ChatGPT Prompts for Cold Email That Actually Work (Tested)

Most ChatGPT cold email prompts are terrible

You’ve seen the posts: “10 ChatGPT prompts to write cold emails.” You copy the prompt, paste it in, and get back something that sounds like every other AI-generated email on the internet.

The problem isn’t ChatGPT. It’s the prompt. Generic input produces generic output.

These 6 prompts work because they’re specific, they include context, and they constrain the output to avoid AI slop.

Prompt 1: ICP research

Use this when: You’re entering a new market segment and need to understand what problems your product solves for them.

The prompt:

Analyze the role of [VP of Sales] at [B2B SaaS companies with 50-200 employees
that recently raised Series A or B funding].

Identify:
1. Their top 3 daily frustrations related to [outbound sales and pipeline generation]
2. The tools they're likely using and paying for
3. What would make them respond to a cold email (what's in it for them?)
4. What language/jargon they use when talking about these problems

Be specific. No generic "challenges." Give me actual scenarios they face
on a Tuesday morning.

Example output (condensed):

Top frustration #1: Board is pushing for 3x pipeline growth post-funding, but the SDR team is 2 people and the outbound stack is held together with Zapier. They’re spending Tuesday mornings debugging why 30% of yesterday’s emails bounced instead of strategizing.

Likely tools: Apollo ($79/mo), Instantly ($97/mo), Clay ($149/mo), Google Workspace (15 seats at $108/mo). Total: $433/mo before the SDR salaries.

What makes them respond: Concrete cost savings with numbers (“cut your stack from $433 to $149”). Not “improve your outbound” — they hear that 20 times a week.

What to tweak: The role, company profile, and problem domain. The more specific you are about the ICP, the more useful the output. “VP of Sales at SaaS” is too broad. “VP of Sales at B2B SaaS companies with 50-200 employees post-Series B hiring their first SDR team” is specific enough to produce actionable insights.

Prompt 2: Subject line generation

Use this when: You need 5-10 subject line variations for A/B testing.

The prompt:

Write 10 cold email subject lines for this scenario:

- Sender: [outbound platform that replaces 5+ tools with one]
- Recipient: [VP Sales at Series B B2B SaaS, 80 employees]
- Core value prop: [cut outbound tool costs from $600/mo to $149/mo]
- Tone: direct, peer-to-peer, not salesy

Rules:
- Under 7 words each
- No questions
- No "quick question" or "touching base"
- No emojis
- No ALL CAPS
- Must feel like a human wrote it to a colleague

Example output:

1. Your outbound stack costs too much
2. $600/mo for cold email is broken
3. The tool consolidation math
4. Outbound costs after Series B
5. Re: your outbound infrastructure
6. 74% less on your email stack
7. Replacing 5 tools with 1
8. Your Apollo + Instantly bill
9. Outbound at 80 people
10. The $450/mo you don’t need to spend

What to tweak: The constraints matter more than the scenario. “No questions” eliminates the lazy “Quick question?” subject lines. “Under 7 words” forces brevity. “Must feel like a human wrote it” prevents AI-sounding phrasing. Adjust constraints based on what your audience responds to.

Prompt 3: First-line personalization

Use this when: You have prospect data (LinkedIn, company info) and need personalized openers at scale.

The prompt:

Based on this information about a prospect, write 3 personalized
opening lines for a cold email.

Prospect info:
- Name: [Sarah Chen]
- Title: [Head of Revenue Operations]
- Company: [CloudSync, Series B SaaS, 120 employees]
- Recent activity: [Posted on LinkedIn about struggling to attribute
  pipeline to outbound vs inbound sources]
- Company news: [Just hired 3 SDRs in the last month]

Rules:
- Reference something specific they did or said (not their job title)
- Under 25 words
- Don't be creepy or stalkerish
- Don't start with "I noticed" or "I saw"
- Sound like a human, not a marketer

Example output:

1. “Pipeline attribution between inbound and outbound is a nightmare when the SDR team triples in a month, Sarah.”

2. “Three new SDRs and no way to tell which channel is actually generating pipeline — I’ve seen that movie before.”

3. “Your LinkedIn post about attribution hit close to home. It gets harder when you’re scaling outbound and inbound simultaneously.”

What to tweak: The “Rules” section is critical. “Don’t start with ‘I noticed’” kills the most overused cold email opener in existence. “Don’t be creepy” prevents ChatGPT from referencing personal details that feel invasive. Add your own constraints based on patterns you want to avoid.

Prompt 4: Pain point matching

Use this when: You have signal data about a company and need to identify which pain point to lead with.

The prompt:

I sell [an all-in-one outbound platform that replaces separate tools
for leads, enrichment, verification, sending, and infrastructure].

Here are signals I've detected about this company:
- [Hiring 3 SDRs (job postings on LinkedIn)]
- [Using Instantly + Apollo (from technographic data)]
- [Raised Series B 4 months ago ($25M)]
- [Company size grew from 60 to 95 in 6 months]

Based on these signals, identify:
1. The #1 pain point they're likely experiencing RIGHT NOW
2. Why this pain point is urgent (what happens if they don't fix it)
3. A one-sentence cold email opener that addresses this pain
4. The specific dollar amount or time cost this pain is creating

Be specific. Use the signals to make real inferences, not generic statements.

Example output:

#1 pain point: They’re about to scale outbound 3x (3 new SDRs) but their tool stack (Instantly + Apollo) doesn’t scale linearly. Each SDR needs their own Apollo seat ($79/mo) and the total stack cost will triple from ~$500/mo to ~$1,500/mo.

Why it’s urgent: The Series B board expects pipeline results in Q3. If the SDRs spend their first month setting up tools and debugging integrations instead of selling, that’s $45K in salary with zero pipeline to show for it.

One-sentence opener: “Scaling from 1 SDR to 4 on Instantly + Apollo is about to cost you $1,000/mo more — and the onboarding headache is worse than the bill.”

Dollar cost: ~$1,000/mo additional tools + ~$15K in SDR productivity loss during first month.

What to tweak: The signals you include determine the quality of the output. More signals = better inferences. If you only have “they use Instantly,” the output will be generic. If you have hiring + tech stack + funding + growth, the output is specific and actionable.

The limitation: This works when you manually paste prospect data into ChatGPT. It doesn’t scale to 500 prospects per day. That’s where built-in AI personalization matters — it runs this analysis automatically for every prospect in your campaign.

Prompt 5: Follow-up sequences

Use this when: Your initial email didn’t get a reply and you need follow-ups that add value instead of just bumping.

The prompt:

Write a 3-email follow-up sequence for this scenario:

Initial email: [Sent to Head of RevOps about replacing their $600/mo
tool stack with one platform at $149/mo. No reply after 4 days.]

Rules:
- Each follow-up must add NEW information (not just "bumping this up")
- Follow-up 1: Add social proof (case study or data point)
- Follow-up 2: Address the most likely objection (switching cost)
- Follow-up 3: Breakup email (last attempt, create urgency)
- Each email under 75 words
- No "just following up" or "circling back"
- 3-5 days between each email

Example output:

Follow-up 1 (Day 4): Sarah — a similar-sized SaaS team (Vertex, 90 employees) consolidated from Instantly + Apollo + Clay to one platform last quarter. Their outbound cost dropped from $720/mo to $349/mo. Reply rates actually went up because enrichment → verification → sending happens in one flow with no data handoffs. Worth 15 minutes?

Follow-up 2 (Day 8): Sarah — the usual concern about switching: “migration is painful.” Fair. But here’s the real cost of staying: 3 new SDRs × setup time across 4+ tools × first month ramp = roughly $15K in lost productivity. Migration to a single platform takes a day. The math only goes one direction.

Follow-up 3 (Day 13): Sarah — I’ve reached out a few times about consolidating your outbound stack. If the timing’s wrong, no worries. If you want to see the Instantly → SendEmAll migration path, I’ll send it over. Either way, I’ll close the loop on my end after this.

What to tweak: The rules constrain the output quality. “Each follow-up must add NEW information” is the most important constraint. Without it, ChatGPT produces variations of “just checking in.”

Prompt 6: Objection handling

Use this when: A prospect replies with an objection and you need to respond thoughtfully.

The prompt:

A prospect replied to my cold email with this objection:
"[We're locked into annual contracts with our current tools until Q4.]"

Context:
- They're currently using [Apollo + Instantly + Clay]
- My product replaces all three for less money
- Their contract ends in [~6 months]

Write a reply that:
1. Validates their situation (don't dismiss the objection)
2. Plants a seed for when the contract ends
3. Offers something useful NOW (not a pitch)
4. Keeps the door open without being pushy
5. Under 80 words

Example output:

Totally fair — breaking annual contracts rarely makes financial sense. When Q4 hits and renewal comes up, the comparison worth doing: what you’re paying across all three tools vs. what a consolidated platform costs. I put together a cost calculator that maps your exact stack to potential savings. Happy to send it over so you have the numbers ready for renewal season. No pitch, just math.

What to tweak: The specific objection changes everything. Common objections to pre-load: “we’re happy with our current tools,” “we don’t have budget,” “now’s not a good time,” “we tried something like this before,” “our team doesn’t want to switch.” Build a prompt library for each.

The manual prompting ceiling

These prompts work. But they hit a ceiling at scale.

The workflow for manual AI personalization:

1. Research the prospect (5-10 minutes)
2. Gather signal data (3-5 minutes)
3. Write and tweak the ChatGPT prompt (2-3 minutes)
4. Review and edit the output (2-3 minutes)
5. Total: 12-21 minutes per prospect

At 50 prospects/day, that’s 10-17 hours. At 200 prospects/day, it’s physically impossible for one person.

What built-in AI personalization does differently:

SendEmAll’s AI personalization runs the equivalent of Prompts 3, 4, and 5 automatically for every prospect. It uses the enrichment data already in the platform — hiring signals, funding data, tech stack, company news — to personalize each email without manual prompting.

You set the angle once (“focus on cost savings for growing teams”). The AI adapts it per-prospect with specific data points. At 500 emails/day, every email is personalized. Time investment: minutes, not hours.

ChatGPT prompts are the right tool for strategy, testing, and small batches. For daily outbound at scale, you need personalization built into the sending platform.

Start sending personalized emails at scale

100 free credits. See how AI personalization works without manual prompting. Compare the output quality against your best ChatGPT prompt.

Start free — 100 credits, no card required →