17 Powerful Claude Prompts for Sales That Drive Results

Whether you are prospecting, reviving a stalled deal, or preparing for the call that decides your quarter, Claude is only as useful as the context you hand it. Claude prompts for sales work by supplying the buyer, the situation and the outcome you are trying to reach — the things a template cannot know.

Below are 17 prompts grouped into prospecting and outreach, discovery, pitching and objections, moving deals forward, and closing and learning. Each has a short note on when to reach for it. Paste any of them into Claude, or run them in Chat Smith, an AI chat app that keeps Claude, GPT and Gemini together so you can compare how each one writes to the same buyer.

Sales rewards precision and contextual awareness. Generic tools hand you templates; a prompt carrying the buyer's role, their company's situation and the specific pain you are addressing hands you something you can send without rewriting it first.

A long-context model such as Claude Sonnet 5 can hold a discovery transcript, an account's public information and your last three emails in one conversation, which is what makes deal-level strategy possible rather than message-level drafting. Three things separate a prompt that returns something sendable from one that returns filler.

Name the buyer, not the segment — "a VP of Sales at a 200-person SaaS company already using a competitor" returns usable copy. "A prospect" returns a template.

State the job of this touch — a reply, a meeting, an internal forward. Naming the one outcome stops the output hedging across three.

Ban the tells — no "I hope this finds you well", no "just circling back", no manufactured urgency. Buyers pattern-match on these faster than on anything you actually say.

Four prompts for the top of the funnel, where volume tempts everyone into generic messaging. Each one forces the specific detail back in — useful whether you draft in an AI email writer or straight into your sequencer.

1. The cold email writer

Cold email is the highest-leverage outbound activity and the easiest to get wrong: too long, too generic, too much about the sender. Give the buyer and one specific pain and the output changes completely.

Write a cold email to a [job title] at a [company size and type]. Our product is [what it does and the outcome it produces]. Keep it under 100 words, open with a pain-point-first hook rather than an introduction, name one specific outcome, and close with a single low-friction ask for [the ask]. Do not use "I hope this finds you well" or any variation. Give me three versions with different hooks.

2. The pre-call research brief

Nothing signals respect to a buyer like a rep who clearly did the reading. This turns whatever you could find into something usable on the call itself.

Here is background on my prospect: [paste their LinkedIn bio, the company About page, and any recent news]. I sell [product and the problem it solves]. Based on this: (1) identify three business priorities this person likely has right now, (2) flag the pain points most relevant to my product, (3) suggest two openers referencing something specific to their situation, and (4) list the risks I should probe for on the call. Mark anything you inferred rather than read.

3. The social touch writer

Email is not the only channel, and a comment or connection note that lands can buy the reply your email did not. Different channel, different register — closer to what a social media post generator produces than to an email.

I want to open a conversation with a [job title] at [company] on LinkedIn rather than by email. Context I have: [paste their recent post, role change, or company news]. I sell [product]. Write: (1) a connection note under 300 characters that does not pitch, (2) a comment on their post that adds something rather than agreeing with it, and (3) a follow-up message for if they accept, referencing the first touch without repeating it. No compliments that could apply to anyone.

4. The ICP reality check

Poor list quality looks like a messaging problem for months before anyone questions the list. This checks who you are actually reaching out to.

Here is who we currently target: [your ICP as written]. Here is what our best customers actually look like: [five best accounts with size, industry, use case, and what triggered the purchase]. Here is who churns or never closes: [describe them]. Tell me where my stated ICP and my real ICP diverge, which segment I am wasting outbound effort on, what trigger event is common across the best customers, and the three qualifying questions that would filter for it earliest.

Discovery decides most deals long before the pitch happens. These three sharpen the questions, extract what the call actually told you, and map who has to say yes.

5. The discovery coach

Surface-level questions get vague answers. A framework plus questions written for one specific buyer role changes what a call surfaces.

I want to improve my discovery questioning. Right now I ask things like [paste the questions you actually use] and get vague answers. I sell [product] to [buyer role and company type]. Give me: (1) a framework for deeper discovery and why it suits this sale, (2) ten questions tailored to this buyer role that surface budget, urgency and internal politics, and (3) two practice scenarios — one engaged prospect, one evasive — to roleplay against you, one question at a time.

6. The call debrief

The gap between what a prospect said and what you heard is where deals are lost. Debrief while it is fresh, before optimism edits the memory.

Here are my notes from a discovery call with a [role] at [company]: [paste them]. Analyse them: (1) what did they actually confirm versus what am I assuming, (2) which of my qualifying criteria are still unproven, (3) what did they say that should worry me, (4) what question should I have asked and did not, and (5) what are the three most likely reasons this deal does not close? Then draft the follow-up email with a specific next step in it.

7. The buying group map

A single champion is not a deal. This works out who else has to be convinced and what each of them personally needs to believe.

Here is what I know about the buying group at [company]: [each person, their role, what they have said, how they have behaved]. Deal size: [amount]. Map it: who is the likely economic buyer, who is my champion, who is a potential blocker and why, and who have I not met but probably need to. For each person, state what they personally need to believe before this closes, and what I should send or say to get them there.

Four prompts for the stretch where positioning matters most. The discipline is the same one behind good AI ad copy generator output: lead with their situation, not your feature list.

8. The tailored pitch builder

Generic pitches lose deals; one built around the buyer's own initiative and language wins them. Worth doing properly on any account you care about.

I am pitching [product] to a [buyer role] at [company description]. Their key challenge is [describe it, with a number if you have one] and they have recently [any relevant initiative or event]. Build a five-part pitch: (1) an opening that acknowledges their problem without mentioning my product, (2) a bridge to what that problem costs them, (3) our solution framed around their specific situation, (4) the single most relevant proof point, (5) a clear next step. Use their industry's language, not ours.

9. The objection rehearsal

Walking in with prepared responses buys a composure buyers notice. Ask for the objections you are dreading, not the easy ones.

I am demoing [product] to a [role] at [company description]. They currently use [incumbent or status quo]. Generate the six most likely objections they will raise — including [name the two you expect] — and write a calm, evidence-backed response to each that acknowledges the concern before reframing it. For each, also give me the question that would reveal whether the objection is the real reason or a proxy for something else.

10. The competitor positioning analysis

Deals where you are not the only option are won on differentiation, not disparagement. Pull the pattern out of the competitor's own messaging and reviews.

Here is the homepage copy and review summary for our main competitor: [paste it]. We sell in the same category: [describe your product]. Analyse their positioning: (1) which benefits do they lead with, (2) which customer complaints recur in the reviews, (3) where are they weakest for a buyer who values [what you are actually better at], and (4) write three differentiation talking points I can use in conversation without attacking them. Flag any claim of theirs I should not try to contest.

11. The pricing conversation prep

"Too expensive" is usually a value or timing problem wearing a price costume. Prepare for the specific version you are going to hear.

I need to prepare for the pricing conversation on this deal: [product, price, deal size, and what the buyer has said so far]. Their likely alternatives: [competitors, or doing nothing]. Prepare: (1) how to present the price so value framing comes first, (2) responses to the three most likely price objections, (3) what I can offer instead of a discount — terms, scope, timing, pilot — ranked by what costs us least, (4) the questions that reveal whether this is a budget problem, a value problem or a tactic, and (5) the point at which I should be willing to lose the deal.

Three prompts for the middle of the pipeline, where deals go quiet. Around them, an AI proposal generator handles the document and an AI math calculator keeps the ROI arithmetic honest.

12. The follow-up sequence

Deals close on consistent follow-up that adds something each time. "Did you get a chance to review?" is not follow-up.

Write a four-email follow-up sequence for a prospect who [attended a demo / went quiet after a proposal] and has not replied in [X days]. What we sell: [product]. What they told me matters to them: [paste it]. Each email should add a new piece of value — a number, a proof point, a resource, a question — and stay under 80 words. Space them [intervals]. The last should be a graceful close that leaves the door open without guilt-tripping them.

13. The business case for the economic buyer

Deals stall when the champion loves the product and the economic buyer cannot see the return. This translates benefits into the language finance uses.

Help me build a one-page business case for a [economic buyer's role]. My product costs [price per year]. Their team currently spends roughly [hours per week] on [the task] across [number] people at a fully-loaded cost of about [rate per hour]. Build the ROI model: annual cost of the current process, annual saving assuming [X%] reduction, net saving after our cost, payback period, and three-year return. State every assumption separately so it can be challenged, and write it so the buyer could put it in front of a board.

14. The stalled deal diagnostic

A stalled deal almost always has an unnamed obstacle. This finds it and produces a plan with dates attached.

This deal has stalled: [the timeline, stakeholders, last interaction, and what was supposed to happen next]. Diagnose it honestly: is this a value problem, a priority problem, a missing stakeholder, or a deal that was never real? Give me the three most likely explanations ranked by probability, the single message or action that would test each one, and a mutual action plan with dated steps I could propose to my champion to restore momentum.

The last three cover the negotiation, and then the part almost everyone skips: working out why the outcome happened, so that it either repeats or does not.

15. The negotiation prep

Negotiations go badly when concessions get invented under pressure. Decide the order and the price of each one in advance.

I am entering final negotiation on this deal: [deal size, product, terms proposed, the buyer's stated position, and my constraints]. Prepare me: (1) what each side probably wants that has not been said out loud, (2) my concession ladder — what I can give, in what order, and what I should ask for in return each time, (3) responses to the three most likely pressure tactics, (4) the terms I should protect even at the cost of the deal, and (5) my walk-away point stated as a specific number and date.

16. The lost deal debrief

Lost deals teach more per conversation than won ones, and the reason the prospect gave you is rarely the real one.

I lost a [deal size] deal to [competitor, or no decision]. Details: [timeline, stakeholders, objections raised, the reason given, what the competitor offered]. Run a structured debrief: (1) what were the likely real reasons we lost, (2) at which stage did we lose control, (3) what should we have done differently at each stage, (4) which discovery questions would have surfaced this risk earlier, and (5) what is the best strategy for re-engaging this account in six months? Be blunt — do not protect my feelings.

17. The win analysis

A win you cannot explain is not repeatable. This extracts the pattern before you forget what actually moved the deal.

I just won a [deal size] deal. Here is the full history: [timeline, stakeholders, what they said at each stage, what I sent, who else they considered]. Work out why this actually closed: which moment was the real turning point, which of my actions mattered and which I am giving myself undue credit for, what the buyer's internal trigger was, and what I got lucky on. Then turn it into a repeatable play: the sequence, the timing, and the qualifying signal that tells me a new deal fits this pattern.

Specificity is the whole game: the buyer's role, the company's situation, what they have already said to you. A prompt carrying none of that returns copy your prospect has read a hundred times this quarter.

Feed it the raw material. Discovery notes, the prospect's own words from an email, the exact objection they raised — pasted in verbatim, these are worth far more than any summary of them you could write.

Match the model to the task. Volume work — outreach variants, sequence drafts, list qualification — runs fine on a light model such as Claude Haiku 4.5. Deal strategy, stakeholder mapping and negotiation prep want the strongest reasoning you have, because you are asking for judgement rather than words.

And check the numbers in any business case yourself. A confident ROI model with an arithmetic error in it is worse than no model at all, particularly in front of a CFO who does this for a living.

Prospecting, discovery, pitching, deal strategy and debriefs each want a different prompt, and the ones that fit your motion are worth keeping. Chat Smith lets you save any prompt here as a reusable template, group them by stage or by account, and launch one in a click — which matters when the prep window before a call is twenty minutes.

It is a multi-model assistant, so the same prompt can run across Claude, GPT, Gemini, Grok and DeepSeek from one model library and the outputs compared. For outreach that is directly useful: three cold email drafts from three models give you a wider spread of hooks to test than six from one, and the one you would never have written yourself is often the one that gets replied to.

Start with whichever prompt matches the deal in front of you — the objection rehearsal if there is a demo this week, the stalled deal diagnostic for the opportunity that has not moved in a month.