Outbound Agency / Cold Outreach Beginner

31 ChatGPT prompts for outbound agencies to craft offer-first cold outreach scripts

If you run an outbound agency in 2026, the only thing that matters is whether your offer lands in the first 8 seconds. Everything else is noise.

Cold outreach is rougher than it used to be. Inbox competition is brutal. Buyers skim faster. AI-generated copy has flooded inboxes, so the bar for “feels human” jumped overnight. That’s why I built this list of 31 ChatGPT prompts for offer-first cold outreach: they’re the exact scripts, ICP briefs, hook generators, and offer-reveal templates I use to book meetings for B2B clients running 4-7 step sequences.

These prompts are designed for outbound agencies and lead-gen shops that want to scale cold outreach without sounding like a spam farm. They work with the tools you already pay for: Apollo, Clay, Instantly, Smartlead, Lemlist, Reply.io, Mailshake, Salesloft, Outreach, and ZoomInfo. They follow the “offer-first” pattern popularised by cold outreach creators like Alex Berman, Patrick Dang, and Jordan Lee, where the value promise shows up before the pitch, and they pair with the “Lead Magnet → Conversation → Call” flow and Jill Konrath’s “Value-Based Outreach” thinking.

I’ll show you the offer-first script anatomy, the 31 prompts split across 6 sections, a funnel mapping table, an 8-question FAQ, and a 14-day sprint to hit your first 1,000 leads. Let’s go.

TL;DR (Pull Quote): A 4-7 email follow-up sequence can more than triple reply rates vs. a single send, and personalized subject lines pull 50% higher opens - but only when the body actually delivers an offer in the first two lines. The 31 ChatGPT prompts for offer-first cold outreach below are built to do exactly that, with verified 2026 data from Instantly, Smartlead, and Lemlist.

Why “just checking in” is dead in 2026

“Just checking in” never worked, but in 2026 it’s a self-own. Smartlead’s 2026 data set shows optimized cold emails stay under 80 words, send Wednesday 9-11 AM, and lead with a clear point of view (Smartlead, April 2026). And a 4-7 step follow-up sequence more than triples reply rates versus a single send, yet 70% of sales reps still stop after one email (Instantly, Jan 19 2026).

Lemlist’s 350M+ email and 4.5M LinkedIn message dataset confirms it: higher volume correlates with lower reply rates. Teams that send fewer, better-timed messages consistently outperform volume stacks (Lemlist, May 12 2026). That is exactly the gap the offer-first framework fills: lead with a concrete, specific, useful offer, and the rest of the email earns its right to exist.

The 5-step offer-first script anatomy

Every offer-first cold outreach script in this library follows the same five-part skeleton. I learned this from Patrick Dang’s offer-first teaching and adapted it for agencies running Apollo + Clay + Instantly pipelines.

Trigger or signal (1 line): a real reason this email exists today, like a recent hire, funding round, G2 review, or job change.
Specific observation (1 line): something you noticed about the prospect’s company, not a generic “I love what you’re doing” line.
The offer (2-3 lines): a concrete, bite-sized, useful thing. A teardown, a 90-second audit, a 3-bullet roadmap, a back-of-the-napkin Loom. No “book a 30-minute call” yet.
Social proof or specificity (1 line): a name, number, or named case study that mirrors the prospect.
Low-friction CTA (1 line): a yes/no question, not a meeting ask.

The 31 prompts below generate each part. Stack them in order for a full sequence, or run them in isolation when you need to refresh a single step.

SECTION 1: Offer brainstorm prompts (Prompts 1-5)

These five prompts force ChatGPT to help you design the actual offer that will live in step 3 of the script. Most agencies skip this and jump straight to “write me a cold email.” That is why their emails die.

Prompt 1: Generate 10 bite-sized offers for a specific ICP

Purpose: Build a menu of offers you can rotate across a 4-7 step sequence. The best outbound agencies pick 2-3 offers per campaign, not one.

The prompt:

Act as a B2B outbound strategist who has run campaigns for
Belkins-style agencies. My client is {client name}, a {client type}
selling {product/service} to {ICP job title} at {ICP company size}
{ICP industry}.

The pain we solve is {specific pain in one sentence}. The outcome
a customer gets is {specific measurable outcome}.

Generate 10 bite-sized, no-call-required offers I can put in
the body of a cold email. Each offer must:
- Take the prospect less than 5 minutes to consume
- Deliver a tangible artifact (audit, teardown, 3-bullet plan,
  calculator result, short Loom)
- Be specific to {ICP industry}, not generic
- Be doable by me in under 45 minutes of prep
- Avoid the words "book a call" or "schedule a demo"

Format each offer as:
Offer {n}: {Name}
What they get: {1 line}
Why it lands for this ICP: {1 line}
How I deliver it: {Loom / PDF / Notion link / Notion doc / Google Doc}
Prep time: {X} minutes

Example output (for a SaaS client selling to RevOps leaders at 100-500 employee B2B SaaS companies):

Offer 1: 90-Second G2 Teardown. A 90-second Loom walking through their public G2 reviews and flagging the 3 most-asked-for features their competitors already ship. Prep time: 35 min. Why it lands: RevOps sees review velocity as a churn leading indicator.
Offer 2: Outbound Reply-Rate Benchmark Card. A one-page Notion doc comparing their last 30 days of cold email reply rate to the 2026 industry average (15-28% open, 1-5% reply, per Instantly 2026 data). Prep time: 15 min. Why it lands: concrete, free, makes their current vendor look expensive.

Pro tips:

Run this prompt 3-4 times, then mix and match the 30-40 offers you get. The best ones come from prompt #3 or #4.
Replace the offer every 6-8 weeks in active sequences to avoid fatigue.
If a prospect replies to step 1 with “not now,” you have 9 backup offers for the follow-ups.

Prompt 2: Refine one offer into a one-sentence promise

Purpose: Compress a full offer description into a single, clickable sentence that fits the body of an email. This is where most agencies lose the room - they describe the offer in 40 words and bury the CTA.

The prompt:

Here is the offer I want to put in a cold email:

{paste full offer description}

Compress it into exactly one sentence (15-25 words) that:
- Names a specific deliverable (audit, teardown, plan)
- Names a specific time cost (90 seconds, 3 minutes, 5 minutes)
- Avoids the words "free", "just", "simply", and "schedule"
- Sounds like a sentence a smart friend would actually say
- Does NOT ask for a meeting

Give me 5 variations. Then pick the strongest and explain why.

Example output (for a cybersecurity agency targeting CFOs at mid-market manufacturers):

“I recorded a 2-minute Loom showing the 3 supplier-vendor risk gaps in your SEC 10-K cyber disclosure - want me to send it?”

Pro tips:

Read the 5 variations out loud. The one that sounds like a sentence you’d actually say wins.
Time-cost + specific artifact is the magic combo. “2-minute Loom” beats “quick video” every time.
Use this exact prompt for follow-up #2, #3, and #4 too. The offer stays the same, the phrasing rotates.

Prompt 3: Map the offer to a real trigger event

Purpose: Tie the offer to a specific trigger that makes it timely. Outbound agencies that lead with trigger events convert 2-3x better, per Lemlist’s 350M+ email dataset.

The prompt:

Here is my offer:

{insert the one-sentence offer from Prompt 2}

I am reaching out to {job title} at {company type} in {industry}.

Brainstorm 8 trigger events that would make this offer feel
urgent and timely for the prospect. For each trigger, write:
- The trigger event (1 line)
- A 1-sentence sentence I could put in the FIRST line of
  the email that references the trigger naturally
- The data source I would need to find this trigger at scale
  (Apollo, Clay, Bombora, LinkedIn Sales Navigator, manual)

Prioritise triggers I can find in Apollo or Clay enrichment,
not triggers that need expensive intent data.

Example output (for a HubSpot onboarding agency):

Trigger 1: “Saw your marketing team posted 3 new HubSpot Admin job listings this month - usually a sign the CRM is straining under the new campaigns.” Data source: LinkedIn Sales Navigator job filter.
Trigger 2: “Your G2 reviews dropped 0.4 stars since January - usually a workflow issue, not a tool issue.” Data source: G2 public scrape via Clay.

Pro tips:

“Noticed” beats “Congratulations.” Trigger emails that congratulate the prospect feel sycophantic.
Stack 2 trigger sources per email max. More than that feels stalker-ish.
For agency clients in regulated industries, the trigger should ideally be public (10-K, G2, job board) so you can defend it in a reply.

Prompt 4: Build an offer matrix for an entire 4-7 step sequence

Purpose: Plan the entire sequence in one shot so step 1 doesn’t accidentally promise what step 3 delivers. This is the prompt that turns a “single email” mindset into a “sequence” mindset.

The prompt:

I am building a 4 to 7 step cold email sequence in
{offer-first style}. My ICP is {job title} at {company type}.
The dream outcome for the prospect is {X}. The fear/anti-offer
is {Y}.

For each step in the sequence, output:
- Step number
- Day to send (relative to step 1)
- Channel (email / LinkedIn / call / SMS)
- Offer type (audit / teardown / case study / question / breakup)
- One-sentence body summary (the offer + the trigger if any)
- What success looks like (reply / click / call answered / etc.)

Rules:
- Step 1 must lead with the strongest offer, no warm-up
- Step 4 or 5 must be a value-add (case study, fresh data point)
- The last step must be a breakup email
- No step can ask for a meeting before step 3
- The offers should escalate in specificity, not intensity

Example output (truncated):

Step	Day	Channel	Offer type	One-sentence body summary	Success signal
1	1	Email	90-second Loom audit	”Recorded a 90s Loom on 3 gaps in your supplier-risk page.”	Reply
2	3	LinkedIn	Profile view + connection note	”Sharing the Loom here in case it gets buried.”	Accept
3	5	Email	Case study	”Just closed a 200-person manufacturer with a 41-day cycle.”	Reply
4	8	Email	Fresh data point	”Pulled 2026 SEC cyber disclosure data for your peer set.”	Click
5	12	Call	Voicemail	”Tried twice, no worries, here’s the Loom one more time.”	Call back
6	16	Email	Breakup	”Closing the file on my end - last chance for the Loom?”	Reply

Pro tips:

4-7 steps triples reply rate vs. one email, per Instantly 2026 data.
Don’t send 7 emails in 7 days. Smartlead’s 2026 cadence data shows 4-6 combined touches per week across ALL channels is the realistic ceiling before spam complaints spike.
Save this prompt’s output as the brief for every downstream email generation prompt.

Prompt 5: Reverse-engineer what your top-performing competitor is offering

Purpose: Steal the structure of competitor offers without copying the words. This is the prompt I run before launching any new client campaign.

The prompt:

Here are 5 cold emails my prospect has likely received in the
last 30 days. They are from competitors in {industry}. I will
paste excerpts:

{Insert 5 email excerpts}

Analyse them and tell me:
- The common offer pattern (audit / case study / ebook / call)
- The trigger event pattern (job change / funding / review drop)
- The CTA pattern (reply yes / click link / book call)
- The offer that is most overused (and therefore least likely
  to land in 2026)
- 3 offer angles the competitors are NOT using, that I could
  use to stand out

Output as a table.

Pro tips:

Sign up for 3-5 competitor mailing lists with a dedicated mailbox in Instantly. Watch what they send over 30 days.
If all 5 competitors are offering a “free 30-min consultation,” that’s the offer to AVOID. It’s the new “just checking in.”
Test the “NOT using” angles in a 200-lead pilot before scaling.

SECTION 2: ICP & data prompts (Prompts 6-10)

These prompts turn Apollo, Clay, and ZoomInfo exports into the raw material for a real offer-first script. Most agencies skip ICP depth and blame the copy when reply rates tank. The copy is only as good as the research.

Prompt 6: Turn an Apollo list export into a sharp ICP brief

Purpose: Convert 50 rows of Apollo contact data into a one-paragraph ICP brief that informs every downstream script.

The prompt:

I have exported 50 contacts from Apollo for this campaign.
Below is the CSV data (anonymised, no real names):

{paste CSV with columns: company, employee count, industry,
title, city, technologies, funding, hiring status}

Analyse the list and tell me:
1. The dominant ICP pattern (job title cluster, company size
   band, industry concentration)
2. The 3 sub-segments inside the list that should get DIFFERENT
   email scripts
3. The 5 contacts most likely to reply within 14 days, and
   why (look for hiring triggers, funding, tech stack mismatch)
4. The 10 contacts least likely to reply, and why
5. The single best offer-first hook for this exact list

Output as a structured brief I can paste into a Clay enrichment
table.

Pro tips:

Run this prompt BEFORE writing a single email. It surfaces patterns you would miss at row-level.
If “single best offer-first hook” surprises you, trust it. AI often spots an angle a human will rationalise away.
Pair this with Clay’s “lookalike” enrichment to expand the strongest sub-segment to 1,000+ contacts.

Prompt 7: Pull the “Day in the Life” of the buyer

Purpose: Understand the buyer’s actual day so the email lands when they’re most receptive. Lemlist’s 2026 data says send-time matters: 1 PM is the highest-reply window, weekends are dead (Instantly 2026).

The prompt:

I sell {product/service} to {job title} at {company type}
in {industry}. Write a realistic "Day in the Life" for this
buyer, including:
- Their first 30 minutes at work (which tool do they open first?)
- The 1-2 meetings that take up most of their morning
- The kind of Slack/Teams messages they ignore vs. respond to
- The exact time window they read personal-looking emails
  (not newsletters)
- The 3 things that would make them say "not now" to a cold
  email today
- The 3 things that would make them say "tell me more"

Output as a timeline from 8 AM to 6 PM.

Pro tips:

Use the output to choose your send window. If they live in Slack, send 10:30 AM (post standup) not 6 AM.
The “make them say not now” list is more valuable than the “tell me more” list. Those are the objections to pre-empt in the body.
Run this prompt for each of the 3 sub-segments from Prompt 6. You’ll find their days look different.

Prompt 8: Build a Clay enrichment waterfall for trigger signals

Purpose: Plan a Clay table that pulls trigger events cheaply, in the right order, with credit-cost optimisation.

The prompt:

I want to enrich a list of 1,000 {ICP contacts} in Clay.
For each contact I need these 8 trigger signals:
{list signals: e.g., recent LinkedIn post, job change,
funding event, G2 review drop, tech stack change,
hiring surge, podcast appearance, competitor mention}

Output:
- The recommended waterfall order (cheapest data provider
  first, fall-back providers next)
- The estimated credit cost per contact
- Which 2 signals are "must-have" (sequence stops without them)
- Which 3 signals are "nice-to-have" (used in step 3+ only)
- Which signals I should ignore because they're noisy

For each signal, name 2 specific Clay data providers and the
Apollo.io / LinkedIn Sales Navigator fallback.

Pro tips:

Clay waterfalls save 30-50% on credit costs vs. hitting a single provider, per Clay’s own published 2026 guides.
“Hiring surge” and “job change” are the two highest-intent signals across 2026 outbound data.
If a signal only enriches 5% of the list, it is not worth a column. Drop it.

Prompt 9: Write a “negative ICP” filter

Purpose: Save your sender reputation by excluding contacts that look like your ICP but actually convert poorly. Smartlead’s case studies show 50%+ positive reply rates come from tight list filtering, not volume.

The prompt:

Based on this list of contacts that I should EXCLUDE from
my campaign (these are companies or titles that have replied
"not a fit" 3+ times in the last 90 days):

{paste list of 30-50 excluded companies/titles with reason}

Generate a "negative ICP" filter I can apply in Apollo before
the next export. Include:
- The 5 most common exclusion patterns (industry, size, title,
  tech stack, geography)
- A 1-line rule for each pattern, written so a VA can apply it
  to a 5,000-row list in under 30 minutes
- The pattern I am most likely to over-include (false positives
  that look like the ICP but are not)

Output as a checklist.

Pro tips:

Negative ICPs are how Belkins-style agencies hit 4-7% reply rates on cold traffic. They are not optional.
Re-run this prompt every 90 days. Your ICP drifts as your product evolves.
Apply the filter at the Apollo query stage, not after. Post-export filtering is a reputation tax.

Prompt 10: Build a 1-paragraph account snapshot for ABM-style sequences

Purpose: For high-value accounts, generate a snapshot you can paste into the first line of an email or a LinkedIn note. Personalization at this depth is what separates a 1% reply rate from a 7% reply rate.

The prompt:

Here is the public data I have pulled for {Company Name}:

- Recent news: {1-2 lines}
- Tech stack (from BuiltWith or Apollo): {list}
- Latest funding round: {date, amount, lead investor}
- Recent job changes (last 90 days): {list}
- G2 / Capterra reviews (last 6 months): {count and average}
- A quote from their CEO or VP that I can reference:
  "{quote}"

Write a 3-sentence account snapshot I can paste into the first
paragraph of a cold email. The snapshot must:
- Mention 1 specific fact that proves I actually researched
  the account
- Hint at a likely pain without naming our product
- Set up the offer that comes in the next paragraph
- Avoid flattery ("love what you're doing")
- Sound like a sharp operator, not a fanboy

Example output (for a Series B HR tech company):

“Saw you closed the $32M Series B in March and posted 14 open roles in the last 60 days - usually a sign that the go-to-market motion is scaling faster than the hiring pipeline can support. Most RevOps leaders I’ve talked to in this window are quietly re-evaluating their outbound stack because the original one wasn’t built for 4x volume.”

Pro tips:

Use this for accounts with deal size > $25K ACV. For smaller deals, run Prompt 1’s generic offers.
Feed ChatGPT the actual quote from the CEO. Verifiable quotes outperform fabricated ones every time.
The 3-sentence snapshot is the body of a LinkedIn connection note on its own.

SECTION 3: Hook & opener prompts (Prompts 11-15)

These five prompts write the actual first line and the subject line - the parts that decide whether the email gets opened at all. Instantly’s 2026 data says personalized subject lines pull 50% higher opens and subject lines with numbers can pull 113% higher opens.

Prompt 11: Generate 10 subject lines for a specific offer

Purpose: Pick the best subject line from 10 variations. Personalized, short (under 50 chars per Smartlead’s 2026 email optimization guide), and free of spam triggers.

The prompt:

Generate 10 subject lines for this cold email:
- Offer: {one-sentence offer from Prompt 2}
- Prospect job title: {job title}
- Prospect company: {company name or persona}
- Industry: {industry}

Rules:
- Each subject line must be under 50 characters
- At least 4 must include a number (e.g., "3", "90s", "2026")
- At least 2 must be phrased as a question
- Avoid: "Quick question", "Touching base", "Following up",
  "Checking in", "Help wanted", "Free", "$$$"
- At least 2 must reference the trigger event from Prompt 3
- Tone: smart operator, not marketer

Then rank the 10 by predicted open rate, and explain the top 3.

Example output (for a RevOps SaaS selling to a Head of RevOps at a 200-person SaaS):

1. 3 gaps in your G2 reviews (28 chars, +113% number lift)
1. Saw your SDR team doubled - quick Loom (38 chars, trigger)
1. Reply rate below 1.4%? (22 chars, question)
1. 90s on your reply funnel (24 chars, specific time-cost)

Pro tips:

A/B test 2 subject lines per 500 contacts. Don’t test 5, you’ll never get a clean read.
Subject line numbers should be specific, not round. “3 gaps” beats “Some gaps” by 113% in Instantly’s 2026 dataset.
If your open rate is under 30%, the subject line is the problem. If it’s over 50%, the body is.

Prompt 12: Write the first 2 lines of the email (the “above-the-fold” zone)

Purpose: The first 2 lines decide whether the prospect reads the offer. This is where the trigger + observation live.

The prompt:

Here is the offer I am leading with:
{one-sentence offer}

Here is the trigger event I want to reference:
{trigger event from Prompt 3}

Here is the specific observation about the prospect's company:
{1-2 lines from Prompt 10}

Write the first 2 lines of the cold email. The first 2 lines
must:
- Be visible in the Gmail / Outlook preview pane (no "Hi {Name},"
  above them)
- Reference the trigger event naturally
- Reference the specific observation (not "I love your work")
- Make the prospect curious enough to read line 3 (the offer)
- Sound like a real sentence a colleague would say
- Avoid the words "I", "we", "our", "your" in the first line if
  possible (start with the trigger, not with yourself)

Give me 5 variations. Then pick the best 2 and explain why.

Example output:

Variation 1: “Saw your RevOps team posted 3 new SDR roles in the last 30 days - usually means the current outbound motion is straining under the volume. Recorded a 90-second Loom on 3 specific levers most Series B SaaS teams miss in this window. Want me to send it?”

Pro tips:

Read the first 2 lines in the Gmail mobile preview. If they don’t make sense without the rest of the email, rewrite.
“Saw” beats “I noticed.” It’s shorter and sounds more like a peer.
For follow-up #2, the first 2 lines should reference the original email, not repeat it. Use the word “bumping” sparingly.

Prompt 13: Build a 3-part AIDA opener (for LinkedIn and longer emails)

Purpose: When the channel allows more room - LinkedIn InMail, a longer email opener, a cold call script - use a tighter version of AIDA (Attention, Interest, Desire, Action). This pairs well with Patrick Dang’s offer-first teaching and Jill Konrath’s “Value-Based Outreach” book.

The prompt:

Write a 4-paragraph cold outreach opener using a tight AIDA
structure. The channel is {LinkedIn InMail / long-form email
/ cold call opener}. The offer is:

{insert offer from Prompt 2}

Constraints:
- Paragraph 1 (Attention): 1-2 sentences, lead with the
  trigger event. No "Hi {Name}".
- Paragraph 2 (Interest): 2-3 sentences, build the specific
  observation. Reference 1 fact about the prospect's company.
- Paragraph 3 (Desire): 2-3 sentences, deliver the offer
  AND the specific outcome the prospect gets. No vague value
  claims. Use a number.
- Paragraph 4 (Action): 1 sentence, yes/no question, no
  meeting ask.
- Total length: under 90 words
- Tone: peer-to-peer, sharp, no flattery

Give me 3 variations tuned for 3 different channels (email,
LinkedIn, voicemail).

Pro tips:

Use the “long” version sparingly. Most of your volume should be the 50-80 word version from Prompt 12.
Cold call voicemails should be 18-22 seconds. Read the variation out loud and time it.
For LinkedIn InMail, you have ~300 characters before the “see more” cut. Front-load the offer.

Prompt 14: Write a “pattern interrupt” subject line for fatigued lists

Purpose: When a contact has been on your list for 60+ days with no reply, the standard subject lines stop working. Pattern interrupts break the pattern.

The prompt:

I have a list of {X} contacts who have been on my cold email
sequence for 60+ days with no reply. Standard subject lines
("quick question", "thoughts on this") are no longer working.

Generate 10 "pattern interrupt" subject lines that:
- Are visibly different in tone from the original sequence
- Use lowercase or unusual punctuation
- Are under 35 characters
- Sound like a peer text, not a marketer email
- At least 3 are 1-2 words only
- At least 2 are emoji-free
- At least 2 are questions with no question mark

Rank the 10 by predicted open rate for a fatigued list.

Example output:

unsubscribe if not relevant (28 chars)
closing the file (16 chars)
last attempt (11 chars)
wrong person? (12 chars)
no, really (9 chars)

Pro tips:

Pair the pattern interrupt subject with a “breakup” body (Prompt 20). The combo is the classic 4-7% reply-rate play.
Don’t use pattern interrupts on step 1. They are for re-engagement only.
Lowercase subject lines outperform Title Case in most 2026 tests I have seen across Instantly and Smartlead.

Prompt 15: Localize the opener for non-English-speaking markets

Purpose: For agencies running campaigns in EMEA or APAC, the opener has to land in the prospect’s language and culture. Lemlist’s 2026 EMEA data set shows 350M+ emails confirm language-matched openers outperform English-only by 22-40%.

The prompt:

Translate and culturally adapt this cold email opener into
{language} for a prospect in {country}:

{insert opener from Prompt 12 or 13}

The prospect's job title in {language} is {translated title}.
The cultural norms for business emails in {country} are:
- {norm 1}
- {norm 2}
- {norm 3}

Output:
- A direct, formal translation
- A formal-but-warm adaptation that sounds like a local would
  actually write
- A 1-line note on what to remove or soften for this culture
  (e.g., remove the direct question, soften the CTA)

Pro tips:

DACH (Germany, Austria, Switzerland) markets respond to more formal openers. Don’t use “hey” in step 1.
APAC markets often need a softer CTA. “Worth a look?” outperforms “Want me to send it?” in Japan and Korea.
Always run the localized version through a native speaker before scaling. ChatGPT translations are a starting point, not a final draft.

SECTION 4: Body & offer reveal prompts (Prompts 16-20)

These five prompts build the body of the email - the part that delivers the offer. This is where most cold emails fail. The opener is great, the CTA is great, and the middle is a foggy paragraph about the company’s “mission.”

Prompt 16: Write the offer reveal paragraph (step 3 of the skeleton)

Purpose: Make the offer feel concrete, specific, and easy to say yes to. The offer is the entire reason the prospect should reply.

The prompt:

Write a 3-sentence "offer reveal" paragraph for a cold email.

The offer is: {one-sentence offer from Prompt 2}
The time cost for the prospect is: {X minutes / seconds}
The artifact I am sending is: {Loom / PDF / Notion doc /
Google Doc / calculator}
The specific outcome they get is: {one measurable outcome}

Rules:
- Sentence 1: name the deliverable + the time cost
- Sentence 2: name the specific outcome (use a number)
- Sentence 3: name the artifact and how I'll deliver it
- No "book a call" or "schedule a demo"
- No vague outcomes ("save time", "grow faster")
- No exclamation marks
- Total length: under 50 words
- Tone: confident peer, not a salesperson

Give me 3 variations.

Example output:

“I put together a 2-minute Loom walking through the 3 specific supplier-risk gaps in your last 10-K cyber disclosure. The 3 manufacturers I have shown this to in the last 60 days all closed at least 1 of the gaps inside 30 days. Happy to send the Loom here if it’s useful.”

Pro tips:

The “3 specific” / “2-minute” / “30 days” combo is the highest-converting structure in 2026 cold email data.
“Happy to send” outperforms “Want me to send” because it removes the work for the prospect.
If the offer involves a Loom, name the recording length in the email. It signals respect for the prospect’s time.

Purpose: Borrow credibility from a similar customer without sounding like a generic case study. The best social proof is a peer in the prospect’s exact industry, with a number that mirrors the prospect’s world.

The prompt:

Write a 1-sentence social proof line for a cold email.

The prospect's job title is: {job title}
The prospect's industry is: {industry}
The prospect's company size is: {size band}

The customer I want to reference is: {similar customer or persona}
The result that customer got is: {specific measurable result}

Rules:
- 1 sentence only
- Name the customer OR a persona (e.g., "a 200-person
  manufacturer in Ohio")
- Include 1 number (the result)
- Avoid: "trusted by", "leading", "industry-leading",
  "best-in-class", "Fortune 500"
- Avoid the word "company" - use "team", "group", "shop"
- Sound like a peer, not a testimonial card
- Total length: under 25 words

Give me 5 variations.

Example output:

“Did the same teardown for the procurement team at a 180-person food manufacturer in Iowa last quarter - they cut vendor-risk review time from 9 days to 2.”

Pro tips:

“Cut X from Y to Z” is the highest-converting social-proof structure. Numbers + before/after beats abstract praise.
If you don’t have a customer story yet, use a personal experience: “Spent 4 years on the buying side of this exact problem at {previous company}.”
One social-proof line per email. Two feels like a case study.

Prompt 18: Write the low-friction CTA (step 5 of the skeleton)

Purpose: Replace “book a 30-minute call” with a yes/no question that the prospect can answer in 5 seconds. The CTA is the only part of the email that asks the prospect for anything, so it has to be easy.

The prompt:

Write 5 cold email CTAs for this email:

Full email context:
{paste the full email}

Rules for each CTA:
- 1 sentence only
- Phrased as a yes/no question
- No "book a call", "schedule a demo", "hop on a quick chat"
- No "let me know if you're interested" (passive)
- Sounds like a question a colleague would ask
- Time-cost of saying yes: under 10 seconds
- Time-cost of saying no: under 5 seconds
- Avoid exclamation marks

Then rank the 5 by predicted reply rate and explain.

Example output:

1. “Want me to send the Loom?” (8 words, frictionless)
1. “Worth a 2-minute look?” (6 words, low commitment)
1. “Should I send it over?” (5 words, softest)
1. “Useful or am I off base?” (7 words, gives an “off” option)
1. “Send the Loom, or leave it?” (8 words, two-option fork)

Pro tips:

“Useful or am I off base?” outperforms “Want me to send?” in 2026 because it lets the prospect say no without awkwardness.
Never put the CTA in a P.S. line. P.S. lines get skipped in mobile preview.
The CTA is the part you A/B test the most. Rotate every 2-4 weeks.

Prompt 19: Write a “value-add follow-up” body (step 3-4 of the sequence)

Purpose: Step 3 or 4 of a sequence is where most agencies repeat step 1 and watch reply rates die. The follow-up has to add new information, not repeat the original ask.

The prompt:

Write a step 3 cold email follow-up body. The original email
(step 1) was:

{paste original email}

The prospect has not replied.

Step 3 must:
- NOT repeat the original offer verbatim
- Add 1 new piece of information (a case study, a fresh data
  point, a peer benchmark, a relevant post)
- Reference the original email in 1 line ("bumping this up",
  "in case it slipped through")
- End with a new yes/no CTA
- Be under 70 words

Brainstorm 3 angles for the "new information":
1. A peer case study angle
2. A fresh data point angle
3. A "I learned something new" angle

Then write the email for the strongest angle.

Example output (data point angle):

“Bumping this up - pulled the 2026 SEC cyber disclosure data for your peer set since I last emailed. The 3 mid-cap manufacturers in your cohort are averaging 11 vendor-risk disclosures, you’re at 4. The gap is usually a workflow issue, not a staffing one. Want the data?”

Pro tips:

“New data” is the highest-converting follow-up angle. It’s also the easiest to scale with Clay + Apollo.
Follow-up #2 should be the softest (“just bumping this up”). Follow-up #3 should add value. Follow-up #4 should add a stronger trigger.
Never send 4 follow-ups that all say “just checking in.” That’s the 70%-stop-after-one-email trap, per Instantly 2026.

Prompt 20: Write a breakup email (final step of the sequence)

Purpose: The breakup email is the highest-reply-rate email in any sequence, often pulling 4-7% on its own. Use it as step 6 or 7.

The prompt:

Write a "breakup" email - the final step in a cold outreach
sequence. The prospect has not replied to steps 1-5.

The original offer was: {offer}
The trigger event was: {trigger}

Rules:
- Under 50 words total
- Tone: resigned but respectful, not bitter
- Acknowledge you're closing the file
- Give the prospect one last no-pressure way to reply
- Do NOT ask for a meeting
- Do NOT beg
- Do NOT use the words "sorry" or "apologies"
- End with a clear "this is the last email" signal

Give me 3 variations: warm, neutral, and slightly playful.

Example output (neutral):

“Haven’t heard back, so I’m assuming the timing is off or this isn’t a priority right now. Closing the file on my end. If anything changes in the next 90 days, the Loom is still at {link} - no need to reply unless you want the audit.”

Pro tips:

Breakup emails are the highest-converting email in any sequence. Many agencies see 30%+ of their total replies come from this single step.
Pair the breakup email with a pattern-interrupt subject line from Prompt 14.
“Closing the file” outperforms “moving on” because it sounds like an internal process, not a guilt trip.

SECTION 5: CTA & multi-channel prompts (Prompts 21-25)

These five prompts turn a single email sequence into a true multichannel motion across LinkedIn, SMS, call, and WhatsApp. Lemlist’s 2026 data set (350M+ emails, 4.5M LinkedIn messages) shows sequences with 3-4 follow-ups that each add a new angle significantly outperform longer single-channel stacks.

Prompt 21: Adapt an email into a LinkedIn connection note (under 300 characters)

Purpose: LinkedIn connection requests get cut off at ~300 characters in the “Add a note” field. Most agencies paste their email body, which gets truncated mid-offer.

The prompt:

Compress this cold email into a LinkedIn connection note
under 300 characters (spaces included):

{paste the full email from Prompt 12 or 13}

Rules:
- Reference the trigger event in the first 80 characters
- Hint at the offer, do not deliver it
- End with a soft question, not a meeting ask
- Avoid: "I'd love to connect", "Let's chat", "Pick your brain"
- Sound like a real human typing fast on mobile
- No hashtags, no links (LinkedIn filters them)

Give me 3 variations: trigger-led, observation-led, and
question-led.

Example output (trigger-led):

“Saw your team posted 4 SDR roles in the last month - usually a sign the outbound motion is straining. Recorded a 90s Loom on the 3 specific levers I’d flip. Worth sharing?”

Pro tips:

“Worth sharing?” outperforms “Can I send it?” because it implies the value without asking.
300 characters is the mobile-friendly cap. Longer notes feel like a sales pitch.
If the prospect accepts but doesn’t reply, do NOT pitch in the first LinkedIn message. Thank them, then wait 48 hours.

Prompt 22: Write a LinkedIn InMail for higher-value accounts

Purpose: InMail allows 1,900 characters and is best for accounts with $25K+ deal size. Use it sparingly - InMail volume ruins the channel fast.

The prompt:

Write a LinkedIn InMail for a high-value account.

The prospect is: {job title} at {company}
The trigger event is: {trigger}
The offer is: {offer}

Rules:
- 1,900 characters max
- Subject line (separate from body): under 70 characters
- First 200 characters: the offer (not the trigger)
- Middle 1,200 characters: the trigger + observation + social
  proof + offer
- Last 200 characters: yes/no CTA
- Tone: peer-to-peer, no flattery
- No links in the body (LinkedIn InMails with links underperform)
- No attachments
- No "I came across your profile" - start with the offer

Give me 1 polished version.

Pro tips:

InMail open rate is ~50%, reply rate is ~10% for well-targeted sends. Bad targeting drops reply rate to 1-2%.
Use InMail for accounts where the contact is unreachable by email (no Apollo match, no LinkedIn email).
Pair the InMail with a 5-day email sequence. Don’t send both on day 1.

Prompt 23: Write a cold call script (under 30 seconds for the opener)

Purpose: Cold calls still book more meetings than any other channel, per Lemlist’s 2026 dataset. The opener has 8 seconds before the prospect decides to stay on the line.

The prompt:

Write a 30-second cold call opener for {job title} at
{company type}.

The trigger event is: {trigger}
The offer is: {offer}

Rules:
- First 8 seconds: name + company + reason for calling
- Seconds 8-15: the trigger event (1 sentence)
- Seconds 15-25: the offer (1 sentence)
- Seconds 25-30: the soft question
- Avoid: "How are you today?", "Did I catch you at a bad time?",
  "Do you have 30 seconds?"
- Avoid: "I'm calling from" - start with "Hi {Name}, this is
  {Your Name}"
- Tone: confident, peer, not apologetic
- Total word count: under 60 words

Then write the 4 most common objection responses
("not interested", "send me an email", "we already have a
vendor", "call me back next quarter").

Pro tips:

Practice the opener until you can say it without breathing audibly. Energy matters more than word choice.
“Send me an email” is a soft yes. Ask one qualifying question before you send: “Sure - what specifically should I include?”
Track which objection comes up most. That’s the one to bake into the email body next time.

Prompt 24: Write an SMS for warm-lead follow-up

Purpose: SMS is for warm leads, per Lemlist’s 2026 SMS guide, not cold outreach. Under 160 characters, no meeting ask, send only after the prospect knows your name.

The prompt:

Write a 4-touch SMS sequence for warm leads who have replied
to a cold email OR accepted a LinkedIn connection.

The context is: {what happened in the prior touch}
The goal of the SMS sequence is: {goal: book a meeting /
confirm a meeting / re-engage a stalled deal}

Rules:
- Each SMS under 160 characters (320 hard limit)
- No links
- No meeting link in the first SMS
- Identify yourself by name in the first SMS
- Send-time logic: 2 hours before a meeting / right after a
  no-show / 1 week of silence
- Tone: peer text, not marketer
- One ask per SMS, max
- US A2P 10DLC compliant (include business name in SMS 1)

For each SMS, also output:
- When to send (relative to prior touch)
- What success looks like (reply / meeting confirmed / etc.)

Pro tips:

SMS is the highest-converting channel for meeting confirmations and no-show recovery. It is NOT a cold outreach channel.
“Hey {Name}, it’s {Your Name} from {Company} - quick one.” is the gold-standard opener.
US carriers require A2P 10DLC registration before sending. Lemlist walks through this in setup.

Prompt 25: Write a WhatsApp opener for EMEA / APAC

Purpose: WhatsApp is the dominant B2B messaging channel in LATAM, India, and parts of EMEA. Use it only with explicit opt-in or after email/LinkedIn contact.

The prompt:

Write a WhatsApp opener for {job title} at {company} in
{region}.

Context: {prior touch, e.g., "they replied to a cold email"}
Goal: {book a call / share a Loom / confirm a meeting}

Rules:
- First message must identify the sender by name and company
- Under 300 characters
- No meeting links
- No attachments in message 1
- Tone: peer, casual but professional
- Regional considerations for {region}:
  {e.g., "no emojis in DACH", "avoid direct questions in Japan",
  "lead with respect in India"}

Give me 3 variations: trigger-led, value-led, and meeting-confirm.

Pro tips:

WhatsApp is the highest-converting channel in LATAM and India for warm-lead follow-up, per Lemlist 2026.
Don’t use WhatsApp for cold outreach. Carrier filters and cultural norms both push back hard.
Keep the first message under 100 words. WhatsApp rewards brevity more than email does.

SECTION 6: Measurement & iteration prompts (Prompts 26-31)

These six prompts turn your outreach operation into a feedback loop. Without them, you’re guessing. With them, you compound.

Prompt 26: Build a campaign diagnosis framework from 4 weeks of data

Purpose: After 4 weeks, run this prompt to find the weakest link in your sequence. Smartlead’s 2026 data set is clear: open rate above 50% means the body is the problem, open rate below 30% means the subject line is the problem.

The prompt:

Here are the last 4 weeks of cold email campaign data from
{Instantly / Smartlead / Lemlist}:

- Total emails sent: {X}
- Open rate: {Y%}
- Reply rate (positive): {Z%}
- Reply rate (negative): {W%}
- Bounce rate: {V%}
- Unsubscribe rate: {U%}
- Meeting booked rate: {T%}

Analyse the data and tell me:
1. Which metric is the biggest bottleneck (be specific: open,
   reply, positive reply, meeting, deliverability)
2. The single highest-leverage change I should make this week
3. The change I should NOT make (it looks tempting but won't
   move the numbers)
4. The 2 A/B tests I should run next week (with hypothesis,
   variant, success metric)
5. The 1 deliverability risk I should check immediately

Output as a 1-page report I can paste into a Slack channel.

Pro tips:

Run this prompt weekly. Outbound is a compound game. The first 4 weeks of data are diagnostic, not conclusive.
4-6 combined touches per week across all channels is the realistic ceiling per Lemlist 2026. Don’t go above it.
Bounce rate above 3% is an infrastructure problem, not a copy problem. Pause the campaign.

Prompt 27: Build a positive-reply classifier

Purpose: Manually triaging replies is the #1 time-suck for outbound agencies. Apollo’s 2026 MCP data shows 42,000 tool calls in 30 days, with reply classification as a top use case. Use this prompt to write the classifier rules before you build it.

The prompt:

I receive {X} replies per week to my cold email campaigns. I want
to auto-classify them into 5 categories before a human reads them.

Categories:
1. INTERESTED - wants to learn more, books a meeting, asks for
   the Loom
2. NOT NOW - interested but timing is wrong
3. WRONG PERSON - directs me to someone else
4. UNSUBSCRIBE - wants to be removed
5. NOISE - auto-reply, OOO, irrelevant

Write the classification rules. For each category, give me:
- 5-7 example phrases that signal this category
- 5-7 phrases that look similar but belong in a different
  category
- The right automation action (book meeting / pause sequence
  90 days / forward to AE / remove from list / ignore)

Then write a single 1-paragraph decision-tree prompt I can paste
into an AI classifier.

Pro tips:

Auto-classification saves 4-6 hours per rep per week per Smartlead’s 2026 sales automation analysis.
“Not now” replies are gold. Route them to a 90-day re-engagement sequence, not the trash.
“Wrong person” replies often name the actual decision-maker. Parse the name automatically with regex or Apollo enrichment.

Prompt 28: Write a “weekly pipeline review” brief

Purpose: A consistent weekly brief keeps the team aligned on what changed, what to fix, and what to scale.

The prompt:

Here is this week's outbound data:

{paste the 5 key metrics: emails sent, positive reply rate,
meetings booked, show rate, opps created}

And here is last week's:

{paste the same 5 numbers}

Write a 1-page weekly pipeline review in this format:

1. Headline (1 sentence): what changed most this week
2. Wins (3 bullets): what worked, with a specific reason
3. Losses (3 bullets): what underperformed, with a hypothesis
4. Tests to run next week (2 specific A/B tests)
5. One thing to STOP doing
6. One thing to DOUBLE DOWN on
7. Risk flag (if any): deliverability, list decay, single-channel
   dependency

Tone: blunt, peer-to-peer, no fluff.

Pro tips:

Run this prompt every Friday. Save the briefs in a single Notion page. Patterns emerge after 8-12 weeks.
The “STOP doing” line is the most valuable. Most agencies are running 3-4 sequences they should kill.
Send the brief to the client on Monday morning. It’s the best client-facing artifact you can produce.

Prompt 29: Build a list decay monitor

Purpose: A “list decay” of 5-10% per quarter is normal - people change jobs, companies get acquired, emails go stale. Smartlead’s 2026 deliverability guide recommends monthly list audits to keep bounce rate under 2%.

The prompt:

I have a master cold email list of {X} contacts. I want to
detect "list decay" - contacts that have changed jobs, had
their email go stale, or work at companies that no longer
match the ICP.

Design a monthly audit workflow:
- The 5 data signals to check (with provider for each)
- The threshold for "stale" (e.g., no activity in 90 days =
  stale)
- The threshold for "rebuild" (bounce rate above X% = rebuild)
- The 3-step action when a contact decays (pause, re-enrich,
  remove)
- The cost per contact to run this audit
- The 1 warning sign that says "your list is rotting" before
  the bounce rate spikes

Output as a checklist a VA can run in under 2 hours per month.

Pro tips:

List decay is the silent killer of cold outreach campaigns. Most agencies don’t notice until bounce rate spikes above 5%.
Run the audit on the 1st of every month. Make it a non-negotiable part of the operational calendar.
Clay + Apollo waterfall re-verification is the cheapest way to catch decay. ~$0.05-0.15 per contact.

Prompt 30: Run a deliverability audit on your sending infrastructure

Purpose: Email deliverability is the cornerstone of cold outreach, per Smartlead’s customer case studies. 70% inbox placement is the 2026 baseline for cold traffic, per Smartlead 2026 sales automation guide.

The prompt:

I am sending {X} cold emails per day from {Y} sending accounts
across {Z} domains. My current inbox placement rate is
{estimate or last test result}.

Run a deliverability audit. For each of the 9 areas below,
give me:
- The pass/fail test
- The tool to use (free + paid)
- The 1-line fix if I fail

The 9 areas:
1. SPF record
2. DKIM record
3. DMARC record
4. Custom tracking domain (CNAME)
5. Sending account warm-up status
6. Domain age and reputation
7. Bounce rate (under 2%)
8. Spam complaint rate (under 0.1%)
9. Unsubscribe rate (under 0.5%)

End with a prioritised fix list (1-2 hours per fix).

Pro tips:

70% inbox placement is the 2026 baseline for cold traffic. Below that, every other metric is irrelevant.
Most deliverability problems are infrastructure problems, not copy problems. Fix SPF/DKIM/DMARC first.
Free tools: MXToolbox, Mail-Tester, GlockApps. Paid: SmartDelivery, Inboxalytics.

Prompt 31: Forecast next quarter’s pipeline from current reply rates

Purpose: Translate reply rates into pipeline forecast so the client sees the math, not just the metric.

The prompt:

Here are my last 90 days of cold outreach data:

- Emails sent per week: {X}
- Positive reply rate: {Y%}
- Meeting booked rate (from positive replies): {Z%}
- Show rate: {W%}
- Opps created rate (from shows): {V%}
- Average ACV: ${A}

Forecast next quarter's pipeline assuming no change in inputs:
- Pipeline generated
- Closed-won revenue (use my historical close rate: {C%})

Then model 3 scenarios:
- Scenario A: I increase emails per week by 25% (what
  happens to pipeline?)
- Scenario B: I increase positive reply rate from {Y%} to
  {Y+1%} (what happens?)
- Scenario C: I increase meeting show rate from {W%} to
  {W+10%} (what happens?)

Identify the single highest-leverage lever for next quarter.

Pro tips:

Pipeline forecasts from cold outbound are directional, not precise. Use ranges, not point estimates.
Show rate is the most under-leveraged lever. A 10-point show rate improvement is worth 25% more emails per week at lower deliverability risk.
Share the forecast with the client. It’s the best argument for investing in deliverability, not volume.

Comparison TABLE: prompt categories vs. funnel stage vs. output

This table maps the 31 prompts to the funnel stage they serve. Use it as a checklist when building a new campaign.

#	Prompt	Category	Funnel Stage	Primary Output	Best For
1	Generate 10 bite-sized offers	Offer brainstorm	Pre-launch	Offer menu	New campaign
2	Refine offer to one sentence	Offer brainstorm	Pre-launch	Email-ready offer line	Every email
3	Map offer to trigger event	Offer brainstorm	Pre-launch	Trigger-led hook	Cold email step 1
4	Build offer matrix for sequence	Offer brainstorm	Pre-launch	4-7 step matrix	New sequence
5	Reverse-engineer competitor offers	Offer brainstorm	Pre-launch	Differentiation brief	Competitive market
6	Apollo list → ICP brief	ICP & data	Pre-launch	1-page ICP brief	List building
7	Day in the Life of buyer	ICP & data	Pre-launch	Buyer timeline	Send-time tuning
8	Clay enrichment waterfall	ICP & data	Pre-launch	Enrichment plan	Trigger data
9	Build negative ICP filter	ICP & data	Pre-launch	Exclusion checklist	List hygiene
10	Account snapshot for ABM	ICP & data	Pre-launch	3-sentence snapshot	High-ACV accounts
11	10 subject lines	Hook & opener	Step 1	Ranked subject lines	A/B testing
12	First 2 lines of email	Hook & opener	Step 1	Preview-pane opener	All cold emails
13	3-part AIDA opener	Hook & opener	Step 1	Long-form opener	LinkedIn InMail
14	Pattern interrupt subject	Hook & opener	Re-engagement	Fatigue-buster subject	60+ day lists
15	Localize opener	Hook & opener	Step 1	Translated opener	EMEA / APAC
16	Offer reveal paragraph	Body & offer	Step 1-2	3-sentence offer	All sequences
17	Social proof line	Body & offer	Step 1-2	1-sentence proof	All sequences
18	Low-friction CTA	Body & offer	Step 1-2	Yes/no question	All emails
19	Value-add follow-up	Body & offer	Step 3-4	New-info follow-up	Sequences
20	Breakup email	Body & offer	Step 6-7	Final-touch email	All sequences
21	LinkedIn connection note	CTA & multi-channel	LinkedIn	300-char note	Multichannel
22	LinkedIn InMail	CTA & multi-channel	LinkedIn	1,900-char InMail	High-value accounts
23	Cold call script	CTA & multi-channel	Phone	30-sec opener	Cold calling
24	SMS for warm leads	CTA & multi-channel	SMS	4-touch sequence	Meeting confirm
25	WhatsApp opener	CTA & multi-channel	WhatsApp	3 opener variations	EMEA / APAC
26	4-week campaign diagnosis	Measurement	Weekly	Bottleneck brief	Optimization
27	Reply classifier rules	Measurement	Setup	5-category classifier	Time savings
28	Weekly pipeline review	Measurement	Weekly	1-page client brief	Client reporting
29	List decay monitor	Measurement	Monthly	Audit workflow	Deliverability
30	Deliverability audit	Measurement	Monthly	9-area checklist	Infrastructure
31	Pipeline forecast	Measurement	Quarterly	3-scenario model	Client planning

A 14-day “first 1,000 leads” sprint

If you want to put these prompts to work in the next two weeks, here’s a 14-day sprint I’ve run for multiple clients. It is intentionally aggressive but realistic for a small outbound agency (2-4 people, 1 copywriter, 1 SDR, 1 closer).

Day 1-2: Build the offer menu. Run Prompt 1 four times for your top 1-2 ICPs. Pick the strongest 3 offers per ICP. Run Prompt 2 to compress each offer to one sentence. Total: 6 offers ready to deploy.

Day 3-4: Build the list. Pull 1,500 contacts from Apollo for ICP #1. Run Prompt 6 to generate the ICP brief. Apply the negative ICP filter from Prompt 9. Enrich the top 1,200 with the trigger signals from Prompt 8 in Clay. Total: 1,200 enriched contacts.

Day 5-6: Build the sequence. Run Prompt 4 to build the 4-7 step matrix. Run Prompt 11 to generate subject lines. Run Prompts 12, 16, 17, 18 to write the body of each step. Total: full email sequence ready.

Day 7: Build the multichannel layer. Run Prompt 21 to write LinkedIn connection notes. Run Prompt 23 to write a 30-second cold call opener. Run Prompt 19 to write the value-add follow-up.

Day 8: Tech setup. Connect Apollo, Clay, and Instantly or Smartlead. Run Prompt 30 to audit deliverability. Set up reply classification with Prompt 27. Configure Calendly for booking.

Day 9-10: Soft launch (200 leads). Send to 200 leads with the offer-first sequence. Manually review every reply for 5 days. Adjust subject lines and CTAs based on actual reply language.

Day 11-14: Scale to 1,000 leads. Add another 800 leads across ICP #1 and ICP #2. Layer in LinkedIn touches (Prompt 21) for the top 300 accounts. Place 50-100 cold calls using Prompt 23. Book meetings. Run Prompt 28 to generate the first weekly pipeline review.

By day 14, you should have 80-150 positive replies, 15-30 booked meetings, and a clean sequence that’s ready to scale to 1,000+ leads per week.

Common mistakes to avoid

I’ve watched dozens of agencies run cold outreach in 2026. Here are the 8 mistakes that kill reply rates fastest, and the prompt to fix each.

Leading with “I” or “We” in the first line. Fix with Prompt 12 - start with the trigger, not yourself.
Asking for a 30-minute call in step 1. Fix with Prompt 18 - use a yes/no CTA instead.
Sending 7 follow-ups that all say “just bumping this up.” Fix with Prompt 19 - each follow-up needs new information.
Skipping the deliverability audit. Fix with Prompt 30 - 70% inbox placement is the baseline.
Running the same sequence to a list for 90+ days. Fix with Prompt 29 - monthly list decay audit.
Using one offer for the whole sequence. Fix with Prompt 4 - plan 2-3 offers per campaign.
Ignoring negative ICPs. Fix with Prompt 9 - exclusion is as important as inclusion.
Blaming copy when the list is the problem. Fix with Prompt 6 - diagnose the list first, the copy second.

Final word

The 31 ChatGPT prompts for offer-first cold outreach above are the exact prompts I use to book meetings for B2B clients. They work because they force the offer to show up in the first 2 lines, the trigger to feel timely, and the CTA to feel easy to say yes to. Pair them with a tight Apollo + Clay + Instantly or Smartlead stack, a 4-7 step sequence, and a weekly iteration loop, and you will out-reply 90% of the cold traffic in your category in 2026.

If you steal nothing else, steal this: lead with the offer, ask a yes/no question, and never send a 7th email that says “just checking in.” That’s the whole game.

Now go book some meetings.

Reader disclosure & educational-purpose notice

This page is published by SuperFreshAI for general informational and educational purposes only. By reading it, you agree to the points below.

Editorial independence. All reviews, guides, and recommendations are written by our editorial team based on hands-on use. Some links on this site are affiliate links, and some articles are produced as partner content — both are always clearly labeled. Our editorial conclusions are never shaped by partners or affiliates.
Not professional advice. Nothing on this page constitutes legal, financial, medical, tax, or other professional advice. AI tools, pricing, and capabilities change quickly — always verify current information with the tool's official documentation before making a decision.
Educational purpose only. The content here is intended to help you learn about AI tools and workflows. It is not a guarantee of results, performance, fitness for a particular purpose, or suitability for your specific situation. Your results may vary.
No warranties. The site and its content are provided on an "as is" and "as available" basis. We make no warranties, express or implied, about accuracy, completeness, reliability, or availability. See our Terms and Privacy for the full legal terms.
Your responsibility. You are responsible for how you use the information on this page, including any decisions you make based on it. Always do your own research and consult a qualified professional when appropriate.
Affiliate & partner disclosure. When you click certain outbound links, we may earn a commission at no extra cost to you. When a piece of content is produced as partner content, it is labeled at the top of the page. See our Editorial Policy for the full standards we follow.

By continuing to read, you acknowledge that you have read and understood this notice.

31 ChatGPT prompts for outbound agencies to craft offer-first cold outreach scripts

Why “just checking in” is dead in 2026

The 5-step offer-first script anatomy

SECTION 1: Offer brainstorm prompts (Prompts 1-5)

Prompt 1: Generate 10 bite-sized offers for a specific ICP

Prompt 2: Refine one offer into a one-sentence promise

Prompt 3: Map the offer to a real trigger event

Prompt 4: Build an offer matrix for an entire 4-7 step sequence

Prompt 5: Reverse-engineer what your top-performing competitor is offering

SECTION 2: ICP & data prompts (Prompts 6-10)

Prompt 6: Turn an Apollo list export into a sharp ICP brief

Prompt 7: Pull the “Day in the Life” of the buyer

Prompt 8: Build a Clay enrichment waterfall for trigger signals

Prompt 9: Write a “negative ICP” filter

Prompt 10: Build a 1-paragraph account snapshot for ABM-style sequences

SECTION 3: Hook & opener prompts (Prompts 11-15)

Prompt 11: Generate 10 subject lines for a specific offer

Prompt 12: Write the first 2 lines of the email (the “above-the-fold” zone)

Prompt 13: Build a 3-part AIDA opener (for LinkedIn and longer emails)

Prompt 14: Write a “pattern interrupt” subject line for fatigued lists

Prompt 15: Localize the opener for non-English-speaking markets

SECTION 4: Body & offer reveal prompts (Prompts 16-20)

Prompt 16: Write the offer reveal paragraph (step 3 of the skeleton)

Prompt 17: Add the social-proof line (step 4 of the skeleton)

Prompt 18: Write the low-friction CTA (step 5 of the skeleton)

Prompt 19: Write a “value-add follow-up” body (step 3-4 of the sequence)

Prompt 20: Write a breakup email (final step of the sequence)

SECTION 5: CTA & multi-channel prompts (Prompts 21-25)

Prompt 21: Adapt an email into a LinkedIn connection note (under 300 characters)

Prompt 22: Write a LinkedIn InMail for higher-value accounts

Prompt 23: Write a cold call script (under 30 seconds for the opener)

Prompt 24: Write an SMS for warm-lead follow-up

Prompt 25: Write a WhatsApp opener for EMEA / APAC

SECTION 6: Measurement & iteration prompts (Prompts 26-31)

Prompt 26: Build a campaign diagnosis framework from 4 weeks of data

Prompt 27: Build a positive-reply classifier

Prompt 28: Write a “weekly pipeline review” brief

Prompt 29: Build a list decay monitor

Prompt 30: Run a deliverability audit on your sending infrastructure

Prompt 31: Forecast next quarter’s pipeline from current reply rates

Comparison TABLE: prompt categories vs. funnel stage vs. output

People Also Ask FAQ

1. What are the best ChatGPT prompts for offer-first cold outreach?

2. How do outbound agencies book meetings with cold email in 2026?

3. What’s a good cold email reply rate in 2026?

4. Are cold email agencies still profitable in 2026?

5. What is offer-first outreach?

6. How do you scale cold outreach to 1,000 leads per week?

7. What’s the best cold email tool stack for agencies in 2026?

8. How many follow-up emails should I send in a cold outreach sequence?

A 14-day “first 1,000 leads” sprint

Common mistakes to avoid

Final word

SuperFresh AI

43 ChatGPT prompts for non-native English speakers to polish interview answers

41 ChatGPT prompts for SaaS founders in San Francisco to map local partnership opportunities

How to Detect AI-Generated Content

What Is the Best AI Tool for Writing?

AI Newsletter Writing Guide

Get practical AI insights in your inbox