOUR REQUEST]If the template and my instructions conflict, stop and ask me which one should win.Use this when you need current, source-backed answers instead of generic explanations.
Turn on search or research mode and investigate:
[TOPIC]
Scope:
- Decision or question: [WHAT I NEED TO KNOW]
- Geography or market: [SCOPE]
- Date range: [SCOPE]
- Preferred sources: [PRIMARY / OFFICIAL / OTHER]
Output:
1. Short answer
2. Key findings with a direct citation beside each claim
3. Source quality or conflicts
4. What is confirmed versus inferred
5. Open questions
Open every cited source. Do not cite search-result snippets. Include the date checked.Use this when you are uploading a file and want the model to stay grounded in that file.
Use the attached document as the primary source.
Task:
[TASK]
First confirm:
- document title and date
- whether all relevant pages are readable
- any missing, scanned, or ambiguous material
Rules:
- Cite the page, section, or table supporting each important answer.
- Quote briefly when exact wording matters.
- Clearly label outside knowledge.
- If the document does not answer something, say so instead of filling the gap.
Based on the document above, complete the task.Use this for screenshots, photos, dashboards, UI bugs, and visual analysis.
I am uploading an image or screenshot.
First:
1. Describe exactly what you see.
2. List any ambiguities or things you cannot read clearly.
Then do this task:
[TASK]
Rules:
- Separate observation from interpretation.
- If text is visible, extract the text first.
- If you are inferring something, label it as inference.
- If the image alone is not enough, tell me what else you need.Use this when different articles, documents, or vendors disagree.
Compare these sources on:
[TOPIC]
Sources:
- [SOURCE 1]
- [SOURCE 2]
- [SOURCE 3]
Output:
1. What all sources agree on
2. Where they conflict
3. Which source appears strongest and why
4. What is still uncertain
5. Your best current conclusion
Rules:
- Link the sources.
- Prefer primary or official sources when possible.
- Separate evidence from interpretation.Use this when you want data you can paste into a spreadsheet, script, or workflow.
Return the answer in [JSON | table] format.
Schema or columns:
- [FIELD 1]
- [FIELD 2]
- [FIELD 3]
Rules:
- Use null for unknown values.
- Do not invent missing values.
- If the schema is unclear, ask me questions first.
- Keep all field names exact.
Task:
[YOUR REQUEST]Return the answer in JSON format.
Schema:
- company_name
- product_name
- category
- target_customer
- pricing_notes
- source_url
- confidence
Rules:
- Use null for unknown values.
- Do not invent missing values.
- Keep all field names exact.
Task:
Research the top embedded DevOps tools vendors and return the results in this schema.Most people stop after one answer. Treat the chat as a short working session, not a vending machine.
Round 1: Understand
Restate the goal, assumptions, missing context, and likely risks.
Round 2: First pass
Create the first version. Optimize for usefulness, not perfection.
Round 3: Check
Critique your own output. What is missing, weak, unsupported, or risky?
Round 4: Revise
Fix the important issues and give me a cleaner version.
Round 5: Human review
List the claims, decisions, or actions I should verify myself before relying on this.