What makes a good sales prospecting prompt? A high-quality sales prompt specifies a role for the AI, provides ICP context, sets constraints (tone, length, banned phrases), and defines the exact output format needed. Without all four, output quality is unpredictable.
Most ChatGPT prompts for sales are too vague to produce usable output. "Write a cold email to a VP of Sales" gives you a generic, unusable template that reads like every other AI-generated message in your prospect's inbox.
The prompts that actually work follow a formula: they give the model a specific role, a constrained context, exact output requirements, and enough detail about your ICP that it can generate something a human would send. Woodpecker's 2026 cold email data shows that advanced personalization achieves reply rates of 17 to 18%, roughly double the 7 to 9% average for basic outreach (checked September 2026). The difference starts at the prompt level.
Below are six prompt templates you can use today. Each has a context header, the prompt itself with [BRACKETED] variables you replace, and a breakdown of why it produces better output than the vague alternatives.
What makes a good sales prospecting prompt? A high-quality sales prompt specifies a role for the AI, provides ICP context, sets constraints (tone, length, banned phrases), and defines the exact output format needed. Without all four, output quality is unpredictable.
Generic prompts produce generic output because the model has no anchor. When you don't specify who you are, what you sell, who you're selling to, or what the output should look like, the model defaults to averaging across millions of examples. That average is mediocre by definition.
A ChatGPT prompt that impresses in a demo rarely survives contact with your actual target market. The gap is usually missing context, not missing model capability.
The fix is the Prompt Precision Formula.
Every sales prompt that produces usable output contains four elements:
Context: Who you are, what you sell, and who you're targeting. The more specific, the better. "B2B SaaS for logistics" is better than "software company." "50 to 500-person manufacturing companies that use SAP" is better still.
Role: Assign the model an expert persona. "You are a senior B2B copywriter specializing in outbound for SaaS companies" gives the model a frame that raises output quality significantly over no role assignment.
Constraints: Explicit limits on what not to do. "Do not use phrases like 'I hope this finds you well' or 'just following up.' Keep the email under 80 words. Use a conversational tone."
Output format: Tell the model exactly what to produce. "Write one subject line and one email body" beats "write an email." "Produce three variations" gets you options to A/B test.
See the full breakdown in our AI in B2B sales prospecting guide and in our AI-generated icebreakers resource for the personalization layer specifically.
Context: Use this before writing any outreach. It generates a detailed buyer persona you can feed into subsequent prompts.
You are a B2B market research analyst specializing in [INDUSTRY] companies.
I sell [PRODUCT/SERVICE] to [TARGET TITLE] at [COMPANY SIZE]-person [INDUSTRY] companies in [REGION].
Analyze this ICP and produce:
1. The top 3 business pain points this persona faces related to [PROBLEM AREA]
2. The internal and external pressures driving urgency to solve those pains
3. The metrics this persona is typically measured on
4. The language they use when describing this problem (not the language a vendor would use)
5. The objections they commonly raise when evaluating solutions like mine
Keep each point to 2 to 3 sentences. Use direct language, not marketing speak.
Why this works: Most cold email fails because it addresses what the seller thinks is important, not what the buyer actually cares about. This prompt forces the model to generate buyer-centric language, which you then use in every downstream prompt and email. Highly personalized emails that reflect genuine buyer language produce reply rates in the 17 to 18% range according to Woodpecker's 2026 data versus 7 to 9% for generic approaches (checked September 2026).
Context: Use when you have a specific target account and want quick background before writing outreach.
You are a senior account executive at a B2B company. I'm about to reach out cold to [COMPANY NAME], a [COMPANY SIZE]-person [INDUSTRY] company.
Using publicly available information, produce a concise account research brief covering:
1. What they do and who their customers are (2 sentences)
2. Any recent news, funding events, product launches, or leadership changes in the past 6 months
3. Likely pain points related to [YOUR PRODUCT CATEGORY] based on their business model
4. One specific, relevant hook I could open a cold email with that references something real about this company
Do not fabricate news or events. If you are uncertain about a detail, say so explicitly. Keep the entire brief under 200 words.
Why this works: Account-specific research turns a cold outreach into a warm one. The instruction to flag uncertainty prevents the model from hallucinating events that would embarrass you. The hook request means you get a ready-to-use opening line for your email. For more on account-level research prompts, see our ChatGPT for account research guide.
Context: Use after running Prompts 1 and 2. Paste the ICP pain points and account hook into this prompt.
You are a senior B2B copywriter specializing in outbound email for [INDUSTRY] companies.
Company I'm emailing: [COMPANY NAME], a [COMPANY SIZE]-person [INDUSTRY] company.
My product: [PRODUCT NAME] - [ONE SENTENCE DESCRIPTION]
Buyer persona: [TARGET TITLE]
Key pain point to address: [PAIN POINT FROM PROMPT 1]
Hook to use: [HOOK FROM PROMPT 2]
Write a cold email first touch that:
- Opens with the hook above (do not use "I hope this finds you well" or any variation)
- Connects the hook to the pain point in one sentence
- States what we do in one sentence, no more
- Ends with a single low-friction CTA (not "let's hop on a 30-minute call" - suggest something smaller)
- Subject line: 4 to 6 words, curiosity-driven, no question marks
- Total length: 60 to 80 words including subject line
Do not use phrases like "revolutionize," "game-changer," "seamless," or "leverage."
Why this works: The constraints do the heavy lifting. Specifying "60 to 80 words" forces brevity. Banning specific phrases eliminates the most overused AI-generated language. The explicit CTA guidance prevents the model from defaulting to "schedule a demo," which is the highest-friction ask possible. Compare against our B2B cold email templates to see what high-performing firsts-touch copy looks like in practice.
Context: Use for steps 2 through 5 of a sequence. Reference the previous touch so the follow-up adds value rather than repeating.
You are a B2B sales copywriter. I sent the email below to [FIRST NAME] at [COMPANY NAME] [X DAYS] ago. They have not replied.
Previous email:
[PASTE PREVIOUS EMAIL]
Write follow-up email #[NUMBER IN SEQUENCE] that:
- Does not say "just following up" or "I wanted to circle back"
- Adds a new angle, insight, or piece of value not mentioned in the previous email
- Stays under 60 words
- Ends with the same low-friction CTA or a softer one
- Subject line: reply thread ("Re: [ORIGINAL SUBJECT]") or a new curiosity hook under 5 words
New angle to introduce: [NEW DATA POINT, CUSTOMER OUTCOME TYPE, OR RELEVANT INSIGHT]
Why this works: Single follow-ups increase replies by 65.8% according to Woodpecker's data, and the gains continue through follow-up three before diminishing (checked September 2026). But follow-ups that just say "checking in" are noise. Injecting a new angle each time keeps the sequence from feeling like a drip campaign. See our email sequence best practices for the full cadence structure.
Context: Use when a prospect replies with a common objection. Replace the variables with the actual reply.
You are a consultative B2B sales rep. A prospect replied to my cold email with this objection:
"[PASTE ACTUAL OBJECTION]"
My product: [PRODUCT NAME] - [ONE SENTENCE DESCRIPTION]
The prospect: [TITLE] at [COMPANY NAME], a [COMPANY SIZE]-person [INDUSTRY] company.
Write a reply that:
- Acknowledges the objection directly without dismissing it
- Reframes without arguing or overselling
- Asks one clarifying question that moves the conversation forward
- Total length: 50 to 70 words
- Do not offer a discount, free trial, or extended access unless I specifically ask for that
Tone: direct and respectful, not defensive.
Why this works: Most objection replies go wrong in two ways: they dismiss the concern or they immediately pile on more selling. This prompt forces acknowledgment first, reframing second, and a question third, which is the conversational structure that keeps deals alive. The constraint against discounts prevents the model from defaulting to a concession as a way out.
Context: Use when you're expanding into a new industry and don't have established messaging yet.
You are a B2B sales strategist with deep expertise in [NEW TARGET INDUSTRY].
I sell [PRODUCT/SERVICE] to B2B companies and I'm expanding into [NEW TARGET INDUSTRY].
For companies in this industry:
1. List the top 5 operational or revenue challenges related to [YOUR PRODUCT CATEGORY]
2. For each challenge, describe: who owns it internally (title), how they typically try to solve it today, and why current solutions fall short
3. Identify the trigger events (funding, hiring surge, regulation change, etc.) that signal urgency to solve this problem
4. Write 3 cold email hooks I could use that reference one of these trigger events and the pain it creates
Keep each answer factual and grounded. Do not describe outcomes I should claim - describe the problems they actually face.
Why this works: Expanding into a new vertical blind is one of the fastest ways to waste outbound budget. This prompt structures your discovery so you're building from real pain patterns, not assumptions. The trigger event section feeds directly into your intent signal strategy. Pair this with our AI lead generation tools guide and outbound automation guide to turn the research into a running sequence.
ChatGPT can write the email. It can research the account. What it can't do is verify that the contact you're emailing is real, still at the company, or reachable at an address that won't bounce.
InboundLabs provides a database of 280M verified B2B contacts with 98% email deliverability on verified contacts. Filter by industry, headcount, region, and title to build the contact list your prompts need to run against. Buyer intent signals layered on firmographic data mean you can also feed the "trigger event" section of Prompt 6 with real-time signal data rather than guesses.
Free to start, no credit card required. Monthly plans, no annual lock-in.
See how InboundLabs finds verified contacts instantly → inboundlabs.app
A precise prompt built on all four formula components produces consistent, usable output every time. A vague prompt produces the same mediocre email your prospects have already seen. Start with the ICP Profile Builder (Prompt 1) before any other prompt. Every downstream output gets better when the model knows exactly who you're targeting.
Pair these prompts with high-quality verified contact data and a structured outreach sequence. For the cold email templates that work alongside AI-drafted copy, see our cold email personalization examples and our guide on how to write a cold email that converts.
Can ChatGPT write cold emails for B2B sales prospecting? Yes, with the right prompts. Generic prompts produce generic output. Prompts that specify role, context, constraints, and output format produce emails close enough to send with light editing. The Prompt Precision Formula structures each component.
Do AI-written cold emails work? AI-written emails perform comparably to human-written emails when personalization is deep and contacts are verified. Woodpecker's 2026 data shows advanced personalization achieves reply rates of 17 to 18% versus 7 to 9% for basic approaches. The quality of the contact list and the depth of personalization matter more than who wrote the copy.
How many prompts does it take to write a full email sequence? Three to four prompts covers most sequences: ICP research, account research, first touch, and follow-up variations. Running Prompt 1 once per ICP and Prompts 2 and 3 per target account is a realistic workflow for a solo SDR or a small team.
Should I edit ChatGPT output before sending? Always. AI output is a first draft, not a final draft. Check for accuracy, remove any fabricated details, adjust the tone to match your voice, and verify that the opening line references something real and recent about the prospect.
What are the limits of ChatGPT for sales prospecting? ChatGPT cannot access live data by default, which means account research outputs can be stale. It also cannot verify contact accuracy, check email deliverability, or manage your sequence sending. It's a writing and research assistant, not a prospecting platform.
LSI keywords: chatgpt sales prompts, AI prospecting prompts, cold email prompt templates, GPT sales copywriting, AI cold email templates, ChatGPT B2B outreach, sales prompt engineering, AI email writing, prospect research prompts, outbound sales AI, ChatGPT outreach, email personalization AI
What is an AI SDR? An AI SDR is software that automates top-of-funnel sales tasks including contact research, account targeting, email writing, and outreach sequencing. It operates without human input on each individual step, scaling prospecting efforts across thousands of contacts at once.
What is an AI SDR? An AI SDR is a software agent that automates the research, outreach, and follow-up stages of B2B prospecting. It sources contacts, writes personalised emails, sends sequences, classifies replies, and routes positive responses to human reps. As of 2026, the best AI SDRs can take a prospect from cold list to booked meeting without human input, though quality depends heavily on the underlying contact data.
What is an AI SDR? An AI SDR (AI sales development representative) is a software agent that automates outbound prospecting tasks traditionally done by a human SDR: finding target accounts, sourcing contacts, crafting personalised outreach, following up, and booking meetings. The best AI SDRs in 2026 can handle the full workflow end-to-end, though human oversight at key steps is still the best practice for quality control.
What is an AI SDR tool? An AI SDR tool automates some or all of the work traditionally done by a Sales Development Representative: finding prospects, researching accounts, writing outreach, and booking meetings. The category spans from tools that require human approval at every step to platforms that run fully unsupervised.
No commitment. No credit card. Just 50 free verified contact lookups.