Sales reps spend more time on admin than selling. Logging CRM notes, drafting follow-ups, rewriting that passive-aggressive prospect email so it sounds less desperate. AI sales prompts can handle most of that grunt work, and they can do it fast.

But here's the thing about sales: the words are the easy part. Knowing when to send them, what the prospect actually cares about, when to push and when to shut up. That's the job. AI can draft. You decide.

This article gives you 10 copy-paste prompts for the real work of sales: discovery summaries, follow-ups, objection prep, CRM updates, and more. Use them as starting points, not finished output.

Before you paste anything: don't include real customer names, confidential pricing, contract terms, security details, or regulated data in any AI tool unless your company's policy explicitly permits it. Anonymize first. Always.

The reusable formula behind every good AI sales prompt

Bad prompts produce bad drafts. The fix is simple: give the AI a role, context, task, constraints, and a format. This is what Dee calls the "smart intern" framework in Don't Replace Me. Treat AI like someone who's fast, tireless, knows a lot in general, and knows nothing specific about your customer or your deal.

Here's the formula:

Role: "You are a B2B sales rep." Context: What happened, who's involved, what stage you're at. Task: What you need the AI to produce. Constraints: What to avoid, what tone to use, what's off-limits. Format: Email, bullet list, short paragraph, whatever you need.

Every prompt below follows this structure. You'll see it immediately.

The most common mistake people make is pasting in a one-liner and expecting usable output. "Write me a follow-up email for a sales call" gets you something generic enough to be useless. Specificity is everything. The more context you give, the closer the first draft lands. You're not saving time by being vague. You're just moving the work from drafting to editing.

10 AI sales prompts you can use today

Prompt 1: Summarize a discovery call

When to use it: Right after a call, before your memory goes soft.

You are a B2B sales rep. I just finished a 30-minute discovery call with a [job title] at a [industry] company with about [company size] employees. Here's what I remember: [paste your rough notes]. Summarize this into: (1) key pain points mentioned, (2) business goals they described, (3) objections or hesitations, (4) questions they asked, (5) agreed next steps. Use bullet points. Keep it under 200 words.

What to check: Did you miss anything? Is anything in the summary a stretch of what was actually said? Don't let the AI invent clarity that wasn't in the call.

Prompt 2: Extract buying criteria and decision factors

When to use it: When you want to build a smarter pitch for the next meeting.

You are a B2B sales strategist. Based on these discovery call notes: [paste anonymized notes], list the apparent buying criteria this prospect seems to care about, ranked by how strongly they emphasized each one. Then flag anything that suggests a potential blocker or political complexity. Be direct.

What to check: These are inferences from your notes, not facts. Treat the output as a hypothesis to test in the next conversation.

Prompt 3: Draft a useful follow-up email

When to use it: After any call, meeting, or demo. The most-used prompt in this list.

You are a B2B sales rep. Write a follow-up email to [role, not name] after a [type of meeting]. We discussed: [3-5 bullet points of key topics]. The next step we agreed on is [specific action and date]. The email should be under 150 words, warm but not sycophantic, and end with one clear ask. No filler sentences.

What to check: Does the "one clear ask" match what you actually agreed? Is the tone right for this specific relationship? If you've only spoken once, "warm but not sycophantic" means something different than if you've been working together for months. For more on writing better work emails, see the AI email prompts guide.

Prompt 4: Rewrite a pushy email so it sounds human

When to use it: You wrote something under pressure and it reads like a ransom note.

You are an editor specializing in sales communication. Here is an email draft I wrote: [paste draft]. Rewrite it so it sounds like a real person talking, not a sales sequence. Remove urgency language, soften anything that sounds like pressure, and cut any sentence that exists purely to fill space. Keep it under 120 words.

What to check: Is the core message still there? Did the rewrite remove something important to the deal? Read it out loud. If you wouldn't say it in a call, don't send it in an email.

Prompt 5: Prepare for common objections

When to use it: Before a pitch meeting or negotiation call.

You are a sales coach. My product is [one sentence description]. My prospect is a [role] at a [type of company]. Based on what I know about this buyer type, what are the five most likely objections they'll raise? For each objection, give me: (1) a likely underlying concern, and (2) a response that addresses the concern without dismissing it. Keep each response under 50 words.

What to check: Do these responses match what your product actually does? Don't let AI write claims you can't stand behind. If the objection involves pricing, contract flexibility, or SLA commitments, verify with your manager before using any response.

Prompt 6: Qualify an opportunity

When to use it: When you're not sure if an opportunity is worth pursuing.

You are a B2B sales strategist. Here is what I know about this opportunity: [paste your notes, anonymized]. Using the MEDDIC framework (Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion), flag what I know, what I'm missing, and what questions I should ask to fill the gaps. Format as a table.

What to check: Is the MEDDIC read accurate, or are you being optimistic? The AI doesn't know your gut feel on this deal. You do. Use this as a sanity check, not a forecast.

Prompt 7: Create a proposal outline

When to use it: Before you start writing a formal proposal or presentation.

You are a B2B sales rep preparing a proposal for a [type of company] with [describe their situation in 2-3 sentences]. The key outcomes they care about are: [list]. Create a proposal outline with: (1) executive summary framing, (2) section headers, (3) a suggested order that builds toward the ROI argument. No filler sections.

What to check: Does the outline reflect what this prospect actually cares about, or is it generic? Personalize before you use it. Numbers, timelines, implementation specifics, pricing, and legal terms must all be verified before they appear in any real proposal.

Prompt 8: Write a CRM update

When to use it: End of day, before you forget everything.

You are a B2B sales rep. Write a CRM activity note for a call I had today. Here's what happened: [paste rough notes]. Format it as: date, contact role, meeting type, key topics discussed, next action, and owner. Factual and concise. Under 100 words.

What to check: Is everything accurate? CRM notes have a way of becoming gospel for colleagues who weren't on the call. Don't let a wrong inference from AI create a misunderstanding down the line. For handling meeting notes more broadly, the AI meeting notes prompts article has more templates.

Prompt 9: Prepare a handoff to customer success

When to use it: When a deal closes and you need to brief the CS or onboarding team.

You are a B2B sales rep writing a handoff document for customer success. The customer is a [type of company, anonymized]. Here's what I know about them: [paste notes]. Write a handoff document with: (1) context on why they bought, (2) what success looks like to them, (3) known risks or sensitivities, (4) relationships to be aware of, (5) agreed commitments made during the sales process. Be specific and honest about risks.

What to check: Is the commitments section accurate? If you promised something outside standard terms, that needs to be in there, clearly marked. Don't let the AI smooth over the awkward bits.

Prompt 10: Plan a no-pressure reactivation message

When to use it: A deal went quiet 3-6 months ago and you want to try again without being annoying about it.

You are a B2B sales rep. I had a deal with a [role] at a [type of company] that went quiet [X months] ago. The last thing that happened was: [one sentence]. Write a short reactivation message (under 100 words) that: acknowledges the gap without apologizing excessively, offers something genuinely useful (a new insight, relevant change, or concrete question), and opens the door without pressuring them to respond. Human, not automated-sounding.

What to check: Is the "genuinely useful" thing actually useful, or did AI make something up? If you're referencing a market development, company news, or product update, make sure it's real before you send it.

How to adapt these AI sales prompts to your specific context

The prompts above are templates. They work as written, but they work better when you tune them to your actual situation. Here's how to get more out of them.

Adjust the role instruction. "You are a B2B sales rep" is a fine default, but if you sell enterprise software with 18-month sales cycles, say that. If you're in a high-volume transactional role, say that too. The more the AI understands your context, the less generic the output.

Add tone anchors. If you have a particular way you communicate with a client, tell the AI. "Match the tone of this previous email I sent: [paste]" is more effective than asking for "professional but friendly." Show it, don't describe it.

Use your own examples as input. Past emails that landed well, CRM notes from a deal you won, a proposal that got a signature. Paste those as reference material. Tell the AI to match the structure or tone. You're training it on what good looks like for you specifically, not what good looks like in general.

Chain prompts for complex tasks. For something like a full proposal, don't try to get it all in one prompt. Run the discovery summary first, then the buying criteria extraction, then the proposal outline, then each section separately. Each step gives the next one better context.

The difference between AI saving you 20 minutes and AI actually making your output better usually comes down to how much context you're willing to put in upfront.

This came from a book.

Don't Replace Me

200+ pages. 24 chapters. The honest version of what AI means for your career, written by someone who actually builds this stuff.

Get the Book →

What AI still can't do in sales

AI can organize your notes, draft your emails, and help you rehearse objections. It can't read the room. It doesn't know that your champion is fighting a political battle internally, or that the prospect went quiet because their budget got cut, not because they lost interest. It can't tell when the right move is to stop emailing and pick up the phone.

There's a reason the best salespeople in any organization tend to be the ones with the deepest relationships and the sharpest instincts about timing. A buyer does not trust you because your CRM note is well-formatted. They trust you because you listened, remembered the real problem, and did not overpromise when the room got uncomfortable. AI can help you prepare for that. It cannot do that part for you.

Trust, timing, and judgment are still yours to own. That's not a motivational poster point. It's a practical one. If you want to understand which parts of your job are genuinely hard for AI to take, the human skills AI can't replace breaks it down without the usual cheerleading.

And on the question of whether to use AI at work openly or quietly, that's worth thinking through. The should you tell your boss you use AI article has the honest answer.

A note on what not to paste into AI tools

This deserves its own section because people skip it.

Do not paste into any AI tool, unless your company's policy explicitly says you can:

Anonymize everything. Replace names with "the prospect" or "the buyer." Replace company names with industry descriptors. Replace specific numbers with ranges or placeholders.

If you're using a consumer AI tool like ChatGPT or Claude without a business agreement in place, assume your input could be used for training. Act accordingly.

The Microsoft Work Trend Index found that 78% of AI users at work bring their own tools rather than using company-sanctioned ones. That number matters because it means most people are operating without any data policy guardrails. If that's you, be careful about what you put in.

Frequently asked questions

Can AI write my sales emails for me?

AI can draft sales emails, but you need to edit them before sending. Check every claim, verify pricing and timelines, and make sure the tone matches your relationship with that specific person. A generic AI draft sent without editing is often worse than no email at all.

Is it safe to paste customer information into ChatGPT?

Not without anonymizing it first. Don't include real names, company names, deal sizes, contract terms, or regulated data unless your company has a data processing agreement with the AI provider. When in doubt, replace specifics with placeholders before pasting anything.

What's the MEDDIC framework and why does it matter for sales prompts?

MEDDIC stands for Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion. It's a qualification framework common in B2B sales. Using it in prompts gives the AI a structured way to help you spot gaps in what you know about an opportunity, rather than just generating generic advice.

Can AI help me handle objections in a live sales call?

Not live, no. AI is useful for preparing objection responses before a call. During the call itself, you're on your own. Use AI to rehearse and prepare, then put the phone down and listen to the actual person in front of you.

How do I use these prompts without sounding like a robot?

Read the draft out loud before you send anything. If you wouldn't say it in a conversation, rewrite it. The prompts include instructions to avoid filler and sound human, but AI still has a tendency toward slightly formal phrasing. Your job is to catch that and fix it.

Will using AI for sales make me better or just faster?

Faster at drafting, organizing notes, and preparing for calls. Whether that makes you better depends on how you use the time you save. If you spend it thinking more carefully about the deal and the relationship, yes. If you just send more volume with less thought, probably not.

Do I need a special AI tool for sales, or will ChatGPT work?

ChatGPT or Claude will handle everything in this list. You don't need a sales-specific AI tool. Most of the "AI for sales" products on the market are wrappers around the same underlying models, often with a CRM integration bolted on. Start with what you already have access to, and only pay for a specialized tool if you hit a genuine limitation.