Découvrez folk, le CRM pour les entreprises axées sur les ressources humaines
You have five minutes before a discovery call. The buyer explained what they need across six emails, and someone on your team added a note that says “timing unclear.” Paste the thread into ChatGPT and ask it to pull out what the buyer confirmed, what’s still open, and the three questions worth asking on the call.
That’s the idea behind these 24 ChatGPT prompts for sales. Each starts with something you already have, like a buyer’s reply, a call transcript, or a list of target accounts. Copy the prompt that fits the job, replace the brackets with your own material, and check the output against the source before you use it.
Quick Overview
- Sales prospecting (5 prompts): Rank accounts against your ICP, pick the right person to contact, and turn a verified buying signal into an outreach angle.
- Cold email and LinkedIn outreach (4 prompts): Write a first touch, build a three-message sequence, or reply to “We already use another tool.”
- Discovery calls and demos (4 prompts): Pull questions from a buyer’s email thread, identify gaps in qualification, and turn discovery notes into a demo agenda.
- Follow-ups and pipeline reviews (5 prompts): Recap a demo, reopen a quiet conversation, or flag deals without a confirmed next step.
- Sales enablement (6 prompts): Practice a tough objection, coach a rep using a call transcript, or build a competitive battlecard from approved material.
Best ChatGPT Prompts for Sales Prospecting
A company announces a new office. Is that a reason to reach out? Maybe. A recruiter and an IT vendor would read that announcement differently. These five prompts help you decide which accounts to prioritize, who to contact, and whether you have a good reason to start a conversation.
1. Rank target accounts against your ICP
✅ Use this when you have a list of companies and need to decide which ones deserve more research before you start outreach.
Here’s what we sell and who typically buys it: [product and buyer].
Here are our must-have account criteria: [criteria].
Here are our deal-breakers: [exclusions].
Here’s my account list with research notes: [paste list].
Put each account into one of three groups:
- Priority: The notes support every must-have and show no deal-breaker.
- Research first: A must-have is unverified, and there’s no known
deal-breaker.
- Skip: The notes confirm a deal-breaker or show that the account
fails a must-have.
Return a table with the account, group, evidence, and next step.
For each “Research first” account, tell me the one thing to verify
before I spend more time on it. Use only the information I supplied.
2. Turn a buying signal into an outreach angle
✅ Paste the original announcement, job listing, or interview excerpt. A headline alone can leave out the detail that makes the signal relevant to your buyer.
I sell [product] to [buyer role]. It helps them [specific outcome].
Here’s the source: [paste the excerpt, link, and publication date].
Here’s what we know from past conversations: [CRM notes or “none”].
Give me four things:
1. What the source confirms, in one sentence.
2. One way it could affect this buyer’s work. Label it as a hypothesis.
3. One question I could ask to see whether that hypothesis is right.
4. One cold email opening line that refers to the verified fact
without assuming the buyer has a problem.
If the source doesn’t give me a credible reason to contact this
person, say “no usable angle” and explain why in one sentence.
Use only the material I provided.
3. Identify who to contact first at an account
✅ Start with the people you’ve already found. This is especially useful when several contacts have similar titles and it’s unclear who owns the workflow you sell into.
Here’s the workflow our product improves: [describe the workflow].
Here’s who usually owns it at our customers: [role and responsibilities].
Here are the people I found at this company: [names, titles, and
verified responsibilities].
Here’s any prior contact we’ve had with them: [notes or “none”].
Recommend one person to contact first and one backup, if the
evidence supports a second choice. For each, point to one specific
responsibility or prior interaction that explains your choice.
If ownership is unclear, tell me what to verify before reaching out.
Don’t assume someone controls the budget because of their title,
and don’t suggest contacts outside the list.
4. Qualify an inbound lead before the first call
✅ Use this to triage demo requests when the form tells you who signed up but gives you little insight into what they need.
Here’s the inbound form submission: [paste submission].
Here’s what we already know about this lead: [CRM notes or “none”].
Here are our qualification criteria and routing rules: [paste rules].
Recommend one route from our rules and explain what supports it.
List any criteria you can’t assess from the information provided.
Then write up to two questions I can send the lead to resolve the
gaps that would actually change the routing decision.
Skip anything the lead has already answered. If the rules don’t
cover this case, flag it for manual review and say what’s unclear.
5. Build a one-page account brief
✅ Use this before handing an account to another rep or contacting a company your team has spoken with before. Include the last buyer reply if you have it. It often tells the next rep more than a company overview does.
Here are the source excerpts I gathered about the account:
[paste excerpts and dates].
Here’s our CRM history with the account: [paste relevant notes,
the latest buyer reply, and any open commitments].
Here’s why we’re considering outreach: [reason].
Write an account brief under 250 words with four sections:
- Account context: Up to three verified facts relevant to our offer.
- Relationship history: Who we spoke with, when, and what came of it.
- Open commitments: Anything either side agreed to do that’s still
pending. Include an owner and date only if the notes confirm them.
- First move: One action the next rep can take and one thing to
verify first, if needed.
If the sources and CRM notes conflict, flag the discrepancy.
Leave out company background that won’t affect the next step.
If your team works through a shared list, keep the source and the next action alongside each account. That gives the next rep enough context to pick up the work. See how folk supports AI sales prospecting.
Best ChatGPT Prompts for Cold Emails and LinkedIn Outreach
A cold email needs a reason for reaching out. A LinkedIn DM after an event already has one. Give ChatGPT the detail that explains why you’re writing, then use the prompt that fits the conversation.
6. Write a cold email around one verified fact
✅ Pick a fact that has something to do with the buyer’s job. A job posting that mentions the workflow you sell into gives you more to work with than a funding announcement on its own.
Write a first-touch cold email to [buyer role] at [company].
Here’s the fact I verified: [fact and source].
Here’s how our product helps someone in this role: [specific use case].
Here’s one claim or customer result we’re approved to share: [proof
point and any context needed to use it accurately, or “none”].
Here’s a short email I’ve written before: [sample for tone].
Give me two subject lines and one email under 90 words. Open with
the verified fact and connect it to the use case in plain English.
If the connection is weak, tell me what detail would make it stronger
before drafting.
Use the proof point only if it supports the use case. End with one
easy-to-answer question. Don’t claim the buyer has a problem that
the source doesn’t establish.
7. Write a LinkedIn DM after a real interaction
✅ Use this after someone replies to your comment, asks a question at an event, or gets introduced to you. Paste the part of the exchange you want to pick up.
Draft a LinkedIn DM to [name].
Here’s where we interacted: [thread, event, or introduction].
Here’s what they actually said: [quote or accurate notes].
Here’s why I want to continue the conversation: [reason].
Write two or three sentences under 60 words total. Refer to one
specific detail from the exchange and ask one question that follows
from it. Match the level of familiarity in the original interaction.
If I said I’d send something, lead with that. Don’t add a meeting
request unless they already expressed interest in talking.
8. Build a three-touch email and LinkedIn sequence
✅ Start with the first email and two pieces of material you can actually share. A useful follow-up could include a relevant customer example or a short answer to a question buyers often ask at this stage. You can then set up the email touches in folk.
Build a three-touch sequence for [buyer role] at [company].
Here’s why I’m reaching out: [verified context].
Here’s my first email draft: [paste draft].
Here are two things I can share in later touches: [resource, proof
point, or practical tip, with the exact details I can use].
LinkedIn DM is available: [yes or no].
Write the initial email, one follow-up email in the same thread,
and one LinkedIn DM. If LinkedIn DM isn’t available, write a brief
final email instead.
Give each touch a different reason to read it. Use one new piece
of material in each follow-up and explain its relevance in one
sentence. Keep emails under 90 words and the DM under 50.
Don’t refer to an email open or claim we’ve spoken before unless
I provided that information. End the sequence if the buyer replies
or asks me to stop.
9. Reply to “We already use another tool”
✅ bThe buyer might be happy with their setup, or they might be telling you that switching is off the table. Their exact wording matters more than the name of the competitor.
Help me reply to this prospect: [paste their exact reply].
Here’s my original message: [paste it].
Here’s what they’ve told us about their current setup: [notes or “none”].
Here’s a verified difference between our product and their setup:
[approved comparison or “none”].
Write a reply under 70 words. Acknowledge what they said in one
sentence. If they left room for a conversation, ask one question
about how they handle the workflow I originally brought up. Mention
a product difference only if it addresses something they raised.
If they clearly declined, give me a short acknowledgment that
closes the thread. Don’t add another question or a meeting request.
Best ChatGPT Prompts for Discovery Calls and Demos
If a buyer already told you why they booked the call, use the first few minutes to dig into that problem. The email thread, your call notes, and the buyer’s own wording give ChatGPT enough context to help you prepare questions and decide what to show in the demo.
10. Prepare for a discovery call from an email thread
✅ Start with the buyer’s latest reply and the messages it refers to. If several people on your team have spoken to the account, pull a recap of recent interactions to catch details that sit outside the thread.
Help me prepare for a [call length]-minute discovery call.
Here’s the email thread: [paste relevant messages].
Here are our notes from earlier conversations: [paste notes or “none”].
Here’s what I want to learn on the call: [goal].
Give me a brief I can scan in one minute:
1. Why the buyer agreed to meet, using their own words where possible.
2. What they’ve already told us about their current process and priorities.
3. Any assumptions in our notes that the buyer hasn’t confirmed.
4. Three questions to ask, in the order I should ask them.
For each question, tell me what the answer would help me decide.
If the buyer already answered it, use that answer to write a more
specific follow-up question.
11. Write questions that uncover the cost of the current workflow
✅ When a buyer says a process “takes forever,” start by finding out what happens each time they run it. You can work out which numbers matter once you understand where the time goes.
Here’s how the buyer described their current workflow: [paste their
description]. Here’s the part of the workflow we can improve:
[describe the relevant step]. Here’s what we already know about
the time, volume, and people involved: [notes or “unknown”].
Write four discovery questions that would help me understand the
impact of this workflow. Start by asking the buyer to walk me through
the last time they ran it. Use the remaining questions to explore
frequency, time spent on the most manual step, and what happens
when that step goes wrong or runs late.
Skip any question the buyer has already answered. After the questions,
list the numbers I’d need before estimating the cost. Use only
the figures the buyer has provided.
12. Turn discovery notes into a buyer-specific demo agenda
✅ If the buyer spent half the call describing a messy handoff, make that handoff the first workflow you show. Paste the features you can actually demo so the agenda stays grounded in what your product does today.
Build a [number]-minute demo agenda for [buyer roles attending].
Here are the discovery notes: [paste notes].
Here are the workflows and features I can show live: [list].
Here’s what the buyer wants to see, if they said: [request or “none”].
Choose up to three workflows based on the buyer’s stated priorities.
For each, give me the buyer’s reason for caring, what to show on
screen, one question to ask afterward, and a time limit in minutes.
Leave time for buyer questions and next steps. Make sure the agenda
fits the total meeting time. If a requested workflow isn’t covered
by the features I listed, flag it so I can address it directly.
13. Find qualification gaps after a sales call
✅ Use this before you update the deal stage or brief your manager. Paste your team’s definitions for the framework you use, whether that’s BANT, MEDDICC, or your own qualification criteria.
Review this call against our qualification criteria.
Here are the criteria and what counts as confirmed for each:
[paste your team’s definitions].
Here’s the call transcript or notes: [paste source].
Return a table with one row per criterion and four columns:
criterion, status, evidence, and question for the next conversation.
Use “confirmed,” “partially confirmed,” or “unknown” for the status.
For every confirmed or partially confirmed item, quote the buyer or
cite the relevant note. Label secondhand information as such. If
two statements conflict, flag both and leave the item open.
Finish with the two gaps that matter most for deciding whether the
deal should move forward. Explain what decision each answer would
help us make.
Best ChatGPT Prompts for Sales Follow-Ups and Pipeline Reviews
A buyer saying “send me the pricing” gives you a next step. “This looks interesting” gives you something to follow up on. These prompts help you draft the right reply and keep your pipeline grounded in what buyers actually said and did.
14. Write a post-demo follow-up with clear next steps
✅ Open with the part of the demo the buyer cared about. If they asked you to confirm a detail or send a document, put that ahead of the meeting recap.
Draft a follow-up email using these demo notes: [paste notes].
Here’s a short email I wrote that reflects my tone: [paste sample].
Here are any links or materials I’m ready to share: [list or “none”].
Open with the buyer’s main priority in their own terms. Then cover
up to three agreed next steps. Include an owner or date only when
the notes confirm it. If we agreed to meet again but left timing
open, ask one direct question about scheduling.
If I promised to send something and it isn’t ready, flag it below
the draft. Mention an attachment or link only if I provided it.
Keep the email under 120 words. Use my sample for tone only.
15. Follow up after a missed meeting
✅ Paste the original invite and the exchange that led to it. A buyer who asked to see a specific workflow is easier to re-engage when you remind them what they wanted to cover.
Write a follow-up email after a missed meeting with [name].
Here’s why we booked the call: [paste the invite or buyer’s message].
Here’s any note the buyer sent about missing it: [paste or “none”].
Here are two times I can actually offer, including the time zone:
[slots or “none”]. Here’s my booking link: [link or “none”].
In under 75 words, remind them what we planned to cover and make
rescheduling easy. Use the two time slots if I gave them. Otherwise,
use the booking link or ask what time works for them.
If they already proposed a time, respond to that first. Keep the
tone neutral and use only what I shared about why the meeting slipped.
16. Reopen a stalled sales conversation
✅ A prospect who asked you to check back after budget planning needs a different message from one who stopped replying after a demo. Include the last buyer message and any date they gave you.
Help me decide whether to follow up with [name] today: [today’s date].
Here’s the last buyer message and when they sent it: [paste message
and date]. Here’s what they said about timing, priorities, or why
they paused: [paste notes or “unknown”]. Here’s anything new I can
share that’s relevant to them: [detail or “none”].
If they asked me to wait until a specific date or milestone, check
whether it has passed based on the information provided. If it’s
too early, tell me when or under what condition to follow up. If
they explicitly declined or opted out, skip the draft.
Otherwise, write a reply under 85 words. Pick up the last thread,
mention any genuinely relevant update I supplied, and ask one
question that helps me tell whether the priority has changed.
If several deals have gone quiet, start with conversations where the buyer asked you to circle back or where your team still owes them something. An AI follow-up assistant can also surface inactive conversations with a pending next step.
17. Turn call notes into a CRM update
✅ Give ChatGPT the actual field names and allowed values. A note that says “legal needs to review” may belong in the deal summary, while a stage change depends on your team’s criteria.
Suggest CRM updates from these call notes: [paste notes or transcript].
Here are the current field values: [paste values].
Here are the fields I can update, their allowed values, and our
rules for changing deal stages: [paste fields and rules].
Return a table with four columns: field, current value, proposed
value, and supporting quote or note. Include only changes supported
by the source and allowed by our rules.
Below the table, list anything I should verify before making an
update. Flag conflicting information. Keep tentative dates and
unconfirmed decision-makers in the notes unless our rules say
otherwise.
If the call supports no changes, say so.
18. Flag deals at risk before a pipeline review
✅ Include the last buyer reply as well as the last activity date. An outbound email can make a deal look active even when the buyer hasn’t responded in weeks.
Review this pipeline as of [today’s date].
Here are the deal records: [paste deal name, stage, expected close
date, last buyer reply and date, confirmed next step, and open
blockers]. Here are our stage criteria and inactivity thresholds:
[paste rules or “none”].
Flag deals with a specific risk supported by the records. Look for
a missed buyer commitment, an unresolved blocker, a close date
that conflicts with the buyer’s timeline, or inactivity beyond our
threshold for that stage.
Return a table with the deal, risk, evidence, and one action for the
owner. Put time-sensitive issues first. If a blank field leaves you
unsure whether there’s a risk, list it separately under “Verify in
the review.” If I gave you no inactivity threshold, show the days
since the last buyer reply and leave the judgment to me.
Best ChatGPT Prompts for Sales Enablement
Pick one thing you want reps to get better at this week. If they tend to jump into a product pitch when a buyer pushes back on price, practice that exact exchange. Ask ChatGPT to play the buyer for a few turns, then review the answer the rep actually gave.
19. Role-play a common sales objection
✅ Use an objection a rep heard this week. Include the buyer’s current setup and where the objection came up. “We don’t have budget” means different things on a cold call and after a pricing review.
Play a [buyer role] evaluating [product category].
Your current setup is [workflow], and we’re at [deal stage].
Start with this objection: “[exact buyer objection].”
Reply as the buyer, one turn at a time. If I make a broad claim,
ask what it would change in your day-to-day work. If I ask a
specific question about your objection, answer it as the buyer.
Stay in character until I type “debrief” or we finish six exchanges.
Then score me from 1 to 5 on how well I clarified the objection,
supported my answer, and handled the next step. Mark anything
we didn’t reach as “not observed.”
Quote one line I should keep and rewrite one response I could
improve. Use my actual replies as evidence for the feedback.
20. Coach a rep using a real call transcript
✅ Choose one skill before pasting the transcript. If the rep is working on discovery, focus the review on how they followed up when the buyer mentioned a problem.
Review this call for one coaching skill: [specific skill].
Here’s what good performance looks like: [rubric or example].
Here’s the call transcript: [paste relevant excerpt].
Find one moment the rep handled well and one moment they could
handle differently. For each, quote the exchange and include
a timestamp if the transcript has one. Explain how the rep’s
response matched or missed the coaching criteria.
For the moment to improve, write one alternative line the rep
could have used with the information they had at the time.
Finish with a five-minute drill they can run before their next call.
If the transcript doesn’t show the skill, say what you’d need
to see to assess it.
21. Build a competitive battlecard from approved sources
✅ A rep should be able to scan the battlecard during a call. Give ChatGPT material your team has approved, including any plan limits or conditions that affect the comparison.
Build a one-page battlecard for [our product] vs. [competitor].
The buyer is [role], and the workflow they care about is [workflow].
Use these approved materials: [paste product docs, comparison
notes, and source dates]. Here are objections we’ve actually
heard in these deals: [paste anonymized examples].
Include:
1. Two discovery questions that reveal which differences matter.
2. Up to three verified differences, when each matters, and its source.
3. A short response to one buyer objection, in language a rep can say aloud.
4. Any claims that need verification before a rep uses them.
Include plan names or conditions where they affect a claim.
If the sources don’t support a comparison, mark it “unverified.”
Keep the main battlecard under 300 words.
22. Turn customer objections into a team playbook
✅ Buyers can use the same words for different concerns. “It’s too expensive” might mean the budget is fixed, or it might mean the buyer still can’t justify the spend. Use the surrounding conversation to keep those objections distinct.
Turn these anonymized buyer quotes into a short objection playbook:
[paste quotes with what came before and after each one].
Here are our approved responses and supporting materials:
[paste responses, proof points, and any limits on what reps can say].
Group quotes when the surrounding conversation points to the same
concern. If the reason behind a quote is unclear, label it “unclear.”
Create up to five entries with:
- One representative buyer quote.
- The concern the buyer stated, if known.
- One question a rep can ask before answering.
- A suggested response using our approved material.
- One next step if the concern is still open.
If we don’t have an approved response, flag the gap for the manager.
23. Create a practical onboarding exercise for new reps
✅ Give new hires a task they’ll do on the job. A 15-minute first-call prep exercise can show whether they know which parts of the product matter to a particular buyer.
Create a 15-minute discovery prep exercise for a new sales rep.
Here’s our ideal customer profile: [paste ICP].
Here’s our product one-pager: [paste text].
Here’s an approved case study or anonymized buyer scenario:
[paste source].
Write a rep handout with the buyer context and three tasks:
identify the likely priority, write three discovery questions,
and choose one relevant workflow to discuss if the buyer confirms
the priority.
Then write a separate manager guide with a scoring rubric, one
sample answer supported by the source, and two mistakes to watch for.
Accept different answers when the rep can support them with evidence.
24. Turn a win/loss debrief into a coaching exercise
✅ Pick a moment the team can replay, such as when a buyer flagged an approval issue or asked for proof a workflow would work. Use the buyer’s feedback to ground the discussion.
Turn this deal debrief into a 10-minute team coaching exercise.
Here are the deal notes in order: [paste notes].
Here’s what the buyer explicitly said about their decision:
[paste feedback or “none”].
Here’s what our team thinks influenced the outcome:
[paste interpretations or “none”].
Choose one moment when the rep had a meaningful decision to make.
Write a short scenario that ends at that moment and includes only
what the rep knew then. Give the team two plausible next moves
and ask which they’d choose and why.
Add a separate facilitator note with what happened next, what the
buyer confirmed, and what remains the team’s interpretation.
End with one behavior the team can practice in a similar deal.
How to Get Better Results From ChatGPT Sales Prompts
When a draft comes back vague, look at what you gave ChatGPT. “Write a follow-up for a good demo” leaves most of the email up to guesswork. “The buyer asked for annual pricing and said finance meets Thursday” gives it a reason to write and a deadline to work around.
👉 Paste the source closest to the task. For a follow-up, use the buyer’s last reply and any commitments from the call. For account research, include the original excerpt and its date. A six-month-old job listing may explain what the company was working on then.
👉 Show what you’d actually use. Paste a short email you’ve sent before, a discovery brief your team uses, or your manager’s coaching rubric. Tell ChatGPT which parts to follow. For an email sample, that might be the greeting, sentence length, and how you ask for a next step.
👉 Give missing information a place to go. Add a line like, “If the buyer hasn’t confirmed a date, leave it blank and ask me what to verify.” Use the same approach for budget, deal stage, and who signs off. In a buyer-facing draft, ask for a separate check below the email so the uncertainty is easy to spot.
👉 Revise the part that’s off. If the first line sounds generic, ask for three openings tied to the buyer’s actual words. If the email ends with two requests, tell ChatGPT which one matters and ask it to cut the other. Keep the rest of the draft if it’s already doing its job.
Before you send the email or update the deal, check the details someone else will act on. Did the buyer agree to meet Thursday, or did they say they’d know more by Thursday? That one distinction can change both the follow-up and what belongs in the CRM.
Foire aux questions
What are the best ChatGPT prompts for sales?
Start with prompts for tasks you repeat every week. Use a discovery prep prompt to pull three questions from a buyer’s email thread, a post-demo prompt to draft a follow-up from agreed next steps, or a pipeline review prompt to flag deals with missed commitments. Each gives ChatGPT a clear job and gives you something concrete to check.
How do you write a ChatGPT prompt for sales prospecting?
Give ChatGPT your ideal customer criteria, the account facts you’ve verified, and the decision you need to make. For example, ask it to compare five accounts against three must-have criteria and show the evidence for each rating. Have it flag anything your research hasn’t confirmed.
What are good ChatGPT prompts for sales enablement?
Use a real objection to run a buyer role-play, a call transcript and coaching rubric to review one skill, or approved product documentation to build a competitive battlecard. For role-play, ask ChatGPT to reply one turn at a time and quote the rep’s actual answers in the debrief.
Can ChatGPT find leads for sales prospecting?
With web search available, ChatGPT can help find companies that fit your criteria or identify potential contacts from public sources. It can also rank accounts from a list you already have. Check that each person still works at the company and owns the relevant workflow before reaching out.
How do you make AI-written sales emails sound like you?
Paste a short email you wrote and ask ChatGPT to match its level of formality, sentence length, and style of asking for a next step. Give it one specific detail to build the new email around. If the opener feels generic, revise that line before rewriting the whole email.
Can you use CRM data in ChatGPT sales prompts?
Yes, if your team has approved that data for use in ChatGPT. A short excerpt is often enough. For a follow-up, start with the buyer’s last reply and any commitments from the call. If you connect your CRM, check which records the integration can access before running the prompt.
Découvrez « folk », le CRM basé sur l'IA qui accompagne l'ensemble de votre processus commercial.
