The short answer
The best ChatGPT prompts for sales give OpenAI's ChatGPT a sales role, real context about the buyer and your product, rules against inventing facts and an exact output format. The 12 prompts here cover account research, buying groups, cold emails, follow-up sequences, CRM updates, discovery calls, objection role-play, battlecards, pipeline reviews and loss reviews. Since 7 October 2026, Plus and higher plans answer with GPT-6 Sol and Free uses GPT-6 Luna.
What makes a good ChatGPT prompt for sales
A good sales prompt reads like a brief you would give a sharp new hire: who the buyer is, what you know about them, what you sell, what outcome you want and the exact format you need back. The 12 ChatGPT prompts for sales on this page follow that pattern. Each one sets a role (account executive, sales coach, sales operations analyst), passes in your real material, adds rules that stop the most common AI failures, and asks for a format you can use straight away, such as a battlecard table or a CRM note.
The rules matter most in sales, where a confident mistake costs you a deal. Several prompts tell ChatGPT to mark missing information as "unknown" or [ADD PROOF] instead of filling the gap. Keep those lines. OpenAI's prompt engineering guide recommends stating clearly what the model should never do, and in sales the thing it must never do is invent a customer result.
The prompts are grouped into four jobs: prospecting, outreach and follow-up, calls and objections, and pipeline and deals. They work on any ChatGPT plan, but quality differs by plan. Since 7 October 2026, Plus, Pro, Business and Enterprise get GPT-6 Sol in the Chat tab and Free and Go get GPT-6 Luna. See our ChatGPT guide for what each plan includes.
ChatGPT prompts for prospecting and account research
Account research is the job where ChatGPT saves sales reps the most time, because it can read a company's site, news and job ads with web search and condense them into a brief in a couple of minutes. The first prompt asks for a link for every point and tells ChatGPT to leave out anything without a source. Open two or three of those links before the call. A citation shows where a claim came from, not that the summary is right.
The buying-group and checklist prompts turn your own deal history into something the team can act on. Paste a real export of won and lost deals. ChatGPT is good at spotting patterns in a table, but it will also find patterns in three data points that mean nothing, which is why the prompt asks it to flag thin evidence.
ChatGPT prompts for cold outreach and follow-up
Most AI cold emails fail for the same reason: they open with a generic compliment and pitch features. The cold email prompt forces one specific fact about the prospect and one verifiable result from a similar customer, caps the length at 90 words and bans the phrases buyers delete on sight. Asking for three versions with different openings gives you a small test instead of one guess.
The follow-up prompts are where reps reclaim the most admin time. Paste your raw notes or a call transcript and you get the recap email and the CRM update in one go. On paid plans, ChatGPT can also take an audio recording directly, since audio uploads arrived on 6 October 2026, so a recorded call can go straight to notes. Check your company's rules on recording and on putting buyer data into AI tools first.
ChatGPT prompts for discovery calls and objection handling
The role-play prompt is the most underused sales prompt there is. ChatGPT plays a sceptical buyer, keeps its concerns hidden the way real buyers do, and scores your answers when you type "feedback". Ten minutes of this before a big call is better practice than rereading a battlecard. Use voice mode on your phone for a more realistic rehearsal.
The objection prompt builds a shareable battlecard, but its value depends on your proof points. Give it real customer results, review quotes and security documents. Anything missing gets marked [ADD PROOF], which doubles as a list of the case studies your marketing team should write next.
ChatGPT prompts for pipeline reviews and closing deals
For the pipeline prompt, upload your CRM export as a spreadsheet. ChatGPT runs code to do the calculations, so ask it to show the formulas and check one total yourself. Data analysis is limited on the free plan, so this prompt works best on Go or above. Give it your own stage probabilities, because its defaults will not match your sales process.
The mutual action plan and loss review prompts are about discipline rather than speed. A dated plan agreed with your champion is one of the simplest ways to protect a close date, and a written, blame-free loss review is how a team stops losing deals for the same reason twice.
How to adapt these sales prompts
- Swap in your sales method. If your team uses MEDDICC, BANT or SPICED, name it in the prompt and list its fields so the CRM update matches your process.
- Keep a product fact block. Write 150 words on what you sell, who it is for, three verified results and your main competitors. Paste it at the top of any prompt, or store it in a ChatGPT project (available on every plan, including Free) so every chat starts with it.
- Add one email that worked. A real reply-winning email from your team teaches tone faster than any description. Anthropic's guide suggests 3 to 5 varied examples when you can.
- Tell it the reason. "Under 90 words because prospects read on their phone between meetings" gets better compliance than "keep it short".
- Never send unchecked. Read every email, verify every name and number, and remove anything you could not defend on a call.
Our prompt kits package prompts like these by role so a whole team can start from the same versions.
Which tool works best for sales prompts, and why
ChatGPT is the best general choice for sales reps: web research, writing, role-play, voice and spreadsheets in one tool. ChatGPT Business, for teams of 2 to 200, connects to Google Workspace, Slack, GitHub and Microsoft 365, and does not train on business data by default.
Claude by Anthropic writes more natural, less salesy emails and handles long call transcripts well. Teams on Salesforce should know that Anthropic released a Salesforce in Claude plugin in beta in September 2026, with accounts, opportunities, pipeline and 37 sales skills. Microsoft Copilot is the obvious pick if your company runs on Outlook and Teams, because a Microsoft 365 Copilot licence lets it draw on your organisation's mail and files through Microsoft Graph, under enterprise data protection. Perplexity is the strongest choice for account research when you want numbered sources on every answer.
A privacy rule for every rep: on personal ChatGPT accounts (Free, Go, Plus and Pro), chats can train OpenAI's models unless you switch off "Improve the model for everyone" in Settings, Data controls. Do not paste buyer contact details or deal terms into a personal account. For the full comparison of sales tools, read the best AI for sales and AI for sales teams.
Prompts to copy
Research an account before the first call
You are a senior account executive preparing for a first call. Use web search to research [company name] ([website]). Today is [date]. Find and summarise, with a link for every point: what the company sells and to whom, any news from the last 6 months (funding, leadership changes, launches, layoffs, expansion), open job ads that hint at priorities, and anything public that connects to the problem we solve, which is [problem our product solves]. Then give me: three likely business priorities ranked by evidence, two smart questions to ask that show I did my homework, and one reason they may not need us. If a point has no source, leave it out.
You get: A sourced one-page account brief with priorities, questions and an honest reason they might not buy
Map the buying group for a deal
Act as an enterprise sales coach. We are selling [product] to [company], a [size and industry] business. The people I know so far: [name, title, what they said or did] For a deal like this, list the roles that usually take part in the decision (economic buyer, champion, technical evaluator, user, blocker, procurement). For each role, say whether I have someone in it, what that person cares about most, the question that will make or break their support, and how I could get access if the role is empty. Return a table, then the two biggest gaps to close before [next milestone].
You get: A buying-group map showing who is covered, what each role needs and the gaps to close
Write a sharp ideal-account checklist for prospecting
Here are our last [number] closed-won deals and [number] closed-lost deals with industry, size, region, deal value, sales cycle and the reason won or lost: <deals> [paste] </deals> Act as a sales operations analyst. Find what separates won from lost deals and turn it into a prospecting checklist of 6 to 8 yes or no questions an SDR can answer in under five minutes using LinkedIn and the company website. Give each question a weight from 1 to 3 and explain it with the data. Flag any pattern based on fewer than three deals.
You get: A weighted, evidence-based checklist your team can use to qualify accounts in minutes
Write a cold email that earns a reply
You are a B2B copywriter who writes outbound emails that get replies, not just opens. Prospect: [name, title, company]. What I know about them: [one specific fact, e.g. a post they wrote, a hire, a launch]. Our product: [one-line description]. The result a similar customer got: [verifiable result]. My ask: [e.g. a 15-minute call next week]. Write three versions under 90 words each: one that opens with the specific fact, one that opens with the customer result, one that opens with a question about their likely priority. Each needs a subject line under 6 words, no flattery, no 'I hope this finds you well', one clear ask, and a plain sign-off. After the three, tell me which you would send and why in one sentence.
You get: Three short, personalised cold emails with subject lines and a recommendation
Build a multi-touch follow-up sequence
Act as a sales development manager. A prospect, [title] at [company], opened my first email but did not reply. Here is that email: [paste]. Write a sequence of four more touches over 14 days across email and LinkedIn. Each touch must add something new: a relevant insight, a short customer story, a useful resource or a direct question. Give the day, channel, subject line if email, and the message (under 75 words for email, under 300 characters for LinkedIn). The last touch is a polite break-up message that leaves the door open. No 'just following up' or 'bumping this up'.
You get: A four-touch, 14-day sequence where every message gives the prospect a new reason to reply
Turn call notes into a follow-up email and CRM update
Here are my notes from a sales call with [name, title] at [company] on [date]: <notes> [paste notes or transcript] </notes> Do two things. 1. Write the follow-up email: thank them in one line, recap their three main problems in their own words, confirm agreed next steps with owners and dates, and attach nothing I did not mention. Under 150 words. 2. Write a CRM update with these fields: pain points, budget, authority, timeline, competitors mentioned, next step, risk level (low, medium, high) and why. Mark anything that was not discussed as 'unknown' rather than guessing.
You get: A ready-to-send recap email plus a CRM note that marks gaps as unknown
Prepare a discovery call plan
You are a sales coach preparing a rep for a 30-minute discovery call with [title] at [company]. What we know: [account notes]. Our product solves [problem] for [type of customer]. Write a call plan split into timed blocks that add up to 30 minutes, with: a 2-sentence opener that sets the agenda and confirms the time, 10 open questions in a logical order (current situation, impact, past attempts, decision process, timing), a follow-up for each question to use if the answer is vague, the signals that this is a qualified opportunity and the signals that it is not, the two questions to keep if the call is cut to 15 minutes, and a closing line that secures a dated next step. Keep each question under 20 words and do not pitch the product before minute 20.
You get: A timed discovery call plan with questions, follow-ups and qualification signals
Practise objection handling with a tough buyer
Let's role-play. You are [title] at [company type], a busy and sceptical buyer. You currently use [competitor or current process]. Your real concerns are [e.g. price, switching effort, security review], but do not reveal them all at once. I am the sales rep selling [product]. Stay in character, push back the way a real buyer would and keep each reply under 60 words. After I type 'feedback', step out of the role and score my answers from 1 to 5 on listening, relevance and confidence, quote my weakest answer and show a stronger version.
You get: A realistic practice buyer, then specific scored feedback on your answers
Write answers to your top objections
Act as a sales enablement lead. Our product: [description]. Our main competitors: [list]. Here are the objections our reps hear most, with notes on what has worked before: <objections> [paste] </objections> For each objection, write: what the buyer is really worried about, a short acknowledgement, a response under 60 words that uses a fact, example or question rather than pressure, and a proof point we can back up. Mark any proof point I did not give you as [ADD PROOF]. Format it as a battlecard table I can share with the team.
You get: A shareable objection battlecard with honest responses and gaps marked for proof
Review the pipeline like a sales manager
I have uploaded our pipeline export as a spreadsheet with these columns: [list columns, e.g. deal, stage, value, close date, last activity, next step]. Act as a sales manager preparing for the weekly forecast call. Calculate pipeline by stage and the total weighted value using these stage probabilities: [list]. Show your formulas. Then list deals that look at risk (no activity for 14 days, close date in the past, no next step, or stuck in one stage for more than [number] days), and give one specific action for each. Finish with a realistic forecast for [period] and the reasoning in three sentences.
You get: A weighted pipeline view with at-risk deals, actions and a reasoned forecast
Write a mutual action plan for a late-stage deal
We are in late stage with [company]. Target signature date: [date]. Steps still needed on their side: [e.g. security review, legal, budget approval]. Steps on our side: [e.g. pilot report, pricing proposal]. People involved: [names and roles]. Create a mutual action plan as a table with columns: step, owner (theirs or ours), due date, status and what 'done' means. Work backwards from the signature date and leave buffer for legal. Then write a short, friendly email I can send to my champion to agree the plan.
You get: A dated mutual action plan and the email to agree it with your champion
Write a deal loss review that teaches something
We lost a deal with [company] to [competitor or no decision]. Here is everything I have: emails, call notes and the buyer's stated reason. <deal> [paste] </deal> Act as a neutral deal reviewer. Build a short timeline of the deal, identify the moment it most likely turned, separate what we could control from what we could not, and list three lessons phrased as specific behaviours (for example 'confirm the economic buyer before sending pricing'). Do not blame individuals. Under 400 words.
You get: A blame-free loss review with a timeline, the turning point and three concrete lessons