ChatGPT Prompts for Sales: 34 for Outreach, Discovery & Closing (2026)
34 copy-ready ChatGPT prompts for sales: prospect research, cold email, discovery, objection handling, follow-up, closing, and where AI shouldn't lead.
Researched with AI assistance, reviewed and edited by Tapabrata Biswas.

In this article
- 01How to prompt as a salesperson
- 02Prospect and account research
- 03Cold email and outreach
- 04Cold calls and call prep
- 05Discovery and qualification
- 06Objection handling
- 07Follow-up and nurture
- 08Closing and negotiation
- 09Pipeline, CRM and admin
- 10Where not to let ChatGPT lead
- 11What this post does not cover
- 12Sources
Most sales prompt lists hand you a template about a fictional prospect and stop there. The trouble is that a generic prompt gives you generic outreach, and generic outreach is the fastest way to get ignored or marked as spam. ChatGPT earns its place in a rep's day not by closing for you, but by taking the research, the drafting, and the prep off your plate so you spend more time actually selling. The prompts below are built for that, each with a clear job, a place to drop your real prospect and offer context, and an output you can send after a quick personal pass.
They run across the whole sales cycle, from research and cold outreach through discovery, objections, and closing, and they work the same in ChatGPT, Claude, or Gemini. There is also an honest section on where AI has no business leading, because in sales that line, especially fake personalization and prospect data, is where reps get themselves into real trouble.
How to prompt as a salesperson
A prompt that saves you time and one that wastes it differ by four things: role, context, task, and format. Tell the model who to be, give it the real details it can't see, state exactly what you want, and specify how the answer should be shaped. Leave the context out and it fills the gap with the same tired sales-speak everyone ignores, the problem behind writing a clearer prompt in any field.
Here is the difference. The vague version:
Write me a cold email for my product.
And the version that actually helps, with role, context, task, and format:
Act as an SDR selling [product], which helps [type of company] to [outcome]. I'm emailing a [job title] at a [company type]. Their likely pain is [pain]. Write a three-sentence cold email: one line that shows I understand their situation, one line on the specific outcome we drive with a proof point, and a soft ask for a 15-minute call. Human tone, no jargon, no 'I hope this email finds you well'.
The first gives you spam. The second gives you a draft you can personalise and send. Every prompt below is written in that second style, and the habit that matters most is replacing the brackets with your real, non-sensitive context before you hit enter.
Prospect and account research
Research is where a rep either earns the right to a reply or wastes everyone's time, and it's exactly the structured, repetitive work ChatGPT speeds up. Use it to organise what you know and shape your angle, but verify anything it states as fact about a real company, because it will confidently fill gaps with invention.
Act as a sales researcher. Based on this public information about a company, [paste website copy, news, or notes], summarise what they do, their likely priorities this year, and three plausible pain points my product, [product], could solve. Flag which of these are informed guesses versus stated facts so I don't repeat a guess as truth.
Help me build an ideal customer profile for [product]. Based on our best current customers, [describe them], define the ICP: firmographics, the trigger events that signal a need, the roles involved in the decision, and the pains that make them buy now rather than later. Keep it specific enough to disqualify a bad-fit lead.
Given this target account, [describe the company generically], map the likely buying committee for a [product] purchase: the probable economic buyer, champion, users, and blocker, what each cares about, and the one message that would land with each role.
I'm preparing to reach out to a [job title] in [industry]. What are the three or four business pressures someone in that role is likely feeling right now, in plain terms, and how would I connect each to the outcome [product] delivers without sounding generic?
Turn this messy prospect research into a one-paragraph pre-outreach brief: [paste notes]. Give me the single most relevant hook to open with, the outcome to lead on, and the one thing I should not say. Keep it to something I can scan in ten seconds before I write the email.
Cold email and outreach
This is the highest-volume writing a rep does and the easiest to get wrong at scale. ChatGPT drafts fast, but the reply comes from a true, specific detail, so the rule is simple: let it write the structure, and you supply the real personalization. Never let it invent a reference to a post, a promotion, or a mutual connection that did not happen, because one fake detail ends the relationship before it starts.
Write a three-sentence cold email to a [job title] at a [company type] about [pain]. Sentence one shows I understand their world, sentence two names the specific outcome we drive with a number if possible, sentence three is a soft ask for a short call. No fluff opener, under 90 words, and leave a [personalise here] placeholder for the real detail I'll add.
Write a four-touch cold email sequence for [product] to a [job title]. Email 1 opens on their likely pain, email 2 adds a proof point or customer story, email 3 offers a useful resource with no ask, email 4 is a short break-up email. Vary the angle each time, keep each under 100 words, and keep them clearly connected.
Draft a LinkedIn connection message and a follow-up message for a [job title] I want to sell [product] to. The connection note is warm and reason-led, not a pitch; the follow-up, sent a few days later, offers a specific insight or resource relevant to their role. Both short and human, no 'quick question' openers.
Give me five cold-email subject lines for an email about [pain/outcome] to a [job title]. Mix curiosity, specificity, and a question, keep them under six words, and avoid anything that reads as clickbait or a mass send. Tell me which one you'd bet on and why.
Rewrite this cold email so it sounds like a human wrote it, not a template: [paste]. Cut the corporate filler and the fake enthusiasm, make the ask clear and low-pressure, and keep only the sentences that a busy buyer would actually read.
Cold calls and call prep
Write a cold-call opener for reaching a [job title] about [pain]. Fifteen seconds, permission-based, that earns the next thirty seconds instead of triggering a brush-off. Then give me two ways to respond if they say 'I'm busy' and one if they say 'what's this about?'.
Prep me for a call with a [job title] at a [company type] considering [product]. Give me a one-page brief: the likely goals of that role, three smart questions to ask, the two objections I should expect, and the single outcome to anchor the conversation on.
Write a short, human voicemail script for a prospect I couldn't reach, about [pain/outcome]. Under 20 seconds spoken, gives a reason to call back, and doesn't sound like every other rep. Then write the one-line follow-up text or email to pair with it.
Help me handle a gatekeeper when trying to reach [role]. Give me a respectful, honest approach that treats them as an ally, not an obstacle, with two or three lines I could actually say, and a fallback if they won't put me through.
Discovery and qualification
Weak discovery is where good-looking deals quietly die, so this is worth prepping properly. ChatGPT can build the question set and the qualification structure; the actual listening, following the thread, and reading what's unsaid stay entirely with you on the call.
Design a discovery call question set for selling [product] to a [job title]. Give me three questions about their current situation, three about the pain and its cost, two about what they've already tried, and one about their decision process and timeline. Open, non-leading, and ordered so the conversation flows naturally.
Turn my discovery notes into a qualification summary using [BANT or MEDDIC]: [paste notes]. Fill in what I know, flag what's missing, and tell me the two most important gaps to close before this can be a real opportunity. Be honest if this looks like a weak deal.
Draft a short agenda to send before a discovery call with [role]. Frame it around understanding their goals rather than pitching, set expectations for the time, and make it easy for them to add a topic. Under 80 words, so it reads as considerate, not corporate.
Help me dig into a pain the prospect mentioned but skated over: [describe]. Give me three follow-up questions that get at the real cost and impact of that problem, and one that surfaces who else it affects, without sounding like an interrogation.
Objection handling
Objections are where a deal gets real, and the reps who handle them well have thought about them in advance, not on the spot. ChatGPT is good at building that library, a sharp response to each objection you hear most, so you walk into the call prepared rather than improvising under pressure with a discount you didn't need to give.
A prospect says my price is too high. My product is [product], and its real value is [outcome, ROI, or savings]. Give me three responses that reframe the conversation around value and cost of inaction instead of discounting, each a couple of sentences I could actually say out loud.
A prospect says they're happy with their current vendor, [competitor or status quo]. Help me respond in a way that respects that, plants one specific reason to look again, and asks a question rather than pushing. Don't trash the competitor.
A prospect says 'just send me some information'. Draft a reply that sends something genuinely useful but keeps the conversation alive with a light next step, so it doesn't become a polite dead end. Short and warm.
A prospect says 'now isn't a good time' or 'call me next quarter'. Help me tell a real not-now from a soft no: give me one question to test which it is, and a graceful way to stay in touch that isn't just chasing.
Build me an objection-handling cheat sheet for [product]. List the five objections I'm most likely to hear from [buyer type], and for each give the underlying concern behind it and a two-sentence response that addresses the concern, not just the words.
Follow-up and nurture
Most deals are won or lost in the follow-up nobody sends, and it's exactly the kind of consistent, structured writing ChatGPT handles well. Use it to draft the sequence and the check-ins, then add the specific detail from your last conversation so it reads as a person who was paying attention.
Write a follow-up email after a discovery call with [role]. Recap the two or three things they said mattered most, confirm the next step we agreed, and add one relevant resource. Reference the specifics so it's clearly about our conversation, not a template, and keep it short.
A deal has gone quiet after [what happened]. Write a short re-engagement email that doesn't guilt-trip or say 'just checking in', leads with something of value or a relevant update, and makes it easy for them to respond with a yes or no.
Design a multi-touch follow-up plan across two weeks for a warm prospect who hasn't replied: [context]. Vary the channel and the angle each touch, front-load the value, and tell me when to stop so I'm persistent, not annoying.
Write a genuine break-up email for a prospect who's gone completely dark. Give them an easy out, leave the door open without desperation, and, done well, occasionally prompt a reply. One short paragraph.
Closing and negotiation
Draft a mutual action plan to send a prospect moving toward a decision on [product]. Lay out the steps, owners, and dates from here to a go-live, framed as a shared plan to hit their timeline, not a pressure tactic. Keep it simple enough to fit on one screen.
Help me ask for the close with a prospect who seems ready but hasn't committed: [context]. Give me two or three direct but low-pressure ways to ask for the business and surface any remaining hesitation, in language I could actually say.
A prospect is pushing for a discount. My position is [describe]. Give me negotiation talking points that hold value, offer a trade rather than a straight cut where possible, and keep the relationship intact whichever way it goes. Tell me my walk-away line.
Summarise our proposal into a one-page recap I can send to a prospect's internal champion to sell it for me: [paste the offer and value]. Lead with their outcome, keep the pricing clear, and give the champion the three bullet points they'll need to defend the decision internally.

Pipeline, CRM and admin
Turn my messy call notes into a clean CRM entry: [paste notes]. Give me a two-line summary, the next step with a date, the current stage, and any risk to the deal. Keep it factual and skimmable so future-me and my manager can read it in seconds.
Help me prep for a deal review with my manager. Here's the opportunity: [describe generically]. Structure it as the situation, why it will close, the biggest risk, and what I need from the team, so I walk in with a clear story rather than a status dump.
I need to clean up my pipeline. Here's a list of my open deals with rough status: [paste generically]. Help me sort them into commit, best-case, and unlikely, ask me the one question that would move each into the right bucket, and flag the deals I'm probably fooling myself about.
Where not to let ChatGPT lead
Here is the honest part, and in sales it's where reps get burned. ChatGPT is a fast researcher and drafter, and it is not a salesperson, so the selling stays yours.
Three lines you don't cross. Never let it fake personalization: it will happily write "I loved your recent post on X" about a post that doesn't exist, and a single invented detail destroys the trust you were trying to build, so verify every personal claim before it sends. Never paste real prospect data, names, email addresses, company financials, or CRM deal values, into a consumer chatbot, because that information can be retained and it's a genuine privacy and compliance risk; describe the situation generically instead. And never send anything unread or spray generic AI outreach at volume, which tanks your reply rate, your sender reputation, and your credibility all at once. Used inside those lines, it clears the busywork so you spend your hours in front of buyers. Outside them, it makes you look like a bot. The cold-email side pairs well with our prompts for email writing when you want to sharpen the writing itself.
What this post does not cover
This is a working set of prompts and how to use them safely, not a sales methodology or a guide to any specific CRM or sales-engagement tool. The prompts are starting points to adapt, and they assume you'll add your own real, non-sensitive context and review every output before it reaches a buyer. Nothing here replaces the discovery, the trust, and the timing that actually close deals, and no AI draft should go out without your eyes on it first. If prompting itself is new to you, the basics of prompting is the place to start.
Sources
- Prompt engineering guide, OpenAI (the role, context, task, and format principles behind the prompts here)
- MEDDIC sales qualification, Wikipedia (the qualification framework behind the discovery prompts)
Frequently asked questions

Written by
Tapabrata Biswas
Tech Researcher
I test AI productivity tools and research home-automation gear the way most people use them. Not in a lab, but on an ordinary desk with an ordinary internet connection. The only test that matters: does it save you time?
Share the Post with Your Besties
Get the plain-English tech brief
One email a week on AI tools and smart-home tech. No jargon, no hype.


