AI Sales Prompts: 30 Ready-to-Use Prompts for Reps

AI sales prompts are written instructions reps feed into tools like ChatGPT or Claude to speed up research, outreach, call prep, and follow-up — a good one specifies the role, the context, the exact output format, and a constraint on length or tone, rather than a vague one-line request. Below are more than 30 you can copy today, organized by where they fit in the sales cycle, plus an honest look at where hand-typed prompting stops scaling and something more automated takes over.

What Makes an AI Sales Prompt Actually Work

A good AI sales prompt gives the model four things: a role, context, a specific task, and a constraint on the output. Skip any one of those and the model fills the gap with something generic, or worse, invented. This matters more in sales than in most other use cases, because a hallucinated statistic in a cold email or a made-up feature comparison in a proposal doesn’t just look bad — it can kill a deal the moment a prospect fact-checks it.

Specificity is the single biggest lever. Asking an AI tool to “write a cold email” produces filler. Asking it to write a three-sentence cold email to a VP of Revenue Operations at a 200-person SaaS company, referencing a recent Series B raise, with a single question as the call to action, produces something usable. The extra sixty seconds of typing saves ten minutes of rewriting.

Role prompting matters almost as much. Telling the model “you are an enterprise AE who has run this exact discovery call two hundred times” changes the register of the output noticeably compared to no role at all — it reaches for the vocabulary and structure an experienced rep would actually use instead of a textbook version of selling. Context is the third piece: pasting in the prospect’s LinkedIn summary, a recent press mention, or your own product’s actual differentiators (not what you assume the model knows) grounds the answer in facts instead of training-data guesses. And a stated constraint — word count, tone, what to leave out — is what keeps the output usable without a rewrite.

One more habit worth building: explicitly give the model permission to say it doesn’t have enough information rather than filling gaps with plausible-sounding fiction. A prompt that ends with “if you don’t have enough detail on this company, say so rather than guessing” measurably cuts down on invented facts, which matters a lot when the output is going straight into an email a real prospect will read.

Account Research and Prospecting Prompts

These prompts work best when you paste in real source material — a LinkedIn profile, a press release, an earnings call excerpt, a job posting — rather than asking the model to research a company from memory, since it has no live access to current company data on its own.

• Company snapshot: “You are a B2B sales rep prepping for a first call. Based on the following company overview and recent news [paste source], summarize in 5 bullet points: what this company does, who their buyer likely is, one recent trigger event, and one plausible reason they’d care about [your category]. Flag anything you’re inferring rather than confirming.”
• Persona pain-point mapping: “Given this job title and this LinkedIn About section [paste], list the three most likely priorities this person is measured on this quarter, and for each one, one question I could ask on a call to surface it without sounding scripted.”
• Trigger-event framing: “This company just [funding round / new exec hire / product launch — paste detail]. Write two sentences connecting that event to a reason they might now have budget or urgency for [your category], written the way an experienced rep would explain it to a colleague, not as marketing copy.”
• Competitive landscape scan: “Based on this list of tools mentioned in their job posting / tech stack page [paste], what does their current stack suggest about how mature their [function] process already is, and where’s the most obvious gap?”
• Account prioritization pass: “Here are ten accounts with these firmographic details [paste table]. Rank them by likely fit for [your ICP criteria] and explain your reasoning for the top three in one sentence each.”

For a deeper walkthrough of turning this kind of research into a repeatable habit rather than a one-off exercise, see our guide on how to use AI for account research.

Cold Outreach and Personalization Prompts

Outreach prompts should optimize for one specific, checkable detail rather than generic flattery. Research on trigger-based messaging — emails that reference a funding round, a hire, or a leadership change — has shown open rates around 57%, reply rates near 9.6%, and positive reply rates close to 5.8%, well above what generic templated outreach typically pulls. The prompts below are built around that same principle: one real hook, not five vague ones.

• First-line personalization: “Write one opening sentence for a cold email to [name, title] that references [specific trigger event] without using the words ‘noticed,’ ‘saw,’ or ‘congrats.’ Keep it under 20 words.”
• Full cold email draft: “Write a 3-sentence cold email to [persona] at [company]. Sentence 1: the trigger-based hook. Sentence 2: a one-line relevant insight, not a pitch. Sentence 3: a low-friction question, not ‘do you have 15 minutes.’ No exclamation points, no buzzwords like ‘synergy’ or ‘revolutionize.'”
• LinkedIn connection message: “Write a LinkedIn connection request under 300 characters referencing [shared context — mutual connection, recent post, shared group]. No pitch, no CTA — just a reason to connect.”
• Break-up email: “Write a final follow-up email for a prospect who’s gone quiet after 3 touches. Tone: direct, no guilt-tripping, gives them a graceful out, and leaves the door open. Under 60 words.”
• Multi-touch sequence outline: “Outline a 5-touch outbound sequence (email, LinkedIn, email, call, email) over 12 days for [persona] at [company type]. For each touch, give the channel, the day, and a one-line angle — not full copy yet.”
• Subject line variations: “Give me 8 subject lines for this email [paste draft], under 6 words each, no title case, no emoji, mixing curiosity and directness.”

Personalized, AI-assisted outreach isn’t just a nice-to-do — teams that deploy it well have reported open rates rising by roughly 42%, meeting booking rates up around 31%, and proposal acceptance improving close to 27% compared with generic sequences. The gap between a prompt-written email and a template is usually that one specific, verifiable detail.

Discovery Call and Meeting Prep Prompts

Call prep is where role prompting earns its keep — telling the model what kind of call this is and what you already know changes the quality of the questions it suggests.

• Discovery question bank: “You are a senior AE running a first discovery call with [persona] at a [company size] [industry] company. Give me 8 open-ended discovery questions organized under three headings: current process, cost of inaction, and decision process. No yes/no questions.”
• Call agenda: “Draft a 30-minute discovery call agenda for a first meeting with a [persona], with time allocations, that leaves the last 5 minutes for next steps rather than a pitch.”
• Pre-call briefing: “Summarize this account’s research notes [paste] into a 6-line pre-call brief: who I’m meeting, their likely priority, one risk to watch for, one good opening question, and what a strong next step looks like.”
• Mock objection roleplay: “Act as a skeptical VP of Sales evaluating [your category]. I’m going to pitch you in a few sentences — respond the way a genuinely busy, slightly skeptical buyer would, including at least one real objection, not a softball.”
• Post-call summary: “Turn these raw call notes [paste] into a structured summary: pain points mentioned, stated timeline, budget signals if any, stakeholders named, and agreed next step. Flag anything ambiguous rather than assuming.”

Objection Handling and Negotiation Prompts

These are best used for rehearsal and structure, not as a script to read verbatim on a live call — buyers can tell.

• Objection reframe: “A prospect just said: ‘[paste objection verbatim].’ Give me three different response angles — one that asks a clarifying question back, one that reframes with a short example, and one that acknowledges and pivots. Keep each under 3 sentences.”
• Pricing pushback: “Prospect says our price is higher than [competitor]. Help me draft a response that doesn’t discount immediately, focuses on total cost of the status quo, and ends with a question rather than a statement.”
• Competitive displacement talking points: “This prospect currently uses [competitor]. Based on these known differences [paste feature comparison], write 3 fair, non-disparaging talking points for why a switch might be worth evaluating, plus one honest scenario where staying with their current tool might make sense.”
• Stakeholder-specific framing: “I need to explain the same value proposition to a CFO and to a VP of Sales. Give me two versions of the same core message — one framed around cost and risk, one framed around rep productivity and pipeline.”
• Late-stage risk reversal: “Deal is stalled at legal review. Draft a short, low-pressure check-in message that offers help removing a blocker rather than asking ‘any update?'”

Follow-Up and Pipeline Management Prompts

• Meeting recap email: “Turn these call notes [paste] into a recap email to the prospect: 3 bullet points on what we discussed, 1 bullet on next steps with an owner and date, professional but not stiff tone.”
• Multi-threading intro request: “Draft a short message I can send my champion asking them to introduce me to [stakeholder role], explaining why briefly and making it easy for them to forward.”
• Deal risk summary: “Based on this deal history [paste stage changes, last activity date, stakeholder engagement], write a 3-line risk assessment I can share with my manager: what’s going well, what’s a yellow flag, what I’m doing about it.”
• Quarterly pipeline narrative: “Summarize this list of open opportunities [paste CRM export] into a short narrative for a forecast call: total pipeline, biggest deal at risk, and one deal that could pull forward.”

For a broader look at automating this stage of the cycle rather than prompting it one email at a time, see our guide on how to automate sales follow-ups.

Why Manual Prompting Hits a Ceiling at Scale

Every prompt above works. The problem isn’t quality — it’s throughput. The typical workflow looks like this: find a lead, paste their LinkedIn profile into ChatGPT, prompt for a cold email, edit out the parts that read generic, copy it into your sequencing tool, and repeat for the next lead. It works for the first ten prospects of the day. It falls apart at the volume a real pipeline target requires, because every step still needs a human copying, pasting, and reviewing.

There’s also a data problem that no prompt can fully solve. ChatGPT and Claude are extraordinary at language, but neither has live access to your CRM, your intent data, or a prospect’s current buying signals unless you paste that information in yourself, every single time. That’s fine for one account. It’s not a workable system for two hundred. This is the exact gap that purpose-built AI sales tools are built to close — not by writing better sentences than a well-crafted prompt, but by removing the copy-paste step entirely and connecting the writing to live account and CRM data automatically.

AI Prompting vs. an AI Sales Copilot or AI SDR

Whether hand-typed prompts are enough or whether you need a dedicated platform depends almost entirely on volume and how much manual assembly you’re willing to do per prospect. The table below breaks down where each approach actually holds up.

In practice, most sales orgs end up using both — prompts for the ad hoc, judgment-heavy moments (a tricky objection, a one-off executive email) and an automated layer for everything that needs to happen at volume and on a schedule. If you’re trying to figure out where that line sits for your own team, that’s exactly the kind of side-by-side you can see directly by booking a demo rather than guessing from a features list — seeing how a platform like SalesWorx’s AI sales copilot handles the same account research and outreach tasks these prompts do, but pulling from live data instead of a pasted LinkedIn summary, usually makes the tradeoff obvious fairly quickly.

It’s worth being fair here: SalesWorx and platforms like it aren’t magic, and they won’t out-write a genuinely skilled rep on a nuanced, relationship-driven email. What they solve is consistency and coverage — the same quality of research and personalization applied to account 200 as to account 1, without a human re-typing prompts each time. For a fuller comparison of what a copilot handles versus a fully autonomous AI SDR, our AI Sales Copilot vs AI SDR breakdown goes into the distinction in more depth.

Common Mistakes Reps Make With AI Sales Prompts

Most bad AI output in sales traces back to one of a handful of habits, and every one of them is fixable without a new tool.

• Asking too broadly. “Write me a sales email” gives the model nothing to work with. Specify the persona, the trigger, the length, and the tone.
• Skipping the fact-check step. Never send an AI-drafted email or proposal that references a statistic, feature, or company detail you haven’t personally verified. Models can and do fabricate specifics that sound completely plausible.
• Reusing the exact same prompt for every account. If your prompt template doesn’t change per prospect, the output won’t either — and buyers can spot a templated email even when the words are different.
• Not giving the model permission to say “not enough information.” Without that instruction, the model will often guess rather than flag a gap, and a confident-sounding guess is more dangerous than an obvious one.
• Treating prompt output as final copy. The best use of these prompts is a strong first draft in 90% less time — not a finished send. A 30-second human read-through catches the last 10% every time.
• Ignoring where volume breaks the workflow. A prompt library is a skill, not a scaling strategy. Know the point where your pipeline goals outpace what one person copying and pasting can sustain.

Frequently Asked Questions

What is the best AI prompt for sales?

There isn’t one universal best prompt — the strongest ones are specific to the task, the persona, and the stage of the deal. The consistent pattern across effective sales prompts is a defined role, real pasted context (not assumed facts), a precise task, and an explicit constraint on length, tone, or format.

Can ChatGPT replace an SDR?

No. ChatGPT can draft strong individual pieces of content — an email, a call script, an objection response — but it can’t autonomously research accounts with live data, sequence multi-channel outreach, or manage a pipeline on its own. That’s the functional gap that AI SDR and AI sales copilot platforms are built to close.

Are AI-written sales emails less effective than human-written ones?

Not inherently — the deciding factor is specificity and personalization, not authorship. Generic AI output performs about as poorly as generic human templates. Well-prompted, trigger-based, specific messaging has shown meaningfully higher open and reply rates than generic outreach, whether a human or an AI tool drafted the first version.

How much does it cost to use AI prompts for sales?

Using ChatGPT or Claude directly for prompting typically costs whatever your existing subscription already covers — often $20 or so per seat per month for a standard plan, sometimes free at lower usage. The real cost is time: every prompt still requires a human to gather context, run the prompt, and edit the output, which is the expense that scales with pipeline volume even when the software itself doesn’t.

Is prompt engineering a useful skill for sales reps?

Yes. Reps who get specific and consistent with how they prompt — providing role, context, and constraints every time — get measurably more usable output with less editing, and that skill transfers across whatever AI tools their stack includes.

When should a sales team move from manual prompting to an AI sales platform?

Generally once account or lead volume outpaces what one person can research and personalize by hand, or once consistency across a team of reps becomes the priority instead of individual skill with a chat window. Teams already running AI outbound sales at scale tend to hit that point earlier than they expect.

Do AI prompts work the same in ChatGPT, Claude, and Gemini?

The underlying principles — specificity, role, context, constraints — carry across all of them, though each model has its own quirks in how literally it follows formatting instructions or how much it pushes back on ambiguous requests. A prompt that works well in one usually needs only minor tweaking to work as well in another.