56 add-ons
These attach to prompts rather than replacing them. Paste one at the end of a prompt you already use and it changes what comes back — how the model reasons, how long the answer runs, whose point of view it takes.
A prompt add-on is a short instruction you paste onto the end of an existing prompt to change the output. It is not a prompt on its own: it assigns no role and asks for no task. It only modifies the prompt it is attached to. Copy one, paste it after your prompt, and send both together.
These change how the model works before it answers. Use them when the reasoning matters more than the speed.
Before you give me the final answer, write out the reasoning that got you there. Then give the answer.
Copy add-onMakes the model expose its logic instead of jumping to a conclusion.
When the answer matters more than the speed, and you need to check the logic before you trust it.
List three different ways this request could be interpreted. Pick the one you think I meant, say why, then answer that one.
Copy add-onSurfaces ambiguity in your own prompt before the model commits to a reading.
When your prompt is short, or the first answer you got solved a problem you did not have.
Start from the outcome I want, then work backwards to the steps that produce it. Give me the steps in forward order.
Copy add-onReverses the planning direction so the output stays tied to the goal.
Planning a sequence — a close plan, a multi-touch cadence, a mutual action plan.
Before answering, list every assumption you are making. Mark any that you are not confident about.
Copy add-onSeparates what the model knows from what it filled in.
When you gave partial context and want to know what got invented to cover the gap.
After your answer, list the two or three pieces of information that would most change it if I gave them to you.
Copy add-onTurns one answer into a list of what to go find out.
Early in a deal, when you are deciding what to ask in the next call.
Make the strongest case for each option, then pick one and say what tipped it.
Copy add-onForces a real comparison instead of a preference stated as a conclusion.
Any either-or call: which persona to lead with, which of two deals to work.
When you are done, point to the single weakest step in your reasoning and say why it is weak.
Copy add-onMakes the model mark its own soft spot instead of presenting flat confidence.
Before you forward the output to someone who will pressure-test it.
These make the model ask before it answers. Use them when the task depends on context it cannot guess.
Do not answer yet. Ask me up to five questions that would let you give a genuinely good answer. Wait for my replies.
Copy add-onConverts a one-shot prompt into a short intake conversation.
When the task depends on context the model cannot guess — your product, your buyer, this specific deal.
Ask me questions one at a time, waiting for each answer before the next. Stop when you have enough to answer well.
Copy add-onPaces the intake so each question can build on the last answer.
On a phone or between meetings, when a wall of five questions will not get answered.
Answer with what you have. Then list what you did not know, and what you would have written differently if you had it.
Copy add-onGives you a usable draft now and the gap list at the same time.
When you need something immediately but want to know where it is thin.
Before you write anything, restate what you think I am asking for in three lines. Wait for me to confirm or correct it.
Copy add-onCatches a misread before the model spends a long output on it.
Long or expensive outputs — a full proposal section, a territory plan.
Only ask me a question if the answer would materially change your output. Otherwise proceed and note the assumption.
Copy add-onFilters the intake down to questions that actually change the result.
When clarifying prompts have been asking you things that did not matter.
Give me three short directions you could take this, one line each. I will pick one, then you write it out in full.
Copy add-onSplits the work into a cheap choice and then one committed draft.
Messaging and positioning work, where the angle matters more than the prose.
Interview me about this opportunity like a manager doing a deal review. Ask follow-ups where my answers are thin. Then summarize what you learned.
Copy add-onRuns a structured extraction of what you know instead of asking you to write it out.
Before a forecast call, or when you need to get a deal out of your head and onto paper.
These change the shape of the answer rather than its content. Most of the time saved by AI is saved here.
Give the answer as a table. Choose columns that make the options genuinely comparable, and keep every cell under ten words.
Copy add-onForces parallel structure so options can be compared at a glance.
Comparing accounts, competitors, packages, or candidates for outreach.
Answer in bullets. Start every bullet with the actual point in bold, then the explanation. No preamble.
Copy add-onPuts the conclusion first so the output survives skimming.
Anything a manager or an executive will read fast.
Give this as a numbered sequence of steps in the order I should actually do them. One action per step.
Copy add-onConverts advice into an ordered list you can work through.
Close plans, onboarding handoffs, any process you will hand to someone else.
Give me two columns: what I have now, and your revised version. Change nothing you cannot justify.
Copy add-onMakes every edit visible instead of returning a silently rewritten block.
Editing your own email, call script, or proposal language.
Output only the message itself, ready to paste. No subject line explanation, no notes, no commentary before or after.
Copy add-onStrips the wrapper so the output goes straight into the send window.
High-volume sending, where trimming commentary every time is the real cost.
Convert this into a checklist of things to verify, each phrased so it can be answered yes or no.
Copy add-onRewrites guidance as binary checks you can actually run.
Pre-call prep, proposal review, deal-desk submission.
Open with a three-line summary that stands on its own. Then the full detail below it under a clear heading.
Copy add-onGives one output that works for both a skimmer and a reader.
Anything forwarded up a level, where the first reader stops after three lines.
These set how much you get back. A length limit is the fastest way to make an answer usable.
Keep the entire answer under 120 words. If you cannot fit it, cut the least important point rather than compressing everything.
Copy add-onSets a hard budget and makes the model choose what to drop.
Follow-ups, internal notes, anything with a real length limit.
Answer in one sentence first. Then, below it, give the longer version for when I need it.
Copy add-onProduces a headline answer and the backup in a single pass.
When you need something quotable now and the reasoning available later.
Identify the single most important point in your answer and expand only that one. Keep the rest to one line each.
Copy add-onSpends the depth budget where it actually pays.
When even coverage is making everything equally shallow.
Give me only the twenty percent that produces most of the result. Say explicitly what you are leaving out.
Copy add-onTrims to the high-yield subset and names the tradeoff.
When you have limited time and want the moves that matter first.
Cover everything briefly, except the part I name — go deep there and treat the rest as context.
Copy add-onAims the detail at the section you actually care about.
When one section of a long output is the only part being reviewed.
Give the shortest useful answer. End by listing what you could expand on, and wait for me to pick.
Copy add-onMakes depth opt-in instead of default.
Exploratory work where you do not yet know which thread matters.
Remove every sentence that does not add information. No restating the question, no summarizing what you are about to say, no closing offer to help.
Copy add-onStrips the scaffolding language models add around the substance.
Every time. Especially anything a prospect will read.
These change the voice. Most default AI writing reads like a vendor, which is the problem these solve.
Here is a sample of how I write. Match its tone, sentence length, and vocabulary. Do not improve on it.
Copy add-onAnchors the voice to a real example instead of an adjective.
When 'professional but friendly' keeps producing something that does not sound like you.
Write this so a smart person outside my industry understands it. No jargon, no acronyms, no internal shorthand.
Copy add-onRemoves the vocabulary that only works inside your company.
Writing to a new stakeholder, a CFO, or anyone outside the technical buying group.
Write this as one professional to another. No selling language, no enthusiasm markers, no thanking them for their time.
Copy add-onDrops the deference that marks a message as a sales email.
Cold outreach and executive follow-up, where sounding like a vendor is the problem.
Be direct. Say the thing plainly, including the part I might not want to hear. No cushioning.
Copy add-onRemoves hedging so the actual assessment is visible.
Internal deal reviews and self-assessment, where softened feedback is useless.
Write this for a reader who is skeptical and short on time. Every claim needs to be either specific or cut.
Copy add-onRaises the evidence bar and removes unsupported claims.
Executive summaries, business cases, anything a CFO reads.
Mirror the sentence length and formality of the message I am replying to. If they wrote two casual lines, do not reply with six formal ones.
Copy add-onMirrors the counterpart's register in a reply instead of defaulting to formal.
Replying in a live thread, where a formality mismatch reads as tone-deaf.
Remove all superlatives and hype. No 'best-in-class', no 'cutting-edge', no 'revolutionary'. State what it does.
Copy add-onStrips marketing intensifiers that reduce credibility with senior buyers.
Any writing that goes to an evaluator who has read fifty vendor emails this month.
These change whose view the answer comes from. Running your own work past a buyer's eyes is the highest-yield use on this page.
Answer this from the buyer's point of view, not mine. What are they thinking, and what would make them stop reading?
Copy add-onFlips the vantage point to the person receiving your work.
Reviewing your own outreach or proposal before it goes out.
Respond as the CFO who has to approve this spend. What do you need to see, and what makes you say no?
Copy add-onApplies the economic buyer's criteria to your material.
Before a business case goes to finance, or when a deal stalls at approval.
Take the position of my main competitor. How would you attack this, and where is it genuinely weak?
Copy add-onRuns your material through an adversarial reading.
Competitive deals, and before any bake-off or evaluation.
Review this the way a demanding sales manager would in a deal review. Push on anything I have not evidenced.
Copy add-onApplies internal scrutiny before external exposure.
Before a forecast call or a deal review you would rather not be surprised in.
Respond as the procurement lead. What do you push back on, and where is my position weakest on commercial terms?
Copy add-onSurfaces the commercial pushback separately from the business case.
Once a deal reaches paper and the buying conversation changes hands.
Take the position of my champion, who has to sell this internally without me in the room. What is hard to explain?
Copy add-onTests whether your case survives being retold by someone else.
Any deal that gets decided in a meeting you are not in.
Respond as a skeptical technical evaluator. Where is this vague, and what would you need to test?
Copy add-onApplies a technical reading to claims written for a business audience.
Security reviews, technical validation, and proof-of-concept scoping.
These make the model check its own work. Cheaper than a second prompt and usually more honest.
Score this out of ten on clarity, specificity, and relevance to the reader. Show each score and the one change that would raise the lowest.
Copy add-onProduces a measured assessment instead of general feedback.
Comparing several drafts, or tracking whether your writing is improving.
First critique your own draft honestly. Then write a revised version that fixes what you found.
Copy add-onRuns an edit pass in the same output as the draft.
Any first draft, when you want the second draft without a second prompt.
Try to take this apart. Find the weakest claim, the assumption most likely to be wrong, and the fastest objection.
Copy add-onAttacks the output instead of polishing it.
Before anything high-stakes: a proposal, a pricing position, a QBR narrative.
Mark each substantive claim as high, medium, or low confidence. For anything low, say what would raise it.
Copy add-onSeparates the reliable parts of the output from the guesses.
Research output, account plans, anything you will repeat to a customer.
After your answer, list what is missing that a thorough version would include.
Copy add-onNames the gaps rather than quietly leaving them.
Account plans, close plans, and anything built from partial information.
Re-read my original request. List anything I asked for that your answer does not actually cover.
Copy add-onAudits the output against the brief instead of against itself.
Long multi-part prompts, where one requirement quietly gets dropped.
Give me your answer, then a weaker version you considered and why you rejected it.
Copy add-onExposes the choice behind the output so you can disagree with it.
When the model keeps making a stylistic call you would not have made.
These set limits. The most important one blocks the model from inventing facts you will later have to defend.
Use only the information I provided. Do not add facts, statistics, names, or examples from anywhere else. If something is missing, say so.
Copy add-onBlocks invented detail from entering the output.
Anything customer-facing that cites specifics — numbers, names, results.
Write this without referring to price, discounts, or commercial terms in any form.
Copy add-onKeeps the conversation on value while pricing is still being sequenced.
Early-stage messaging, and any thread where pricing goes through a different channel.
Keep this under 300 characters total. Count them and confirm the number at the end.
Copy add-onEnforces a hard limit for channels that impose one.
LinkedIn connection notes, SMS, and subject-line work.
Use the exact terms the customer used in the notes I pasted. Do not substitute my company's vocabulary for theirs.
Copy add-onKeeps the customer's own language instead of translating it into your category terms.
Recaps, proposals, and business cases built from call notes.
Do not use the following words anywhere in the output, including variations of them: [list].
Copy add-onEnforces a banned-term list on the output.
Brand guidelines, or when one phrase keeps reappearing in every draft.
The message may contain exactly one request. Everything else is context. Make the single ask unmistakable.
Copy add-onRemoves competing calls to action that split the reply.
Any outbound message where reply rate is the metric.
Do not ask me anything. Make reasonable choices, note them briefly at the end, and give me the finished output.
Copy add-onSuppresses clarifying questions when you need a result in one pass.
Batch work, or any moment you cannot sit in a back-and-forth.
Add-ons combine. Two or three on one prompt is usually the point where the output stops needing an edit. Three worked examples:
Write a cold email to a VP of Sales at a mid-market logistics company that just posted three SDR roles. Write this as one professional to another. No selling language, no enthusiasm markers, no thanking them for their time. Keep the entire answer under 120 words. If you cannot fit it, cut the least important point rather than compressing everything. The message may contain exactly one request. Everything else is context. Make the single ask unmistakable.
Peer to peer removes the deference, the word cap forces a choice about what matters, and the single ask stops the reader having to decide which question to answer.
Here are my notes from the last three calls on this opportunity. Summarize where the deal actually stands. Before answering, list every assumption you are making. Mark any that you are not confident about. Review this the way a demanding sales manager would in a deal review. Push on anything I have not evidenced. After your answer, list what is missing that a thorough version would include.
The assumptions list separates what you were told from what you inferred, the manager lens finds the unevidenced claims, and the gap list becomes the agenda for your next call.
Draft the executive summary for this proposal using the discovery notes below. Write this for a reader who is skeptical and short on time. Every claim needs to be either specific or cut. Use only the information I provided. Do not add facts, statistics, names, or examples from anywhere else. If something is missing, say so. Open with a three-line summary that stands on its own. Then the full detail below it under a clear heading.
This is the combination that matters most on a document someone forwards. The skeptic lens forces specificity, the no-invention rule keeps you from citing a number you cannot defend, and the standalone summary is the only part a CFO reads.
A short instruction you paste onto the end of an existing prompt to change what comes back. It assigns no role and asks for no task of its own, so it is not a prompt — it only modifies the prompt it is attached to.
Copy the add-on, paste it on a new line after your prompt, and send both together as one message. Nothing else changes about how you work.
Yes, and stacking two or three is where most of the value is. Avoid combining two that pull in opposite directions, such as a length cap and a request to expand — the model will pick one and you will not know which.
They are written to be tool-agnostic and carry no product-specific syntax, so they work with Claude, ChatGPT, Gemini, Grok, Microsoft Copilot, and Perplexity. Models differ in how strictly they follow a length cap.
A prompt is a complete instruction that stands on its own and produces an output. An add-on produces nothing by itself. Sent alone it will confuse the model, because it is an adjustment to a task rather than a task.
Yes. Every add-on on this page is free to copy with no account. The prompt library itself is where Pro applies. Current plans and prices are on the pricing page.
The broad areas of sales work these attach to.
The specific task level, if you know what you are doing.
The 10 stages of a deal, if you know where yours is.