25 ChatGPT Prompts for Sales Reps and SDRs (2026)

The sales reps doing well with ChatGPT in 2026 are not the ones generating 200 cold emails a day from a template. Those reps got their domains blocked in the first quarter, and their prospects added them to the do-not-call list of three competitors before lunch. The reps doing well are the ones using ChatGPT to compress the 90 minutes of pre-call research, post-call admin, and CRM hygiene into 20 minutes, so the rest of the day is the actual selling work that pays.

This post is 25 prompts a working B2B sales rep can paste today, organised by where they sit in the deal cycle. The audience is account executives, SDRs, and senior sellers carrying a quota of $400k to$ 3M ARR a year, working two to four channels of outbound and a CRM that is half-organised on a good day. Every prompt has four parts: role, task, constraints, output spec. Copy the prompt, substitute the bracketed brief, paste into ChatGPT. We ran one of them live at the end of the post and reproduced the actual response so you can see what you are working with.

How to use these prompts

Treat ChatGPT as the junior researcher who never sleeps and never sends anything for you. The model is reliable for first-draft outreach, structured pre-call briefs, and post-call note synthesis when you give it real account context. It is not reliable for anything that goes into a prospect's inbox without a human edit. Two operational rules before you paste.

First, never paste private account data into the free tier of ChatGPT. Buyer names, deal values, contract terms, anything covered by your CRM's data-handling policy, all belong in your CRM or in an enterprise LLM with explicit data terms. The prompts in this post use bracketed placeholders so you can either anonymize the input or run them inside a paid product. Second, every output is a draft your judgment has to sit on top of. The prospect on the other end will tell you in 30 seconds whether the email sounded like a human or a model. Edit before sending.

Cold outreach and prospecting

These five prompts cover the top of the funnel: the first email, the first call, the second touch, the bump message, and the breakup. The model is reliable here because cold outreach is highly templated and the brief constraints are short.

Prompt 1: First-touch cold email to a named prospect

Prompt: "You are a senior B2B sales rep at a software company. Write a cold first-touch email to a named prospect. Brief: prospect name and title: [name, title]. Company: [company and one-sentence on what they do]. Trigger event you are using as the hook: [news, hire, funding round, product launch, content they published]. Specific pain point you believe their team has based on the trigger: [pain point]. Your product in one sentence (no jargon): [product]. One peer customer they would recognise: [customer name]. Constraints: subject line under 50 characters and specific to the trigger (not 'Quick question'). Body under 90 words. Three short paragraphs maximum: hook tied to trigger, peer-customer proof point in one sentence, one specific question that earns a reply. No 'Hope this finds you well'. No 'I came across your profile'. No m-dashes. Output: subject line on its own line, then the email body."

Prompt 2: Cold email follow-up after no reply

Prompt: "You are a senior B2B sales rep writing the second follow-up to a cold first-touch email that received no reply. Brief: prospect name: [name]. First email subject: [subject]. First email date: [date]. The hook from the first email: [hook in one line]. A new angle for this follow-up (not a guilt trip, a new piece of value): [new angle, e.g., a piece of research, a customer outcome number, a new product capability]. Constraints: under 70 words. Single paragraph plus a one-line PS. The PS contains the new angle. Does not ask 'did you see my last email'. Does not include a calendar link. No m-dashes. Output: subject line and body."

Prompt 3: LinkedIn connection request after a relevant post

Prompt: "You are a sales rep sending a LinkedIn connection request to a prospect after they posted something relevant to your product. Brief: prospect name and title: [name, title]. Their post topic in one sentence: [topic]. The specific thing they said you want to reference: [quote or paraphrase]. Constraints: under 280 characters total (LinkedIn limit). Names the specific thing from their post. Asks one open question. Does not mention your product or company in this message. No m-dashes. Output: the connection request message only."

Prompt 4: Bump email after a discovery call no-show

Prompt: "You are a sales rep sending a polite bump email after the prospect no-showed a discovery call. Brief: prospect name: [name]. Scheduled date and time: [date, time]. The original meeting topic: [topic]. Whether they have responded at all since (yes/no): [yes/no]. Constraints: under 80 words. Single paragraph. No guilt-trip language ('I waited for you'). Offers two specific reschedule windows (a morning and an afternoon). No m-dashes. Output: subject and body."

Prompt 5: Breakup email after a cold sequence with no replies

Prompt: "You are a sales rep sending the final breakup email after a cold outreach sequence with no replies. Brief: prospect name: [name]. Number of touches sent in the sequence: [count]. The original value proposition: [value prop in one sentence]. Constraints: under 70 words. States the rep is closing the file on the outreach but leaves the door open. Asks one closing question that allows a yes/no/maybe response. Includes a 'reply with STOP if you want me to never reach out again' line. No passive-aggressive tone. No m-dashes. Output: subject and body."

Discovery and qualification

These five prompts help the rep walk into a discovery call ready to ask sharper questions than the rep on the other side of the chair would expect from a vendor.

Prompt 6: Pre-discovery research brief on a target account

Prompt: "You are a senior B2B sales rep preparing for a 30-minute discovery call with a prospect. Brief: company name and industry: [company, industry]. Headcount band: [band]. The two job titles you are meeting: [titles]. What you already know from the inbound or outbound trigger: [paste 4-6 bullets]. What you sell in one sentence: [your product]. Constraints: produce a 1-page pre-call brief covering: the company's likely current state, the three pain hypotheses worth testing on the call (ranked by likelihood), the two unknowns you cannot infer and must ask about, the two questions to ask if you suspect this is a low-fit deal. No filler. No m-dashes. Output: structured pre-call brief."

Prompt 7: Discovery call question list (open-ended only)

Prompt: "You are a senior sales rep generating a discovery call question list for a 30-minute call. Brief: prospect industry and likely pain: [industry, pain]. The sales methodology you use (MEDDIC, SPIN, BANT, command-of-the-message): [methodology]. The buyer's seniority: [seniority]. Constraints: produce 12 questions in the order you would ask them. All open-ended (no yes/no). Group them: 2 warm-up on their role and how they spend their week, 5 on the current state and problem severity, 3 on how they would evaluate a solution, 2 on internal politics and decision process. No leading questions. Each question gets a one-sentence rationale for what answer you are listening for. No m-dashes. Output: numbered list."

Prompt 8: Qualification-call summary email to send the same day

Prompt: "You are a sales rep writing the same-day summary email after a 30-minute qualification call. Brief: prospect name: [name]. Three things they said on the call (paraphrased): [list]. The next step you agreed on (date, format, attendees): [next step]. One open question that came up on the call: [question]. Constraints: under 220 words. Names the three things back to them so they see you listened. Confirms the next step and attendees. Asks the one open question explicitly. Includes a clear close: 'Reply if I got anything wrong'. No 'great call today'. No 'as discussed'. No m-dashes. Output: subject and body."

Prompt 9: Champion-building question for a working session

Prompt: "You are a senior sales rep preparing questions for a 60-minute working session with a likely champion at a target account. Brief: champion name and title: [name, title]. What they care about based on prior conversations: [list]. The deal value at stake: [ACV]. The decision the champion is trying to influence internally: [decision]. Constraints: produce 8 questions designed to surface what the champion needs from you to win the internal argument. Questions get progressively more specific. Two questions must surface political risks (who pushes back, who the champion needs to convince, who has veto). No m-dashes. Output: numbered list with a one-sentence rationale for each."

Prompt 10: Disqualification email when the deal does not fit

Prompt: "You are a sales rep writing a polite disqualification email to a prospect who you have decided does not fit your product. Brief: prospect name: [name]. The reason it does not fit (size, use case, timeline, budget): [reason]. Whether you want to leave the door open for a future fit (yes/no): [yes/no]. Constraints: under 130 words. Direct but warm. Names the reason it does not fit specifically. If leaving the door open, names the change that would make it a fit (e.g., 'when you cross 50 reps') and offers to reconnect then. No fake compliments. No m-dashes. Output: subject and body."

Demos and call prep

The model is useful for compressing pre-call research and demo prep. It is not useful for running the demo. The rep does the demo.

Prompt 11: Demo script tailored to a specific pain

Prompt: "You are a senior sales rep building a 20-minute demo script for a prospect with a specific pain. Brief: prospect name and role: [name, role]. The single pain you are solving in the demo: [pain]. Your product features that map to that pain (in order of relevance): [list of 3-5 features]. What you already know is a non-pain for them (so do not show): [list]. The success criterion they named on the discovery call: [criterion]. Constraints: produce a 20-minute timed demo script with 4 sections. Each section: timing, what to show on screen, the specific outcome moment to point at, the verbatim line to say at the end of the section (under 25 words). Ends with a clear ask: book the technical-fit call. No feature-tour structure. No m-dashes. Output: timed demo script."

Prompt 12: Demo objection-handling brief

Prompt: "You are a senior sales rep preparing for demo objections from a specific prospect. Brief: prospect role: [role]. The three objections you most expect from this profile: [list]. Your honest counter for each (not corporate spin): [list]. Constraints: produce an objection-handling brief with one row per objection. Columns: Objection (verbatim how they will say it), Why they are raising it (one sentence), Your one-sentence response, The follow-up question to ask back. No defensive language. No m-dashes. Output: markdown table."

Prompt 13: Pre-demo agenda email to the prospect

Prompt: "You are a sales rep sending the pre-demo agenda email to a prospect 24 hours before a 30-minute demo. Brief: prospect name and the names of others attending: [list]. The two things you will show on the call: [list]. The single decision the prospect is trying to make from the demo: [decision]. The two things you need from them in advance: [list, e.g., 'sample data', 'a current workflow screenshot']. Constraints: under 160 words. Names the others attending so they see you noticed. Includes the decision-frame explicitly. Asks for the two things with a clear deadline. No 'looking forward to'. No m-dashes. Output: subject and body."

Prompt 14: Demo recap email with the proposed next step

Prompt: "You are a sales rep writing the demo recap email same-day after a successful demo. Brief: prospect name: [name]. The two things from the demo that the prospect responded to most: [list]. The one open concern still on the table: [concern]. The proposed next step with date, format, and attendees: [next step]. Constraints: under 220 words. Leads with the two things they responded to, names them specifically (not generic features). Acknowledges the open concern in one sentence and proposes how to address it. Confirms the next step explicitly. No 'great demo today'. No m-dashes. Output: subject and body."

Prompt: "You are a sales rep preparing a one-page deal review for your sales manager. Brief: account name and ACV: [name, ACV]. Stage in the pipeline: [stage]. The buyer's stated outcome they want: [outcome]. The three things working in your favour: [list]. The three risks: [list]. The single thing you need from the manager (a co-call, a discount approval, a customer reference, an exec sponsor intro): [ask]. Constraints: one page. Five sections: Deal Snapshot, Buyer Outcome, Working in Our Favour, Risks, The Ask. No filler. No m-dashes. Output: markdown one-pager."

Proposals, negotiation, and close

The model produces serviceable first drafts of proposal language and negotiation prep. The final wording on price and terms always gets edited.

Prompt 16: Proposal email with pricing options

Prompt: "You are a senior sales rep sending a proposal email with two pricing options to a qualified prospect. Brief: prospect name and company: [name, company]. The outcome they told you they want: [outcome]. Option A (pricing and what it includes): [option A]. Option B (pricing and what it includes): [option B]. Your recommendation between A and B (and why, in one sentence): [recommendation]. Constraints: under 250 words. Leads with the outcome they told you (not with the price). Presents both options in a markdown table. Names your recommendation explicitly. Closes with a one-line ask for a 20-minute conversation to walk through, not for an email reply. No 'amazing partnership'. No m-dashes. Output: subject, body, and the pricing table inside the body."

Prompt 17: Mutual action plan document for a high-value deal

Prompt: "You are a senior sales rep drafting the mutual action plan (MAP) document for a $250k+ deal. Brief: account name: [name]. Decision date the prospect named: [date]. The decision criteria the buyer agreed to: [list]. The internal stakeholders on the buyer side and what each owns: [list with role and what they evaluate]. Your team's side of the work: [list]. Constraints: one page. The plan is a markdown table with columns: Date, Step, Owner (Buyer or Seller), Decision Criterion this Step Validates. Reverse-engineered from the decision date. Includes the kickoff conditions if signed. No m-dashes. Output: MAP markdown table preceded by a one-paragraph intro."

Prompt 18: Discount-request response

Prompt: "You are a senior sales rep responding to a prospect who has asked for a 25% discount on a signed proposal. Brief: account name and ACV: [name, ACV]. Their stated rationale for the discount: [rationale]. The maximum discount you are authorised to give: [percentage]. Two concessions you can offer that are not price (longer term, smaller initial commitment, deferred payment, additional services): [list]. Constraints: under 200 words. Does not concede price in the first paragraph. Asks one question about the rationale before discussing options. Presents the non-price concessions before any price concession. No 'I will need to talk to my manager' unless you actually will. No m-dashes. Output: subject and body."

Prompt 19: Close-or-kill email to a stalled deal

Prompt: "You are a senior sales rep writing a close-or-kill email to a deal that has stalled for 45 days after a strong demo. Brief: prospect name: [name]. The deal value and stage: [ACV, stage]. The last meaningful interaction (date and what happened): [date, what]. The last three follow-up attempts and outcomes (or non-outcomes): [list]. Constraints: under 130 words. Names the silence directly without guilt-tripping. Asks one of: 'should we keep talking', 'should we revisit in [quarter]', or 'should we close the file'. Does not include a calendar link. No m-dashes. Output: subject and body."

Prompt 20: Contract-sent confirmation email

Prompt: "You are a sales rep writing the contract-sent confirmation email to a prospect who has verbally agreed to sign. Brief: prospect name and signer name (if different): [list]. Contract sent date and via what platform (DocuSign, PandaDoc, etc.): [list]. The agreed terms in 3 bullets: [list]. The target signature date: [date]. The two things the prospect needs to confirm before signing (e.g., billing email, AP contact): [list]. Constraints: under 180 words. Confirms the three terms back to them explicitly. Names the signer if different. Asks for the two pre-sign confirmations with a clear deadline. No 'congratulations'. No m-dashes. Output: subject and body."

Pipeline hygiene and post-call admin

The work that does not directly close a deal but quietly determines whether the rep hits the number. ChatGPT shines here because the audience is the rep's own CRM and manager.

Prompt 21: CRM note from raw call notes

Prompt: "You are a senior sales rep cleaning up raw call notes into a structured CRM entry. Brief: paste the raw notes from a 30-minute call: [paste notes]. Constraints: produce a CRM entry with these fields filled in: 1) Outcome of the call in one sentence, 2) Three things the prospect said verbatim (with quotes), 3) The pain points they confirmed, 4) The objections they raised, 5) The next step agreed (date, format, attendees), 6) MEDDIC fields filled in where the call surfaced them (Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion). No invented information. If a field cannot be filled from the notes, mark it as 'not surfaced'. No m-dashes. Output: structured entry."

Prompt 22: Weekly pipeline review note for the manager

Prompt: "You are a senior sales rep writing the weekly pipeline review note for your manager. Brief: top 5 deals in the pipeline with stage, ACV, and the one thing that moved (or did not move) this week: [paste list]. Your forecast for the quarter: [number]. The one deal you need help on: [name and what kind of help]. Constraints: under 250 words. Number-led opening sentence ('Five active deals, $X total,$ Y forecast for the quarter'). Names the deal that moved most. Names the one ask explicitly. No 'kicking off another great week'. No m-dashes. Output: prose note."

Prompt 23: Lost-deal post-mortem entry

Prompt: "You are a sales rep filling out the lost-deal post-mortem for a recently closed-lost deal. Brief: account name: [name]. Lost ACV: [amount]. Stage at which the deal died: [stage]. Stated reason for the loss: [reason]. What you believe is the real reason: [your read]. The competitor (or no-decision): [competitor]. Three things you would do differently if you ran the deal again: [list]. Constraints: one page. Six sections: Account Snapshot, Stated vs Real Reason, Decision Process Timeline (date / step / what happened), What We Missed, What the Competitor Did Better (if applicable), Operational Takeaway for the Team. No blame language. No m-dashes. Output: markdown one-pager."

Prompt: "You are a senior sales rep drafting your quarterly territory plan one-pager. Brief: quota for the quarter: [number]. Current pipeline by stage: [list]. The five accounts you believe are most likely to close this quarter, ranked: [list]. The five accounts to prospect into this quarter that are not yet in pipeline: [list]. Two operational gaps that are slowing you down: [list]. Constraints: one page. Five sections: Quota, Pipeline Health, Top 5 Likely Close, Top 5 New Prospects, Operational Gaps. No filler. No m-dashes. Output: markdown one-pager."

Prompt 25: Handoff note to customer success after a closed-won deal

Prompt: "You are a sales rep writing the handoff note to the customer success team after closing a deal. Brief: account name and ACV: [name, ACV]. Product they bought: [product]. Their stated business outcome: [outcome]. The promise made during the sales process that CS needs to honour: [promise]. The two stakeholders to be aware of (one champion, one skeptic): [list with names and roles]. The political risk to know about: [risk]. Constraints: under 250 words. Six labelled paragraphs: Outcome, Promise, Stakeholders, Political Risk, First-30-Day Recommended Plays, Open Red Flags. No 'great win for the team'. No m-dashes. Output: structured handoff note."

What this prompt actually returns (tested 2026-05-13)

We ran Prompt 1 (first-touch cold email to a named prospect) on free ChatGPT with the example brief baked in. The substituted brief was: prospect is Priya Mehta, VP of RevOps at Atlas Logistics (a mid-market freight brokerage, ~280 employees, $180M annual revenue); trigger event is Atlas's CFO LinkedIn post last week announcing a 2026 initiative to consolidate the sales tech stack and cut tooling spend by 25%; specific pain is protecting forecast accuracy through a stack consolidation, which usually lands on the RevOps lead's bag; product is a forecasting platform that plugs into HubSpot and replaces three tools (forecasting spreadsheet, pipeline analytics, board-reporting dashboard) with one; peer customer is Redwood Freight, a similar-sized brokerage that consolidated four tools into ours last year and held forecast accuracy above 92% through the cutover. Reproduced verbatim:

Subject: 25% tooling cut without forecast risk

Priya,
Saw Atlas Logistics' 2026 initiative to cut sales tooling spend by 25% while consolidating the stack. The hard part is usually protecting forecast accuracy during the transition, and RevOps tends to own the fallout if visibility slips.
Redwood Freight, a similar-sized brokerage, replaced four forecasting and reporting tools with our platform last year and kept forecast accuracy above 92% through the cutover.
Would it be worth a 15-minute look at how they consolidated without losing forecasting confidence?

Editor's notes. This is closer to a "send today after light edits" cold email than anything we have seen from an unprimed model. The subject is 38 characters and number-anchored to the trigger event, the body is 75 words against the 90-word cap, the opening is just "Priya," with no filler greeting, the hook in paragraph one ties directly to the CFO's post, and the Redwood proof point includes a specific number (92%). The model also avoided every banned phrase the prompt named. That is the win.

Four things to edit before this goes out. First, the closing question ("Would it be worth a 15-minute look at how they consolidated without losing forecasting confidence?") is a yes/no question. The prompt asked for "one specific question that earns a reply", and a yes/no question is the lower-quality version of that. A sharper close: "Who on your team owns the consolidation evaluation, and would they want to see how Redwood handled the cutover?" That asks for a name, which is harder to ignore than a yes/no.

Second, the pain line ("RevOps tends to own the fallout if visibility slips") generalises. It is the right pain, but the model wrote it as a generic claim rather than a Priya-specific one. Rewriting to "you usually own the fallout if visibility slips" is one word longer and lands much harder. Cold email lives on the second person.

Third, the subject ("25% tooling cut without forecast risk") is fine but does not surface the Redwood reference. A test alternative worth A/B-ing: "How Redwood cut 25% without forecast loss". Naming a peer in the subject often opens more emails from a buyer who has heard of them.

Fourth, the apostrophe in "Atlas Logistics'" came out as a curly right-single-quote character rather than a straight apostrophe. Most email clients render it fine, but some Outlook templates and CRMs render the curly quote as a different glyph or as "Atlas Logistics’". On a real send, normalise to straight quotes before pasting.

The model also missed one move worth adding by hand: there is no PS line. In cold outreach the PS often gets more opens than the body, and the brief did not explicitly ban one. A short PS like "PS: happy to send the Redwood case study before any meeting" gives a no-meeting fallback path that often returns a reply by itself.

FAQ

Will prospects know I used ChatGPT to write the email?

If you send the model's first draft without editing, yes. Buyers in 2026 have read enough AI-drafted outreach to recognise the rhythm. The tell is usually a too-clean transition, a three-noun list ("scale, speed, and efficiency"), a hedge phrase, or a corporate compliment that no human would write to a stranger. Edit those out and the email reads as yours. If you cannot find anything to edit in the model's output, the model has not yet matched your voice.

Are AI-drafted cold emails getting blocked or marked as spam?

The model's text itself is not the spam signal. The signals are sender reputation, sending volume, link-to-text ratio, and prior bounce rate from your domain. ChatGPT does not change any of those. What does change spam risk is using ChatGPT to send 200 cold emails a day from a single mailbox, because that volume is what gets domains blocked, not the wording. Keep daily send volume per mailbox under the platform's recommended limit and warm new mailboxes before scaling.

Should I paste my CRM data into the free tier of ChatGPT?

No. Account names, contract values, contact details, prior call notes, anything you would not want surfaced as a training example, should stay inside your CRM or in an enterprise LLM where the data terms have been reviewed by your security team. The prompts in this post use bracketed placeholders for a reason: anonymise the input, or run the prompts in a paid product that explicitly excludes your data from training.

What is the highest-payoff use of ChatGPT for a working sales rep?

Post-call note synthesis (Prompt 21) and pre-call research briefs (Prompt 6). Both compress 20 to 40 minutes of admin into 3 to 5 minutes, several times a day. The compounding savings across a week are larger than any single outbound email the model drafts. The reps who get the most value from ChatGPT use it for the surrounding work, not for the customer-facing wording.

Should sales reps pay for ChatGPT Plus?

For most full-time reps carrying quota, the Plus tier pays for itself in saved time within a month, mostly through longer context windows that let you paste raw notes from a 45-minute call without splitting them up. Reps at companies with an enterprise LLM contract (ChatGPT Enterprise, Claude for Work, an internal model) should use that instead so account data stays inside the data-handling envelope your security team has approved.

Where to go from here

The 25 prompts above cover the formats most working B2B sales reps spend time on. The savings compound when you save the three or four prompts that fit your motion into a file you open every day. Pick three from this list: one outbound prompt, one post-call prompt, one deal-review prompt. Run them on real account work this week. The ones that produced something you could ship after a 2-minute edit, you keep. The rest, you adjust or drop. After two weeks you will have a personal prompt library that is sharper than any list a blog can give you.

The single most important habit to build: the model writes the draft, you write the message that hits send. If your name is on the email, you wrote the email. The model just took the blank-page tax off the top.