AI Prompts for Sales Reps (2026)

Each prompt is a template, and the quality of the output depends entirely on the quality of the input you give it. The pattern that works: give the model a **role** ("You are a senior B2B account executive"), the **research** (the trigger, the transcript, the prospect's words — the more specific, the better), and a tight **output format** (length, tone, structure, what to avoid). This is the standard role-context-format approach from the major prompt guides; see What Is Prompt Engineering, How to Write a System Prompt, and the OpenAI prompt-engineering guide.

The single biggest lever is specificity. "Write a cold email to a VP of Sales" produces filler; "Write a cold email referencing their May funding round and the hiring spike that usually follows it" produces something a human would read. Always paste the real artifact — the 10-K line, the LinkedIn post, the call notes — rather than asking the model to imagine it. Two efficiency tips: keep a reusable "product block" (one paragraph on what you sell and the outcome it drives) to prepend to every prompt, and read What Is a Token in AI if you're pasting long transcripts and watching cost.

Personalization at the moment a buying signal appears — the highest-leverage outreach there is.

**1. Trigger-based cold opener** — "You are a senior B2B SDR. Write a 4-sentence cold email to [first name], [title] at [company], who just [trigger event with date, e.g. announced a funding round / hired a new [role] / launched [product]]. S1: name the specific trigger and where you saw it, no flattery. S2: name the operational problem this trigger usually creates. S3: state one outcome we deliver that maps to it. S4: a soft question, not a meeting demand. Under 75 words, plain text, no 'I hope this finds you well'. What we do: [product block]."

**2. LinkedIn connection + first message** — "Write a LinkedIn connection note (under 200 characters) and a follow-up first message to [first name], [title] at [company]. Reference this specific thing they posted/did: \"[paste post or detail]\". Be peer-to-peer, curious, and non-salesy. The first message should open a conversation, not pitch. No links in the connection note."

**3. Referral / warm intro request** — "Draft a short message asking [name of mutual] to introduce me to [prospect], [title] at [company]. Make it easy for them to forward: include a one-line reason the intro makes sense and a two-sentence blurb they can paste. Respect that they're busy. Max 90 words."

**4. Re-engage a cold lead** — "Write a re-engagement email to [first name], a lead who went quiet [timeframe] ago after [what happened, e.g. a demo / a pricing conversation]. Lead with one new, genuinely relevant piece of context ([new feature / customer outcome / industry shift]), connect it to their original interest, and make a small, easy ask. No 'just checking in' or 'circling back'. Max 80 words."

Use AI to prepare sharper questions and to mine your own call recordings for what matters.

**5. Generate discovery questions** — "You are a sales coach. Generate 10 open-ended discovery questions for a first call with [title] at a [industry] company sized about [headcount/segment]. Group them by: current state, pain and impact, decision process, and success criteria. Make them conversational, not interrogating, and avoid leading questions that pitch our product. Context on what we sell: [product block]."

**6. Pre-call research brief** — "Create a one-page pre-call brief from these inputs: company overview [paste], recent news [paste], the contact's role and background [paste]. Output: likely priorities for this person's role, two hypotheses about their current pain, three smart questions to ask, and one risk to watch. Use only the facts I provided — do not invent details. Flag anything you're unsure about."

**7. Summarize a discovery call into next steps** — "Summarize this discovery-call transcript into: the prospect's top 3 pains in their own words, the decision process and timeline, budget signals, objections raised, and the agreed next step. Then draft a recap email to the prospect confirming what we heard and the next step. Transcript: \"[paste transcript]\"."

**8. Build a mini business case** — "From these call notes, draft a short value hypothesis I can share with [prospect]: the problem they described, the cost of leaving it unsolved (qualitatively or using only numbers they gave me), and the specific outcome we'd target. Do not fabricate ROI figures — if a number isn't in my notes, describe the impact qualitatively. Notes: \"[paste notes]\"."

The motions where most deals are won or lost — timely, specific, and never lazy.

**9. Follow-up that adds new context** — "Write a follow-up email to [first name] after [event, e.g. our demo on [date]]. Do not use 'just following up', 'checking in', or 'circling back'. Share one new relevant item ([resource / customer story / answer to a question they raised]), tie it to their stated priority [priority], and restate the next step with a slightly smaller ask. Under 90 words."

**10. Multi-touch nurture sequence** — "Draft a 4-email nurture sequence over [timeframe] for [first name], who is interested but not ready. Each email must stand alone, lead with value (insight, not pitch), and escalate the ask only gradually. Vary the angle: industry insight, customer outcome, a relevant resource, then a soft re-offer. Keep each under [length]. Context: [product block]; their situation: [notes]."

**11. Break-up / last-touch email** — "Write a calm, non-pushy final email to [first name] after no reply to [number] previous touches. State plainly it's the last email, name the likeliest real reason for silence (timing, not a priority, wrong contact), give a graceful low-friction option if they ARE the right person, and close warmly. No guilt trips. Max 80 words."

Rehearse responses before the call and tighten replies after it.

**12. Handle a specific objection** — "A prospect said: \"[paste exact objection]\". Give me three ways to respond: (1) acknowledge-and-reframe, (2) ask-a-question-to-uncover-the-real-concern, (3) provide-evidence (using only facts I supply: [facts]). For each, one line on when to use it. Keep responses conversational, not scripted. Do not invent statistics or customer names."

**13. Prep for a pricing conversation** — "Help me prepare for a pricing discussion with [prospect]. Based on their stated priorities [paste], draft: how I'd anchor on value before price, two likely pushbacks and calm responses, and one question to understand their budget reality. Do not state or invent any actual prices — leave [our pricing] as a placeholder."

**14. Competitor displacement talk track** — "A prospect currently uses [competitor]. Without disparaging them, draft a talk track that focuses on the specific gap [gap] and the outcome we'd improve. Be factual and confident, not negative. If I haven't given you a verified fact about the competitor, do not assert it — frame as a question to confirm instead."

Do not let AI invent facts to make outreach sound impressive. It will fabricate statistics, ROI numbers, customer names, case studies, competitor claims, and prices if you let it — and a buyer who catches one invented detail stops trusting everything you say. Feed the model only verified facts, and instruct it (as the prompts above do) to describe impact qualitatively when a number isn't available.

Protect data and relationships. Don't paste confidential customer information, signed-contract terms, or another company's private data into a public chatbot; redact to brackets, and use an enterprise instance with the right data agreement for anything sensitive — check your company's policy first. And don't send raw AI output. Generic, over-polished AI cadence is easy to spot and reads as automation, so cut any sentence without a specific noun or number, use contractions, and read it aloud before sending. For sequencing and reply-rate craft beyond these templates, see the Cost Per Token, All Major Models (2026) guide for budgeting long-transcript prompts, and the OWASP LLM Top 10 for handling untrusted pasted content.

Any current flagship drafts outreach well; the practical choice is about speed and cost for high-volume drafting versus careful reasoning over long transcripts and account research. The table compares durable dimensions only — verify live pricing on each provider's page.