25 professional, fill-in-the-blank prompts for email, meetings, writing, research and planning — each with guidance on how to use it, what to ask next, and what to watch out for. Works in ChatGPT, Claude and Gemini.
Every prompt follows the same 5-part structure we teach in class. Fill in the [square brackets] with your specifics — the more context you give, the better the result.
RRoleWho should the AI be (expert, tone, style)
CContextBackground, constraints, and what's at stake
TTaskExactly what you want done (numbered steps help)
FFormatShape of the output (bullets, tables, word count)
CConstraintsTone, what to avoid, when to flag uncertainty
Golden rules
Give examples: show the AI what good looks like with 1–2 real samples.
Iterate, don't iterate endlessly: refine the prompt 2–3 times, then move on.
Never paste confidential data into consumer tools—redact names, numbers, and proprietary info.
Verify facts and numbers: AI can sound confident while being wrong.
Save your winners: keep 3–5 best prompts in a personal library for reuse.
Email & Inbox
Turn email from a time sink into a five-minute decision engine. Extract action items, drafts replies in your voice, and triage without reading everything.
01
Summarize any email thread in 30 seconds
Beginnersaves 15–20 min per thread
Use it when: You receive a long back-and-forth and need the key points and action items without reading every message.
prompt.txt
ROLE:
You are an executive assistant skilled at distilling email threads to their essence.
CONTEXT:
Email threads often contain repetition, tangents, and emotional language. Your job is to extract signal, not noise. The reader is busy and needs to know: what happened, what's the decision, what's next, and who needs to do it.
TASK:
1. Extract the main topic and outcome (1 sentence)
2. List all action items with owner and deadline (bullets)
3. Flag any decisions that affect the reader directly (separate section)
4. Note any missing information that could cause problems
FORMAT:
Outcome: [1 sentence]
Actions:
- [Owner]: [Task] by [date if mentioned]
Decisions: [if any]
Red flags: [if any]
CONSTRAINTS:
Use the sender's tone and terminology. If a deadline is implied but not stated, say "deadline unclear." If you cannot find an action item, say "no explicit action items." Do not invent urgency. Flag anything you're unsure about.
INPUT:
[Paste the email thread here]
How to fill it in
Paste the entire thread, starting from the oldest email—the AI needs context.
Include thread subject and participants if they're not in the text.
Paste exactly what you see; don't clean it up first.
Then ask
Draft a reply to [person] confirming we'll ship by [date].
Create a meeting agenda to discuss the decision in point 3.
Write a one-line Slack message to [team] flagging the red flag above.
Pro tip Paste the raw thread as-is—messy threads actually help the AI spot important details faster. If you get an action item wrong, paste just the relevant 3–4 messages and ask the AI to re-extract.
Watch out Redact names, customer names, or revenue figures before pasting if sharing with team members.
02
Draft a reply that sounds exactly like you
Beginnersaves 10–15 min per email
Use it when: You need to reply to an email but don't have time to write it, and you want it in your voice.
prompt.txt
ROLE:
You are a skilled ghostwriter who writes in a specific person's voice and communication style.
CONTEXT:
Voice matters. A formal person using casual language sounds fake. A direct person hemming and hawing sounds weak. You'll study the example emails below and match tone, sentence structure, and how the writer handles disagreement, enthusiasm, and deadline pressure.
TASK:
1. Read the three example emails to understand voice and style
2. Draft a reply to the new email that matches that voice
3. Keep the draft to 2–3 sentences if it's a quick answer, or 4–5 if it needs explanation
4. Do not add emojis, excessive exclamation marks, or phrases the person never uses
FORMAT:
[Draft email reply, ready to paste into the reply field]
CONSTRAINTS:
Match the person's actual communication style—do not sanitize or improve it. If the original emails are casual, be casual. If they're formal, match that. Do not change the message's substance; just adopt the voice. Ask me to clarify if the request conflicts with the person's usual values or tone.
INPUT:
Example 1:
[Paste an email this person wrote]
Example 2:
[Paste another email]
Example 3:
[Paste a third email]
---
New email to reply to:
[Paste the email that needs a response]
What's your high-level intent for the reply? (e.g., "agree and add a suggestion," "decline politely and offer an alternative," "confirm receipt and ask one question")
How to fill it in
Grab 3 recent emails the person actually wrote—the more varied, the better (casual, formal, disagreeing, enthusiastic).
State your intent clearly (agree, decline, clarify, escalate, etc.) so the AI doesn't guess.
If you always include your name or a sign-off, mention it so the AI includes it in the draft.
Then ask
Make this reply a bit more [direct / warm / brief].
Add a sentence about [specific topic] to the reply above.
Rewrite the opening to reference our conversation on [date] instead.
Pro tip Include emails from different contexts (replying to a peer, a direct report, a client). The AI will blend them into a flexible voice that works for any situation.
03
Decline a request without sounding like a jerk
Beginnersaves 8–12 min per decline
Use it when: Someone asks you for something you can't or shouldn't do, and you want to say no without damaging the relationship.
prompt.txt
ROLE:
You are a diplomatic communicator who says no while preserving relationships and leaving the door open.
CONTEXT:
Bad declines feel cold, evasive, or aggressive. Good ones acknowledge the person's request, explain the real constraint, and offer an alternative when possible. The goal is to be clear, kind, and memorable.
TASK:
1. Acknowledge the request and why it matters
2. State your real reason for declining (time, priority, capability, policy—pick the true one)
3. If there's a way to partially help or suggest an alternative, include it
4. End with a door-opener (future timing, related help, stay in touch)
FORMAT:
[2–4 sentence email, ready to send]
CONSTRAINTS:
Be honest about your reason; vague declines frustrate people. Do not over-apologize or sound weak. Match the formality of the original request. If you can help in a small way, mention it. Do not commit to anything you cannot deliver.
INPUT:
[Paste the request or email you need to decline]
What's your real constraint? (time / priority / capability / policy / budget / other)
Who is this person and what's the relationship? (peer / boss / client / job applicant / vendor)
How to fill it in
Be specific about your constraint; 'I'm swamped' is weaker than 'we're in peak season and my team is at capacity.'
If there's a timing where you could help (after a project, next quarter), say it.
Suggest an alternative person or approach if you know one.
Then ask
Warm this up a bit—add a sentence about [their project / their situation] to show I understand.
Add a timeframe: can I revisit this in [3 months / after Q4]?
Include my alternative suggestion: they could try [approach] instead.
Pro tip The fastest way to be forgotten is to ignore a request. Declining with a genuine reason and a small alternative takes 30 seconds and keeps the relationship warm.
04
Triage your inbox in 10 minutes
Beginnersaves 30–45 min of inbox stress
Use it when: Your inbox has 200+ unread emails and you don't know where to start. You need a quick scan of what's urgent, important, and what can wait.
prompt.txt
ROLE:
You are an inbox strategist who quickly categorizes emails by urgency and importance.
CONTEXT:
Most email feels urgent but isn't. A good triage surfaces real deadlines, relationship maintenance, and information worth filing, while separating status updates, newsletters, and filler.
TASK:
1. Review all subject lines and senders provided
2. Group into buckets: URGENT (requires action today), IMPORTANT (this week), FYI (skim or skip), UNSUBSCRIBE (consider removing from mailing list)
3. For URGENT and IMPORTANT, state why and suggest a 1-sentence action
4. Do not read the full body of every email; subject lines and senders are enough
FORMAT:
URGENT (read these today):
- [Sender]: [Subject] – [Why] – [Action]
IMPORTANT (this week):
- [Sender]: [Subject] – [Why]
FYI (skim or archive):
- [Sender]: [Subject] – [Archive or read if curious]
UNSUBSCRIBE:
- [Sender]: [Reason]
CONSTRAINTS:
Err on the side of caution: if a deadline might be in the subject, mark it URGENT. If the sender is your boss or a key client, bump it up. Ask me if you see a pattern that doesn't match my actual priorities.
INPUT:
[Paste a list of unread subject lines and senders, roughly this format:
From: Sender Name | Subject line
From: Another Person | Another subject
...]
Context: What are my current top 3 priorities? (e.g., launch Q4 campaign, close hiring for Engineering, onboard new client)
How to fill it in
Grab a screenshot of your unread email list, or copy the sender and subject columns.
Tell the AI your actual priorities so it can filter by what matters to you.
Include the date if visible; recent emails are more likely urgent than old ones.
Then ask
Of the URGENT list above, which 3 should I handle first and why?
Draft replies to the top 3 URGENT emails in my voice.
Unsubscribe me from all FYI senders with a one-line reason.
Pro tip Do this triage weekly, not daily. If your inbox hits 100+ unread, you've lost signal. Use this to reset.
05
Write a follow-up that actually gets a response
Intermediatesaves 5–10 min, plus recovered answers
Use it when: You followed up on an important request 3+ days ago and haven't heard back. You need to nudge without sounding annoyed.
prompt.txt
ROLE:
You are a persistent but warm communicator who nudges without nagging.
CONTEXT:
Follow-ups get ignored because they sound needy. The best ones are brief, reference the original ask, assume the person is busy (not ignoring you), and make the next step tiny.
TASK:
1. Reference the original request by date or topic (proves you're following up, not sending a new ask)
2. Assume good intent: "I know you're slammed, so just a gentle bump on..."
3. Restate what you need in 1 sentence
4. Give them an easy out if they can't do it (suggest an alternative, delegate option, or deadline)
5. Lighten the tone with one authentic detail (avoid false casualness)
FORMAT:
[2–3 sentence follow-up email]
CONSTRAINTS:
Do not imply they're bad for not replying. Do not repeat the entire original ask. Do not send more than 2 follow-ups in 10 days. If a deadline is firm, state it clearly.
INPUT:
Original email sent [date]:
[Paste your original request]
Why do you need this? (decision you're waiting on, project deadline, other)
How urgent is this really? (today / this week / by end of month / no real deadline)
How to fill it in
Be honest about urgency; a fake deadline kills trust.
Mention what might be blocking them ('I know you mentioned [reason] last time') if relevant.
Give them a real alternative if it helps ('Or if this isn't a priority right now, just let me know and we can pause it').
Then ask
If I don't hear back in 2 days, write a final escalation to [person's boss].
Rewrite this to go to [different person] instead and explain why I'm looping them in.
Draft a fallback plan assuming I don't get this answer.
Pro tip The magic phrase is 'I know you're busy'—it's not patronizing, it's human. People reply faster when they don't feel guilty for being slow.
Watch out Don't follow up more than twice on low-stakes requests. If someone doesn't reply after 2 nudges, let it go or pick up the phone.
Meetings
Turn meetings from time wasters into decision engines. Create agendas that move, turn transcripts into action tables, and recap decisions so nobody has to wonder what happened.
06
Create a meeting agenda in 5 minutes
Beginnersaves 20–30 min of prep time
Use it when: You're running a meeting in an hour and need an agenda that actually moves decisions, not just lists topics.
prompt.txt
ROLE:
You are a meeting architect who designs agendas that surface decisions and move at the right pace.
CONTEXT:
Bad agendas list topics. Good agendas state the outcome for each topic and allocate time tight enough that people stay focused. A 30-minute meeting should have 3–4 topics, not 7.
TASK:
1. For each topic, state what decision or input you need (not just "discuss X")
2. Allocate time: longer for decisions, shorter for FYI items
3. Order by: urgency first, then dependencies (need decision A to discuss B)
4. Add a time buffer for the top decision if the meeting is longer than 30 minutes
5. End with a next steps call-out so someone owns follow-up
FORMAT:
[Meeting title] – [Date, time, duration]
[Minutes] | [Topic] | [Decision/input needed]
[Minutes] | [Topic] | [Decision/input needed]
...
Next steps owner: [name]
CONSTRAINTS:
Keep the agenda to 1 page. If someone will present, note that. If data is needed, call it out in advance so people come prepared. Ask me if the topics don't fit in the time.
INPUT:
Meeting: [Title]
Duration: [e.g., 30 min]
Attendees: [Names and roles]
Topics to cover: [List what you want to discuss]
Key decision needed: [What's the actual decision point?]
How to fill it in
State the real outcome you need; 'discuss hiring timeline' is vague. 'Approve 3-person team for Engineering' is clear.
If this is a standup or regular check-in, note that; the agenda can be lighter.
List any pre-reads or data people should bring.
Then ask
Send this agenda 24 hours before the meeting and ask people to come with input on decision 2.
If we run over on the first topic, write a 1-sentence summary and bump decision 2 to a 15-min follow-up.
After the meeting, turn the outcomes into action items.
Pro tip The tightest agendas list the decision first, then the topics that inform it. 'Approve Q4 budget (need Sales forecast and headcount plan)' moves faster than 'Sales forecast, Headcount, Budget.'
07
Turn meeting notes into an actions table in 60 seconds
Beginnersaves 10–15 min per meeting
Use it when: Your meeting transcript or notes are a wall of text and you need a clear table of who's doing what by when.
prompt.txt
ROLE:
You are a meeting analyst who extracts action items from messy notes and organizes them for clarity.
CONTEXT:
Meeting notes are gold but only if anyone reads them. An actions table forces clarity: who, what, by when. If an action is vague, flag it.
TASK:
1. Read the notes and extract all action items (stated or implied)
2. For each action, identify: owner (the person who will do it), task (specific, not vague), deadline (if stated), and priority (high/medium/low based on impact and deadline)
3. Flag any action that's unclear or has no clear owner
4. Group by deadline (due today, this week, next week, TBD)
FORMAT:
Create a table with columns: Owner | Action | Deadline | Priority | Notes
Order by deadline, then priority.
CONSTRAINTS:
If an action is ambiguous (e.g., "follow up on budget"), state exactly what 'follow up' means. If no owner was stated but it's obvious (e.g., Finance will review budget), assign it. If you really cannot determine who owns it, flag it as UNASSIGNED. Use exact dates, not "soon."
INPUT:
Meeting: [Title, date, attendees]
[Paste notes, transcript, or recording transcript]
How to fill it in
Paste the raw notes exactly as written; the AI can parse messy input.
If someone said 'I'll do it,' mention their name so the AI assigns it correctly.
Include any timestamps or context about priorities mentioned in the meeting.
Then ask
Send this table to all attendees with a note asking them to confirm ownership.
For the UNASSIGNED items above, who should own each one?
Which 3 actions are blockers for other work?
Pro tip The best actions tables are sent within 1 hour of the meeting. People remember what they said and will self-correct if needed.
Watch out If the meeting involved confidential business data (revenue, salary, names), redact before sharing the actions table with a large group.
08
Recap a meeting in writing so nobody has to wonder what happened
Beginnersaves 15–20 min of recap writing
Use it when: Your meeting just ended and you need a brief written recap for attendees and people who couldn't make it.
prompt.txt
ROLE:
You are a communications specialist who writes clear, brief meeting recaps that answer three questions: what happened, what's next, and what's changed.
CONTEXT:
Meeting recaps are sent to attendees (confirmation) and non-attendees (context). They should be skim-able in 2 minutes. Nobody wants a play-by-play; they want decisions, action items, and what it means for them.
TASK:
1. State the decision made (1 sentence, clear and specific)
2. List the reasoning in 2–3 bullets (why we chose this, what data informed it)
3. List action items with owners and deadlines
4. Note who's impacted and what changes for them
5. Mention the next milestone or review date
FORMAT:
Subject: [Meeting] Recap – [Decision summary]
Decision:
[1 sentence stating what was decided]
Reasoning:
- [Point 1]
- [Point 2]
Actions:
- [Owner]: [Task] by [date]
Impact:
[Who's affected and how, in 1–2 sentences]
Next review: [Date, if applicable]
CONSTRAINTS:
Do not include dissenting views unless explicitly asked. Keep tone neutral and professional. If a deadline is uncertain, say "TBD pending [blocker]."
INPUT:
Meeting: [Title, date, attendees]
Main decision: [What was decided?]
[Paste notes or transcript, or tell me what happened]
How to fill it in
State the decision clearly and specifically; 'We're moving forward with X' is clear. 'We decided to table this' is vague.
If there were strong opinions against the decision, decide whether to mention them or handle separately.
Include timeline or dependencies that people need to know.
Then ask
Make this recap read a bit more formal / casual (depending on your team culture).
Add a section for people who want more detail: 'Full notes: [link]'.
Send this to [broader team] with a note that [decision] affects their [area].
Pro tip Send recaps within 4 hours of the meeting. The longer you wait, the less useful they are because people have moved on.
09
Prep stakeholders before a big meeting
Intermediatesaves 20–30 min of prep thinking
Use it when: You have a meeting with a key decision-maker or stakeholder and you want to set them up to say yes.
prompt.txt
ROLE:
You are a strategic communicator who helps prepare a stakeholder for a decision meeting by priming them for success.
CONTEXT:
The best meeting outcomes are decided before the meeting starts. You're writing a memo that surfaces the key issues, aligns on priorities, and makes the decision seem obvious.
TASK:
1. Identify what the stakeholder cares about most (budget, risk, timeline, impact)
2. Build a narrative that connects your proposal to their priorities
3. Anticipate and address their likely objections with data or logic
4. Propose the decision you want them to make, assuming they agree
5. Ask for any input they want to give before the meeting
FORMAT:
Brief memo (3–4 paragraphs):
- What we're deciding and why timing matters now
- Why this aligns with [their priority] and how it impacts their goals
- Quick evidence that this is the right move (1–2 key data points or comparisons)
- What we need from them in the meeting
- Open question: anything they'd like us to address beforehand?
CONSTRAINTS:
Be honest about tradeoffs. Do not oversell or hide risks. Frame the decision as a choice between real alternatives, not as inevitable. Ask clarifying questions if you don't know their priorities well.
INPUT:
Meeting: [Title, date, stakeholder name]
Decision you want: [Specific: e.g., "Approve a 50k budget for tools" or "Delay launch to March"]
Stakeholder's role: [e.g., CFO, VP of Ops, Board Chair]
Their top 3 priorities: [e.g., "cost control, speed to market, risk management"]
Their likely objections: [What might they say no to?]
Key evidence (data, comparisons, case studies): [Paste or list]
How to fill it in
Research their actual priorities; ask peers if unsure. CEO cares about growth, CFO about runway, Head of Ops about team capacity.
Include 1–2 pieces of evidence: a comparison, a case study, a cost-benefit analysis, or a quote from a trusted peer.
Frame objections as 'I anticipate you might wonder...' rather than 'You'll probably object to...'
Then ask
Rewrite this to emphasize [their priority] more and tone down [another area].
Add a data point showing [competitor / similar company] did this and saw [result].
Soften the opening; this is too aggressive.
Pro tip The best stakeholder memos are specific to the person. Generic memos get skimmed and ignored. What does this stakeholder actually care about?
10
Write a decision memo that settles a debate
Intermediatesaves 45–60 min of decision drafting
Use it when: You need to make a complex decision with competing options and you want to document it clearly so the team moves forward together.
prompt.txt
ROLE:
You are a decision analyst who structures complex choices so the right path becomes clear.
CONTEXT:
Good decision memos aren't arguments; they're frameworks. They lay out what matters, compare options fairly, and explain the choice in a way the team can accept even if they disagree.
TASK:
1. State the decision needed and why timing matters (deadline, risk, opportunity cost)
2. List the criteria that matter most (e.g., speed, cost, quality, team bandwidth)
3. Compare 2–3 realistic options against those criteria (use a table or bullets)
4. State your recommendation and the logic (which criteria mattered most, why other options rank lower)
5. Name the key risks of your choice and how you'll mitigate them
6. Invite dissent: offer a clear process for if the team disagrees
FORMAT:
Title: Decision Memo – [Decision]
Decision: [What choice are we making?]
Deadline: [When do we need to decide?]
Why now: [What changes if we delay?]
Criteria (ranked by importance):
1. [Criterion 1] – [Why it matters]
2. [Criterion 2] – [Why it matters]
Options analysis:
[Table comparing options against criteria, or bullet bullets for each option]
Recommendation: [Option X], because [primary reasons].
Risks and mitigants:
- Risk: [Specific downside] | Mitigate: [How]
Dissent process: If you disagree, reply with [one specific alternative] or [the criteria you'd weight differently]. We'll discuss on [date].
CONSTRAINTS:
Be fair to options you're not recommending; don't strawman them. Use data or past examples where possible. If you're uncertain about a factor, say so. Avoid presenting personal preference as objective fact.
INPUT:
Decision: [What are we choosing?]
Timeline: [Deadline]
Options: [List 2–3 real alternatives, with brief description of each]
Criteria: [What matters? e.g., cost, speed, quality, team capacity]
Evidence: [Data, case studies, examples that inform the choice]
Who do you need to convince? (team / board / key stakeholder)
How to fill it in
List the actual options people are discussing, not strawman versions. If someone advocates for Option C, understand it fairly.
Weight the criteria honestly. If you say cost and speed matter equally but cost is really your priority, the memo will feel manipulative.
Include 1–2 precedents or data points that inform the choice (past projects, industry benchmarks, similar companies).
Then ask
Rewrite this to emphasize [criterion] more heavily; it's our actual constraint.
Add a section comparing this to how we handled [similar decision] last year.
For the key risk you mentioned, what's our contingency plan if it happens?
Pro tip The best decision memos end with an invite to dissent. 'If you disagree, tell me what criteria I weighted wrong.' That signals you're open to being wrong, which gets the team's actual input.
Writing & Content
Write faster, better, and with less back-and-forth. Draft from rough notes, edit without rewriting, and turn raw material into polished content in half the time.
11
Draft polished writing from messy bullet points
Beginnersaves 20–30 min of writing time
Use it when: You have rough notes, an outline, or scattered thoughts and need them shaped into a draft email, article, or memo.
prompt.txt
ROLE:
You are a skilled writer who transforms rough notes into polished prose while preserving the author's voice and intent.
CONTEXT:
Great writing flows. Bullets are scaffolding. You'll take loose ideas, order them logically, add transitions, and clean up the grammar while keeping the original thinking intact.
TASK:
1. Read the notes and understand the core message
2. Organize into a logical sequence (problem, solution, evidence, action or similar)
3. Write in complete sentences with smooth transitions
4. Preserve any specific phrases or examples from the original notes
5. Match the formality level requested (email, article, memo, etc.)
FORMAT:
[Polished draft, ready to paste and edit minimally]
CONSTRAINTS:
Do not change the meaning or add ideas not in the original. If a bullet is unclear, ask for clarification rather than guessing. Keep sentences short and active. Flag anything that contradicts other points.
INPUT:
Format: [email / article / memo / other]
Length: [e.g., 3 paragraphs, 200 words, 1 page]
Tone: [formal / casual / persuasive / explanatory / other]
Original notes:
[Paste your bullets, outline, or rough notes here]
How to fill it in
Don't clean up your notes first; paste them as-is. The AI will parse them.
If you have a specific phrase or example you want included, put it in quotes or highlight it.
State your tone: 'like we're talking to peers' is different from 'like we're pitching investors.'
Then ask
Make the opening sentence punchier; it's too abstract.
Add an example of [specific topic] from [context].
Tighten the second paragraph; it's wandering.
Pro tip Paste your notes in the order you thought of them, not the order they should appear. Let the AI reorganize. You're less attached to the structure than you are to the ideas.
12
Edit without rewriting—show every change and why
Beginnersaves 15–20 min of editing back-and-forth
Use it when: You have a draft and want feedback on specific improvements, with a clear explanation of each change so you can learn or push back.
prompt.txt
ROLE:
You are a developmental editor who improves writing one change at a time, explaining the reasoning so the author can learn and decide.
CONTEXT:
Rewritten drafts feel like someone else wrote them. Phrase-level feedback preserves authorship while improving clarity, rhythm, and impact. You'll suggest specific rewrites with reasons.
TASK:
1. Read the draft and identify 3–5 strongest improvements (not 20 nitpicks)
2. For each change, show the original phrase and your suggested version
3. Explain why in 1 sentence (clarity, rhythm, impact, word choice, tone)
4. Offer your suggestion as an option, not a mandate ('Consider...', 'You might...')
5. Flag tone issues or contradictions, but ask before changing voice
FORMAT:
1. Original: [Phrase]
Suggestion: [Improved version]
Why: [One sentence reason]
2. Original: [Phrase]
...
Questions:
[Anything unclear in the draft or worth discussing]
CONSTRAINTS:
Do not rewrite paragraphs; suggest changes at phrase or sentence level. Do not change voice or add new ideas. If you find more than 5 strong improvements, pick the ones with highest impact (clarity, persuasion, readability). Ask if you're unsure about the author's intent.
INPUT:
[Paste the draft]
Context: What's the goal of this piece? (persuade, inform, build credibility, entertain, other)
Audience: [Who is reading this?]
Anything you're worried about: [Specific passages or tones you want feedback on]
How to fill it in
Paste the full draft, even if it's rough. The AI needs to see it all.
State the goal clearly: 'This should convince [person] to [action]' guides better feedback than 'make it better.'
Mention anything you're unsure about; the AI will focus there.
Then ask
Incorporate all 5 suggestions above and read it aloud to me (paste the edited version).
For change #3 above, I'm not sure it's better. Can you explain the rhythm change?
Any other phrases that feel stiff or unclear?
Pro tip Phrase-level feedback takes longer to explain but makes you a better writer. You see why the change works, not just that something is 'better.'
13
Simplify jargon into plain language that anyone understands
Beginnersaves 15–25 min of rewriting
Use it when: Your writing is full of industry jargon, corporate speak, or complex sentences that make it hard for non-experts to understand.
prompt.txt
ROLE:
You are a plain-language editor who makes complex ideas accessible without dumbing them down.
CONTEXT:
Jargon feels professional but confuses readers. Plain language is actually harder to write because it requires knowing the idea deeply enough to explain it simply. You'll replace jargon and untangle complex sentences.
TASK:
1. Identify jargon, acronyms, and complex phrases
2. Replace with simple equivalents (use a 10th-grade reading level as a target)
3. Break long sentences into shorter ones
4. Preserve the rigor and accuracy; don't oversimplify
5. Keep the same tone (formal memos stay formal, emails stay casual)
FORMAT:
[Fully rewritten version in plain language, ready to paste]
CONSTRAINTS:
Do not sacrifice accuracy for simplicity. If a term is standard in the field, keep it but explain it once in simple terms. Ask me if the tone should shift (e.g., if this is for exec leadership vs. a broader employee audience).
INPUT:
[Paste the draft]
Current audience: [Who is reading this?]
What's the core idea you're explaining: [In your own words, simply]
Tone: [Formal / casual / other]
How to fill it in
Mark any jargon you know is unclear, or let the AI find it.
State the core idea in your simplest language; that's the target for the AI.
If you must keep certain technical terms, highlight them and note why.
Then ask
Can you explain what [jargon term] means in one simple sentence?
This explanation is good but the second paragraph still feels technical.
Keep the term '[industry specific]' but add a one-sentence definition.
Pro tip Read the plain-language version aloud. If you stumble over a sentence, it's still too complex.
14
Write headlines and hooks that actually get clicked
Intermediatesaves 10–15 min per headline
Use it when: You're writing email subject lines, social media posts, or article headlines and want them to stand out without clickbait.
prompt.txt
ROLE:
You are a copywriter who writes headlines and hooks that make people want to read more.
CONTEXT:
Good headlines are specific, promise a benefit or insight, and hint at something unexpected. Clickbait feels cheap. The sweet spot is intriguing without being manipulative.
TASK:
1. Understand the actual benefit or insight (not the surface topic)
2. Generate 5–7 headline options that vary in approach (curiosity hook, benefit-led, specific/surprising, question, contrast, etc.)
3. For each, note the approach and why it works
4. Recommend the strongest 1–2 based on your audience and goal
FORMAT:
Option 1: [Headline]
Approach: [Curiosity / Benefit / Question / Contrast / Specific insight / Other]
Why: [Why this works]
Option 2: [Headline]
...
Top recommendation: [Option X], because [audience insight].
CONSTRAINTS:
Avoid superlatives (Amazing, Incredible, Shocking). Use specific numbers and insights instead. Do not promise something the content doesn't deliver. Match the tone and platform (Twitter is different from email subject lines).
INPUT:
What's the content about? [Topic, main idea]
What's the real benefit or insight? [What's worth reading?]
Audience: [Who's reading? What do they care about?]
Platform/medium: [Email subject, social post, article, etc.]
Tone: [Professional / casual / edgy / other]
What would make YOU click? [Your honest take]
How to fill it in
Be specific about the benefit: 'How to write better emails' is vague. 'Why your emails aren't getting replies (and how to fix it)' is specific.
Name your actual audience: developers, small business owners, parents, etc. They respond to different hooks.
Note any constraints: character limit for social, word limit for headlines, etc.
Then ask
Generate 5 more options focusing on [specific angle, e.g., fear, curiosity, novelty].
Which of these would resonate most with [audience segment]?
Rewrite option 3 to be less clickbait-y but keep the intrigue.
Pro tip The best headlines promise something specific and true. Test 2–3 variations if you can; use the one that gets the highest click rate.
15
Repurpose one piece of content into three formats
Intermediatesaves 30–40 min of repurposing
Use it when: You've written something valuable and want to turn it into email, social posts, and a short article without writing from scratch three times.
prompt.txt
ROLE:
You are a content strategist who finds the core idea in one piece of content and adapts it for different formats and audiences.
CONTEXT:
Repurposing isn't copying. It's extracting the key insight and reshaping it for the medium (thread vs. email vs. article) and the audience's context (skimming vs. reading vs. listening).
TASK:
1. Read the original and extract the core insight (the one thing people should remember)
2. Identify the strongest supporting points
3. Adapt for three formats: [email / social post / short article / other]
4. For each, adjust depth, tone, and structure (email needs a CTA, social needs brevity and hooks, articles need more context)
5. Do not duplicate text across formats; each should feel native to the medium
FORMAT:
Format 1: [Email version - 100–150 words with subject line and CTA]
Format 2: [Social post version - under 280 characters or 1–2 short paragraphs]
Format 3: [Article/blog version - 300–400 words]
CONSTRAINTS:
Do not change the core idea. Do not pad formats to hit a word count. Social posts should be snackable. Emails should have a clear CTA. Articles should add context or examples the original might not have. Ask me if formats don't fit the content.
INPUT:
[Paste the original content]
Core insight: [In one sentence, what's the takeaway?]
Goal for each format: (reach, persuade, educate, entertain)
Distribution plan: [Where will each version go? e.g., newsletter, Twitter, blog]
How to fill it in
Paste the original content exactly as-is.
State the core insight in one sentence; everything else should support it.
Note where each version will be published; that guides tone and depth.
Then ask
Turn these three into a weekly content calendar (give me a template).
Expand the article version to 600 words and add one more example.
Shorten the social post to fit a tweet thread format.
Pro tip Your best repurposing insight came from someone who noticed you repeated the same idea in different contexts. Pay attention to what you keep saying.
Research & Decisions
Make better decisions faster by researching thoroughly, comparing options fairly, and stress-testing your thinking before you commit.
16
Get a clear explanation of any complex topic in 5 minutes
Beginnersaves 30–45 min of research and reading
Use it when: You need to understand a concept, trend, or technology quickly so you can make a decision or talk about it confidently.
prompt.txt
ROLE:
You are an expert explainer who makes complex topics understandable to smart people who aren't experts in that field.
CONTEXT:
Good explanations start with the core idea, use an analogy or example, then zoom out to context and implications. Bad ones either oversimplify or pile on jargon.
TASK:
1. Explain what this is in one clear sentence
2. Use an analogy or real-world example that makes it intuitive
3. List 3–4 key points that matter (not everything, just what's important to know)
4. Address: where is this used, who cares, what's the upside / risk
5. Flag any debate or uncertainty in the field
FORMAT:
What it is: [One sentence]
Analogy: [How to think about it]
Key points:
- [Point 1]
- [Point 2]
- [Point 3]
Why it matters: [Where it's used, who cares, upside/risk]
Debate: [If experts disagree, note that]
CONSTRAINTS:
Use plain language. If you need a technical term, explain it once. Do not oversimplify so much that the explanation is wrong. If this is a new or fast-moving area, note what you're certain about vs. uncertain about.
INPUT:
Topic: [What do you want to understand?]
Context: [Why do you need to know? What decision does this inform?]
Your background: [What do you already know about? What should I not explain?]
How to fill it in
Be specific: 'machine learning' is too broad. 'How does ChatGPT know when to stop generating text?' is specific.
State your actual context; if you're evaluating a vendor, say so. That shapes what's important to know.
Mention what you already understand; the AI won't waste time on basics.
Then ask
Explain [subtopic from above] deeper.
Give me 3 good sources to learn more about this.
How does this differ from [related concept]?
Pro tip The best explanations start with an analogy. 'Think of it like [familiar thing]' makes even complex ideas stick.
17
Compare options using a criteria table
Intermediatesaves 45–60 min of comparison and spreadsheet building
Use it when: You're choosing between vendors, tools, approaches, or investments and want to compare them fairly against what actually matters.
prompt.txt
ROLE:
You are a decision analyst who helps people compare options fairly by making the criteria explicit and visible.
CONTEXT:
The best comparisons surface tradeoffs. By weighting criteria, you see that Option A is better for speed but worse for cost. That clarity beats gut feel.
TASK:
1. Identify the criteria that matter most (not all possible criteria, just the important ones)
2. Weight them: which matters more, price or features? Speed or reliability?
3. Research or gather what you know about each option against each criterion
4. Create a comparison table
5. Note where you're certain vs. where you're guessing
6. Summarize which option ranks highest overall
FORMAT:
Criteria (weighted 1–5 stars):
- [Criterion]: [Why it matters]
- [Criterion]: [Why it matters]
Comparison table:
[Table with options as rows, criteria as columns, scores/notes]
Overall ranking: [Option X] scores highest, [Option Y] has an edge in [specific area]
Tradeoffs to consider:
- [If you choose X, you gain [benefit] but lose [tradeoff]]
Uncertainties:
- [What we don't know or haven't tested]
CONSTRAINTS:
Weight criteria by importance, not by how easy they are to measure. Be honest about missing information. Do not force a winner if the options are genuinely tied. Ask clarifying questions if criteria conflict.
INPUT:
Decision: [What are you choosing between?]
Options: [List 2–4 options you're considering]
Criteria that matter to you: [e.g., cost, speed, ease of use, reliability, support, flexibility]
What you already know: [Facts, pricing, experiences you've had]
Questions:
- What's more important to you: [Criterion A] or [Criterion B]?
- Are there any deal-breakers? (e.g., "Must integrate with [system]")
How to fill it in
List your actual criteria, not generic ones. 'Cost' is vague. 'Fits within a 50k budget' is clear.
Weight the criteria honestly. If speed matters more than cost, say so. Don't pretend they're equal.
Include any constraints: 'Must work with [system]' or 'Team is already trained on [tool].'
Then ask
Can we get a 30-day trial of Option A to test [criterion] before deciding?
What's the contract commitment for each option? (1-year, month-to-month, etc.)
If we choose Option X, what's the migration cost to switch later?
Pro tip The comparison table is useful, but the tradeoffs section is gold. That's where you see what you're actually giving up.
18
Fact-check your own thinking before you decide
Intermediatesaves 20–30 min of second-guessing and preventing regret
Use it when: You're about to make a decision but you want to sanity-check your reasoning and spot any flawed assumptions before you commit.
prompt.txt
ROLE:
You are a critical thinking partner who asks hard questions and flags flawed reasoning without being mean about it.
CONTEXT:
Confirmation bias makes us believe our own thinking. A good critic asks 'What if you're wrong?' and 'What would change your mind?' without making you defensive.
TASK:
1. Listen to the decision and the reasoning
2. Identify the core assumptions (often unstated)
3. For each assumption, ask: Is this definitely true? What if it's false?
4. Surface any tradeoffs the person might not have considered
5. Note whether the decision is reversible (if not, more caution is warranted)
6. Offer a specific question or test that would falsify the thinking
FORMAT:
Your reasoning:
[Summary of the decision and why you think it's right]
Assumptions I spot:
1. [Assumption 1]: What if this is false? [consequence]
2. [Assumption 2]: What if this is false? [consequence]
Tradeoffs you might not have mentioned:
- [Tradeoff 1]
- [Tradeoff 2]
Reversibility: [Is this decision easy to undo? If yes, lower stakes. If no, higher stakes.]
The question that would prove you wrong:
[One specific, testable question]
CONSTRAINTS:
Be respectful. This is about testing thinking, not attacking the decision. Focus on assumptions and logic, not gut feel. If the person has already considered something, acknowledge it.
INPUT:
Decision you're about to make: [What are you deciding?]
Why you think it's right: [Your reasoning, assumptions, evidence]
What would make you change your mind: [If anything]
Timeline: [When do you need to decide?]
How to fill it in
State your reasoning out loud. Write down what you think is true and why.
Be honest about your confidence level. 'I'm 80% sure' guides different follow-ups than 'I'm 100% sure.'
If there's a sunk cost involved ('We've already spent 30k'), mention it so the AI can account for it.
Then ask
I think assumption 2 is actually true because [evidence]. Does that change your critique?
Can we test [the question above] before we commit fully?
What's the worst-case outcome if assumption 1 turns out to be false?
Pro tip The best red team is someone who doesn't have skin in the game. If you can't find one, ask the AI to be meaner. Tell it to actively try to poke holes.
19
Build a pre-mortem: what could go wrong?
Intermediatesaves 1–2 hours of future regret and crisis management
Use it when: You're about to launch, hire, ship, or commit to something big and you want to stress-test it by imagining it fails.
prompt.txt
ROLE:
You are a pre-mortem facilitator who runs a hypothetical failure analysis. Your job is to imagine this thing failed and work backward to find the cracks.
CONTEXT:
A pre-mortem is not pessimism; it's realism. By imagining failure now, you spot risks while you can still do something about them. Research shows pre-mortems improve decision quality.
TASK:
1. Imagine it's one year from now and the initiative failed spectacularly
2. Work backward: what went wrong? (List 5–7 plausible failure modes)
3. For each failure mode, trace the cause (what could cause this to happen)
4. Rate by likelihood and impact
5. For the top 3 risks, suggest a mitigation (how to prevent or reduce this risk)
FORMAT:
Initiative: [What are you doing?]
Pre-mortem findings:
1. [Failure mode]: [What went wrong?]
Cause: [What led to this?]
Likelihood: High / Medium / Low
Impact if it happens: [What would be the consequence?]
Mitigation: [How to prevent or reduce this risk?]
2. [Next failure mode]...
Top 3 risks to address before launch:
- [Risk 1]: Mitigation [how to reduce]
- [Risk 2]: Mitigation [how to reduce]
- [Risk 3]: Mitigation [how to reduce]
CONSTRAINTS:
Focus on realistic failure modes, not extreme black swans. Be specific (not 'people won't like it,' but 'the onboarding flow is too complex and 40% of users drop off before activation'). Ask clarifying questions if you don't know the context well.
INPUT:
What are you launching, hiring, or committing to: [Description]
Success looks like: [What do you hope happens?]
Timeline: [When do you launch?]
What scares you most: [Your gut fear, if you have one]
Key stakeholders who need to buy in: [Who's dependent on this succeeding?]
How to fill it in
Be specific about the initiative. 'Launch a feature' is vague. 'Launch the mobile app to new geographic regions' is specific.
State what success looks like; it helps the AI imagine realistic failure modes.
Mention your gut fear. Pre-mortems often validate gut instincts you're trying to suppress.
Then ask
For risk 2 above, what's a cheap test we could run to catch this before launch?
Which of these risks is actually a dealbreaker (must solve before we go) vs. acceptable (can manage if it happens)?
Can we do a smaller pilot to stress-test risks 1 and 3?
Pro tip The pre-mortem is most useful 4–6 weeks before launch, when you still have time to fix things but you're deep enough to see real risks.
20
Get a research brief you can actually trust
Intermediatesaves 1–2 hours per topic
Use it when: You need to get up to speed on a market, competitor, vendor or trend before a meeting or decision.
prompt.txt
ROLE:
You are a senior research analyst who writes concise, well-sourced briefs for busy decision-makers.
CONTEXT:
I need to understand [topic, company or market] because [the decision or meeting this is for]. My audience is [who will read this, e.g. my leadership team]. What I already know: [1–3 bullets, or "nothing yet"].
TASK:
1. Summarize the current state of [topic] in plain language.
2. List the 5 most important facts or numbers, each with where it comes from.
3. Identify the 3 key players or options and how they differ.
4. Note what is changing right now and what to watch over the next 6–12 months.
5. List the open questions I should ask or verify before deciding.
FORMAT:
A one-page brief with these headings: Bottom line (2 sentences), Key facts, Players, What is changing, Open questions. Bullets, not paragraphs. Under 400 words.
CONSTRAINTS:
Separate verified facts from your own estimates and label estimates clearly. If you are not certain a fact is current, say so and suggest how to check it. Do not invent sources, statistics or quotes. Ask me up to 3 clarifying questions first if the topic is ambiguous.
INPUT:
[Optional: paste articles, notes or links you want included]
How to fill it in
Name the real decision this research supports; it focuses the brief on what matters.
Paste any articles or reports you trust so the AI works from them instead of memory.
Use a tool with web search (ChatGPT search, Claude with web, Perplexity, Gemini) for anything recent.
Then ask
Which of these facts are you least confident about, and how would I verify each one?
Turn this into 5 questions I should ask [vendor / expert / my team] in our meeting.
Rewrite the bottom line for a skeptical CFO in 3 bullets.
Pro tip Ask for the brief first, then ask the AI to grade its own confidence in each fact from 1–5. The low scores tell you exactly what to double-check.
Watch out AI can state outdated or made-up numbers confidently. Verify every statistic you plan to repeat in a meeting or put in writing.
Planning & Productivity
Turn chaos into a plan. Break overwhelming projects into sequences, write playbooks so work repeats smoothly, and build habits that stick.
21
Distill a document into an executive summary
Beginnersaves 20–30 min of summary writing
Use it when: You have a long report, proposal, or memo and need a 1–2 page summary for busy leaders who won't read the whole thing.
prompt.txt
ROLE:
You are an executive summarizer who extracts the most important information from long documents.
CONTEXT:
Executives need three things: what is this, why does it matter, and what do you want from me. Everything else is supporting detail. A great summary is 80% shorter than the original but answers those three questions completely.
TASK:
1. Identify the core recommendation or finding (1 sentence)
2. State why it matters (the business implication, problem it solves, opportunity it represents)
3. List the top 3–4 evidence points or data supporting the recommendation
4. State what decision or action is needed and by when
5. Link back to the original document for anyone who wants to dig deeper
FORMAT:
Executive Summary: [Title]
Recommendation: [What we're proposing or finding]
Why it matters: [Business impact, problem solved, opportunity]
Evidence:
- [Key data point 1]
- [Key data point 2]
- [Key data point 3]
What we need: [Decision / action needed, by when]
Full document: [Link]
CONSTRAINTS:
Use numbers and specifics, not vague language. Do not oversimplify or change the meaning. If the original has dissenting views, mention them if they're material to the decision. Ask me if the main recommendation is unclear.
INPUT:
[Paste the full document or report]
Who is the audience for this summary: [e.g., board, CFO, executive team]
What's the main ask or recommendation: [In your words]
Any context they should know: [What's the background?]
How to fill it in
Paste the full document; don't summarize it first. The AI will find the key points.
Tell the AI who's reading it. A board wants different info than a CFO.
State your main ask clearly: 'Approve a 500k budget' or 'Delay launch to Q2' is clear.
Then ask
Make this 1 page instead of 2; cut the detail and keep only the decision points.
Add a costs / benefits comparison table.
This is for a [different audience]; rewrite it for them.
Pro tip The best executive summaries fit on one page. If you need more, you haven't distilled enough.
22
Build a weekly plan from a chaos list
Intermediatesaves 30–45 min of planning and decision-making
Use it when: You have 20+ tasks and meetings for the week and need to prioritize ruthlessly and sequence them so you actually get important work done.
prompt.txt
ROLE:
You are a time architect who builds weekly plans that balance urgent work, important work, and the space to think.
CONTEXT:
The best weekly plans are not packed. They have white space for unexpected issues, focus time for important work, and meetings clustered so deep work has protection. You'll help sort by urgency and importance, then schedule realistically.
TASK:
1. Separate tasks by urgency (today, this week, later) and importance (moves the needle vs. housekeeping)
2. Identify deep work (tasks requiring focus: 2+ hours uninterrupted) vs. shallow work (emails, meetings, admin)
3. Group meetings and shallow work to protect deep work blocks
4. Build a realistic weekly schedule that includes: (1) top 3 priorities, (2) urgent items, (3) at least one deep work block
5. Highlight risks: What might not fit? What's the contingency?
FORMAT:
This week's priorities:
1. [Top priority and why]
2. [Second priority]
3. [Third priority]
Weekly schedule (draft):
Monday: [Morning] ... [Afternoon] ...
Tuesday: [Deep work block: 2 hours for X]
...
Risks and contingencies:
- If [urgent issue] comes up, we'll drop [task] to make room.
Backlog (will fit next week if we stay on track):
- [Task 1]
- [Task 2]
CONSTRAINTS:
Do not schedule more than 6 hours of meetings per day (leaves time for work). Identify the task that adds most leverage, and protect focus time for it. If something is both urgent and important, do it early in the week. Ask me if the priorities conflict or if something has a hard deadline you haven't mentioned.
INPUT:
Your full task list for this week: [Paste everything: tasks, meetings, deadlines]
Top 3 priorities: [What would you say matters most?]
Fixed meetings: [e.g., "Monday 10am standup, Wednesday 2pm board prep"]
Deep work needed: [Any tasks requiring 2+ uninterrupted hours?]
Constraint: [Any fixed limitations, e.g., "Friday is light," "Afternoon meetings only"]
How to fill it in
Paste your full task list exactly as you have it; don't pre-sort.
State your top 3 priorities clearly. If you're not sure, pick the one that has the biggest impact if not done.
List any hard deadlines (client calls, meeting preps, deliverables).
Then ask
This assumes everything gets done. What's the priority if we only finish the top 7 items?
Where's the white space if something urgent comes up mid-week?
Which 3 tasks could we move to next week without consequence?
Pro tip The best weekly plan has at least one deep work block (2+ hours uninterrupted) for your most important task. Block it like a meeting or it won't happen.
23
Break down an overwhelming project into a sequence
Intermediatesaves 1–2 hours of project planning
Use it when: You have a big, vague project and you don't know where to start. You need a step-by-step plan with milestones and dependencies.
prompt.txt
ROLE:
You are a project planner who breaks big goals into bite-sized steps and surfaces dependencies.
CONTEXT:
Overwhelming projects feel unmanageable because they're too abstract. Once you sequence the work, list what depends on what, and identify milestones, it becomes executable.
TASK:
1. Take the goal and work backward to the first step (what must happen first?)
2. List all steps in sequence (you can't do step 5 before step 3 if 3 unlocks 5)
3. Identify dependencies (what's waiting on what?)
4. Group into phases and add milestones (completion markers)
5. Estimate rough effort for each step so you can plan realistic timeline
6. Highlight risks: What could block progress?
FORMAT:
Project: [Goal]
Timeline estimate: [Based on effort below]
Phase 1: [Name]
- Step 1: [Task] [effort: 2 days] [owner: if known]
- Step 2: [Task] [effort: 1 day] Depends on Step 1
Milestone: [Completion marker]
Phase 2: [Name]
...
Dependencies graph:
Step 1 → Step 2 → Step 4
Step 3 → Step 4
Risks:
- [What could block progress] Mitigation: [How to prevent]
CONSTRAINTS:
Estimate effort realistically, including time for review and rework. Break tasks to the level of 'I can do this in one session.' If you don't know effort, guess and note uncertainty. Ask me if the goal is vague or if dependencies aren't clear.
INPUT:
Project: [What are you building / launching / completing?]
Goal: [Specific outcome you want]
Timeline: [When do you need it done?]
Constraints: [Team size, resources, dependencies on others]
Any pieces you've already started: [Work already done that informs this project]
How to fill it in
Be specific about the goal. 'Rebuild the website' is vague. 'Migrate the website to Webflow and update all [X] pages with [Y]' is specific.
State your timeline honestly. If you don't know, say 'No hard deadline but sooner is better.'
Mention who's involved. 'Just me' shapes the plan differently than 'Me + designer + developer.'
Then ask
Which 3 steps are on the critical path (must be done to unblock others)?
If we have 10 hours per week, how many weeks will this take?
What's the minimum viable version we could do in half the time?
Pro tip The critical path is your friend. Focus on steps that unblock other work. Doing the hard-but-fast step first moves the whole project forward.
24
Document a process so someone else can do it
Intermediatesaves 30–45 min of future repetition and training someone new
Use it when: You do something regularly and you want to write it down so someone else can do it, or so you don't have to think about it next time.
prompt.txt
ROLE:
You are a process documenter who writes Standard Operating Procedures (SOPs) that are so clear a new person can follow them.
CONTEXT:
Good SOPs are step-by-step, specific, and account for edge cases. They save time, reduce errors, and make it easier to hand off work.
TASK:
1. Talk through the process step-by-step (what do you do first, second, etc.?)
2. For each step, note: what tool or system, what decision point, what could go wrong
3. Highlight any judgment calls (steps where you need to think, not just follow)
4. Add a checklist for pre-work (what needs to be true before starting?)
5. Add a checklist for post-work (how do you know it's done right?)
FORMAT:
Process: [Name]
Purpose: [Why is this done?]
When to use this: [Trigger or schedule]
Time needed: [Estimate]
Pre-work checklist:
- [ ] [Requirement 1]
- [ ] [Requirement 2]
Steps:
1. [Action]: [Use tool X to...]. If [edge case], then [alternate step].
2. [Action]: ...
Decision points:
- If [condition], do [X]. If [other condition], do [Y].
Post-work checklist:
- [ ] [Verification 1]
- [ ] [Verification 2]
Common mistakes:
- [Mistake 1]: How to avoid it.
CONSTRAINTS:
Be specific. 'Update the spreadsheet' is vague. 'Open [file name], go to tab 'Q4 Sales', update row 3 columns B–D with [data]' is clear. Include screenshots or links if helpful. Flag any steps that require special access or training.
INPUT:
Process: [What do you want to document?]
How often do you do this: [Daily / weekly / as needed]
Who might do this next: [Yourself later, a new hire, a team member?]
Walk me through the process step-by-step: [Paste your explanation, messy is fine]
How to fill it in
Explain the process out loud as if teaching someone. That's easier than writing it. Paste exactly what you said.
Include edge cases: 'Usually we do X, but if [situation], we do Y instead.'
Mention any passwords, logins, or special access needed so new people know what they're missing.
Then ask
Add a screenshot of [step 3] so new people can see what it looks like.
Is there a template we should use for [part of the process]?
Test this with [team member] to see if it's clear.
Pro tip The best SOPs are version-numbered and updated when the process changes. Add 'Last updated [date]' at the top so people know if it's stale.
25
Turn a goal into a habit that sticks
Intermediatesaves 60+ min of willpower and intention-setting
Use it when: You have a goal (exercise, learn something, focus on strategic work) and you want to make it a habit so it happens automatically.
prompt.txt
ROLE:
You are a habit designer who builds systems that make behavior automatic rather than dependent on willpower.
CONTEXT:
Willpower fails. Systems work. Good habit design links the new behavior to an existing routine, makes it easy, and creates quick wins. You'll help design a system that sticks.
TASK:
1. Identify a current habit or routine you do every day (anchor habit)
2. Link the new behavior to it (if you do X, then immediately do Y)
3. Make the first version easy (lower the bar for starting)
4. Identify friction points that might block it
5. Design a tiny reward or verification (how do you know you did it?)
6. Plan a review cycle (check in weekly, adjust if needed)
FORMAT:
Goal: [What do you want to be true in 3 months?]
Anchor habit: [What do you do every day? e.g., "Drink my morning coffee"]
New habit design:
- Trigger: [After I X, I will immediately Y]
- Minimum version: [Easiest version of this, e.g., "5 minutes, not 30"]
- Location and tools: [Where and with what]
- Verification: [How do you know you did it?]
Friction points and solutions:
- Friction: [What might block this] Solution: [How to reduce friction]
Review cycle:
- Check in every [frequency] on [day]: Is this working? Do I need to adjust?
- If I miss [X days], I'll reset and recommit.
CONSTRAINTS:
Start stupidly small. A 5-minute habit you do is better than a 30-minute habit you skip. Ask me if the goal is vague or if the anchor habit doesn't exist yet.
INPUT:
Goal: [What behavior do you want to establish?]
Timeline: [How long to establish it?]
Current habits: [What do you do every day without thinking? (coffee, shower, commute, etc.)]
Why it matters: [What changes if you do this for 3 months?]
What's blocked you before: [If you've tried this goal before, what stopped you?]
How to fill it in
Pick an anchor habit you already do daily. It's the most reliable trigger.
Start small: 5 minutes is better than 30. You can increase later.
Be specific about the trigger: 'After I close my laptop at 5pm, I go for a 5-minute walk' is better than 'Exercise more.'
Then ask
What's my minimum version if I'm tired or pressed for time?
Can we set up a [calendar reminder / app notification] to trigger this until it's automatic?
After 3 weeks, I'll check in: did this stick? Do we need to adjust?
Pro tip The first week is about showing up, not perfecting. A tiny win day 1 is better than skipping because you didn't have 30 minutes.
Want prompts built for your exact job?
In 90 minutes live, we'll set up your AI properly and build prompts around your real work — with feedback on the spot.