Cold Call Script Generator

Generate personalized cold call openings, voicemails, and cold email opening lines for B2B sales motions. Combines real prospect research, ranked pain hypotheses, pattern-interrupt opening frameworks, and a full objection-handling library into ready-to-use scripts and 14-day multi-touch sequences.

Usage

Invoke this skill when you have a list of prospects (or one named account) and need a script that actually books meetings instead of getting hung up on.

Basic invocation:

Write me a cold call script for selling [product] to [ICP] Draft a voicemail for a CTO at a Series B fintech Build a 14-day outbound sequence for [target persona]

With context:

Here's the prospect's LinkedIn and the company's funding announcement, write the opener They already use Salesforce — write me an opener that doesn't sound like every other displacement pitch I'm a founder selling to heads of product, write three openers in different tones

The agent produces opener variations, full call scripts with branches, voicemail templates, objection rebuttals, and a touch-by-touch cadence — calibrated to the ICP's seniority, function, and likely buying stance.

How It Works

Step 1: Prospect Research Checklist

Before writing a single line, the agent gathers (or asks for) signal. Cold scripts that reference nothing specific feel like spam — the goal is one concrete reference point per call.

The 9-point research checklist:

Minimum viable research: name, title, company, one specific signal. Anything less and the call is a coin flip.

Step 2: Pain Hypothesis Generation

For each ICP, the agent generates exactly three pain hypotheses, ranked by signal strength.

Signal strength tiers:

• A-tier (named in public): prospect or company has publicly stated this pain (job post, podcast, earnings call). Use as a direct quote.
• B-tier (inferred from stage / stack): logical given their growth stage or known tools. State as observation, not fact.
• C-tier (category-default): the pain everyone in their ICP has. Use only when A and B are unavailable.

Example — ICP: VP Engineering at 200-person Series C SaaS:

Hypothesis 1 (A-tier): "You posted about on-call burnout three weeks ago on LinkedIn"
  Signal: direct public statement
  Opener angle: empathy + specific tool fix
  Confidence: 95%

Hypothesis 2 (B-tier): "Series C scale typically breaks the homegrown CI pipeline around now"
  Signal: stage-pattern inference
  Opener angle: pattern-recognition expertise
  Confidence: 60%

Hypothesis 3 (C-tier): "Engineering productivity metrics are probably hard to defend in board meetings"
  Signal: category-default for the role
  Opener angle: peer-benchmark conversation
  Confidence: 30%

The agent always offers all three and recommends starting with A-tier; B-tier is fallback when A signals are absent.

Step 3: Opening Line Framework — Pattern Interrupt > Permission > Value Tease

The opener has three jobs in 8 seconds:

1. Pattern interrupt — break the "this is a cold call" auto-script in their head
2. Permission — give them an off-ramp so they don't feel trapped (paradoxically increases continuation rate)
3. Value tease — tell them why staying on the call is worth their next 30 seconds

Eight tested opener variations:

1. The Honest Cold Call

"Hi [Name], this is [You] — I'll be transparent, this is a cold call. Can I have 27 seconds to tell you why I picked up the phone, and then you can tell me to go away?"

Works because: explicit honesty pattern-interrupts the "is this a sales call?" frame. The specific number ("27 seconds") signals you've thought about respecting their time.

2. The Permission-First

"Hey [Name], it's [You] from [Company]. Did I catch you at a terrible time?"

Works because: prospects expect "good time" — flipping it earns a "no, what's up?" 70% of the time (per Gong data on 100K+ calls).

3. The Reference-Their-World

"Hi [Name], I saw your post on [specific topic] — that's actually why I called. Got 30 seconds?"

Works because: signals you did homework before the call, not after they answer.

4. The Peer-Drop

"Hi [Name], I just got off a call with [peer at similar company] — they had this exact problem with [pain]. Wanted to ask if it's on your radar too."

Works because: lateral peer pressure is the strongest B2B social proof. Only use if the peer reference is real and verifiable.

5. The Disqualification Opener

"Hi [Name], I'm not sure if this is relevant to you — we work with engineering teams using [stack]. Are you running [stack] today, or have you moved off?"

Works because: framing as "not sure if relevant" lowers their guard. Asking about stack gives them an easy first answer.

6. The Provocative Insight

"Hi [Name] — quick question: how are you measuring [specific metric] today? Most VPs of [function] we talk to are flying blind on this and didn't realize it until they saw what we were measuring."

Works because: implies they might have a blind spot without insulting them. Triggers curiosity.

7. The Trigger-Event Opener

"Hi [Name], I saw [Company] just [funding/launch/hire] — congrats. The reason I'm calling: most companies hitting this stage start running into [pain] within 90 days. Wanted to check if you're seeing that."

Works because: ties the call to a specific event in their life. Creates relevance and time-pressure framing.

8. The Founder-To-Founder (founder-led sales only)

"Hi [Name], this is [You], I'm the founder of [Company] — calling personally because [specific reason]. I'll keep it under 90 seconds. Worth hearing?"

Works because: founder-on-founder asymmetry — heads of product / VPs accept founder calls they'd reject from SDRs. Only authentic when actually the founder.

Opener anti-patterns (the agent flags these):

• "How are you today?" — instant rapport-fail signal, prospects hang up
• "Did I catch you at a good time?" — primes a "no"
• "I'm calling to introduce..." — sounds like a script, is a script
• "We help companies like yours..." — every cold call says this
• Reading a 30-second pitch before letting them speak — guarantees hangup at second 12

Step 4: Voicemail Script — The 17-Second Structure

80%+ of cold calls go to voicemail. A great voicemail gets a callback rate of 3-8%; a bad one is 0%. Target length: 17 seconds, no more.

The 17-second structure:

[0-3s]   Name + company:    "Hi [Name], this is [You] from [Company]"
[3-8s]   Specific reason:   "Calling because I saw [specific signal]"
[8-13s]  Value tease:       "We help [peer companies] with [specific outcome]"
[13-15s] Callback ask:      "Worth a 15-min look?"
[15-17s] Number, slowly:    "[Number], again [Number]"

Example voicemail (Series B fintech CTO, hiring SREs):

"Hi Maria, this is Jordan from Sentinel. I saw you're hiring three senior SREs — most fintechs hitting that scale waste six months on alerting noise before they fix the root cause. We helped Plaid cut their on-call pages by 60% in 90 days. Worth a 15-minute look? 415-555-0142, again 415-555-0142."

Word count: 47. Spoken at normal pace: ~17 seconds.

Voicemail rules the agent enforces:

• Never apologize ("sorry to bother you")
• Always say the phone number twice, slowly, at the end (not the start)
• One specific reference, never two — adding more dilutes the hook
• Don't say "I'll try you again" — it removes their reason to call back
• Match the tone of the day (Monday 8am ≠ Friday 4pm — see tone calibration below)

Step 5: Objection Handling Library

The agent provides a scripted response for every common objection, with the underlying principle named so the rep can adapt.

"Not interested."

Principle: they're not rejecting you, they're rejecting the unknown. Acknowledge, narrow the claim, redirect to a question.

"Totally fair — you don't know me yet. Quick question, just so I don't waste your next call: are you handling [specific pain] in-house today, or have you outsourced it? If neither is a fit I'll get out of your way."

"Send me an email."

Principle: this is the polite hangup. Send the email AND keep the conversation going.

"Happy to — what's the best email? While I have you, just so I send something actually relevant and not a generic deck: are you more focused on [option A] or [option B] right now?"

Then: send the email within 5 minutes referencing the conversation. Subject line: "[Name] — as discussed, the [option they picked] piece"

"We already use [competitor]."

Principle: don't trash the competitor. Compliment-then-pivot.

"[Competitor] is a solid product — half our customers came from there. Most of them switched because of [specific gap]. Are you running into that, or has [Competitor] been working well on that front?"

Anti-pattern: "Oh, [Competitor] is terrible because..." — instantly destroys trust. The prospect picked Competitor; insulting it insults them.

"No budget."

Principle: budget objections are usually priority objections in disguise. Reframe.

"Got it — budget's tight everywhere right now. Just so I understand: if we could show that [outcome] would [save X / earn Y] within [timeframe], is this the kind of thing you'd find budget for, or is it genuinely not on the priority list this year?"

This separates "no money" (call back next quarter) from "not a priority" (disqualify).

"Call me back later / next quarter."

Principle: vague timing = soft brush-off. Pin down a specific reason or move on.

"Happy to. Just so the next call is useful — what would change between now and then? Is it [budget cycle / project finishing / hiring this role]? That way I can come back with something tied to whatever's actually shifting."

If they can't name a specific change, the timing is fake — politely disqualify and don't waste a follow-up.

"How did you get my number?" (hostile)

Principle: be honest, brief, and don't grovel.

"Public profile — your direct line was on [source]. Happy to remove you from our list if you'd prefer, or I can take 30 seconds to tell you why I called and you can decide?"

"Just put me on your do-not-call list."

Principle: comply immediately, do not pitch. This is a TCPA / GDPR moment. Confirm and close.

"Done — you're off the list now. Sorry for the interruption."

Then add to DNC list immediately (see Compliance section).

Step 6: Multi-Touch Sequence Design — 14-Day Cadence

A cold call is not one event; it's the second or fourth touch in a sequence. Connect rates jump 3-5x when calls follow LinkedIn engagement and email warmup.

The 14-day, 11-touch sequence:

Cadence design principles the agent applies:

• Channel diversity per touch: never two calls in a row without email/LinkedIn between
• Time-of-day rotation: call at 8am, 11am, 4pm across attempts (most reps only call at 10-12)
• Different angle per email: email #1 = pain, email #2 = peer story, email #3 = breakup
• Real engagement on LinkedIn: comment substantively on their post, not "great insights!"
• No more than 11 touches in 14 days: beyond this is harassment and hurts reply rates
• The breakup email outperforms all others: "I'll close the loop on this — let me know if I should circle back next quarter" generates the highest reply rate of the sequence

Step 7: Tone Calibration by ICP

The same words land differently with different buyers. The agent tunes vocabulary, pace, and stance per persona.

By function:

Calibration examples:

CTO opener: "Most platform teams at your scale are seeing p99 drift past 800ms on the deploy hot path — is that landing on your plate?"

CMO opener: "We just helped [peer brand] move from 'attribution is broken' to closing the loop on first-party data inside 60 days — wanted to ask if that story is one you're trying to tell to your CEO."

CFO opener: "I'll be quick — every dollar in [category] tooling at companies your size has a measurable payback window. We average 4.2 months. Worth 15 minutes to see if your numbers would line up?"

Founder opener: "Hey [Name], founder-to-founder — I won't waste your time. We do [one sentence]. Specifically for early-stage teams like yours. 10 minutes this week worth it, or not yet?"

Step 8: Compliance Essentials

Cold outreach is regulated. The agent always includes the relevant compliance reminders for the rep's region and prospect region.

TCPA (US — Telephone Consumer Protection Act):

• Calling mobile phones with auto-dialers requires prior express written consent for B2C; B2B calls have looser rules but vary by state
• Calling hours: 8am-9pm in the prospect's local time zone
• Maintain a DNC (do-not-call) list — honor requests within 30 days (in practice: same-day)
• Florida and Oklahoma have stricter "mini-TCPA" laws — one-call-can-cost-$500-per-violation
• Never use prerecorded messages for B2B cold without consent

GDPR (EU/UK prospects):

• B2B cold calling is allowed under "legitimate interest" basis, but the prospect must be relevant to their job role
• Document your legitimate-interest assessment (LIA) before calling
• On request, delete personal data within 30 days
• Cold email to EU prospects: include sender identity, business address, and one-click unsubscribe — even on a "first email"
• Right to object is absolute — when a prospect says "stop," stop on all channels

CAN-SPAM (US email):

• Sender identity must be accurate (no fake "from" names)
• Subject line cannot mislead
• Physical mailing address required in every commercial email
• Honor unsubscribes within 10 business days
• "Cold email is illegal" is a myth — it's legal under CAN-SPAM if you follow the rules

Universal compliance hygiene the agent recommends:

• Maintain a single DNC list across all channels (call, email, LinkedIn)
• When a prospect opts out on one channel, suppress on all
• Keep call recordings only with two-party-consent disclosure where required (CA, FL, IL, MA, MD, MT, NH, PA, WA)
• Never spoof caller ID (illegal under TRACED Act)

Step 9: Measurement and Benchmarks

The agent ends every script delivery with the metrics to watch and what "good" looks like.

The four cold-call metrics:

Cold email benchmarks (B2B, well-targeted):

• Open rate: 40-60% (anything below 30% = subject line problem)
• Reply rate: 3-8% (positive replies); 1-3% positive replies is the median for most teams
• Meeting-booked rate: 1-3% of total emails sent

Sequence-level benchmark:

• Total reply rate across an 11-touch sequence: 8-15% positive
• Meeting-booked rate per prospect entering the sequence: 3-6%

Diagnostic logic the agent uses:

Symptom: Connect rate < 8%
  Likely causes: bad list, wrong time of day, spam-flagged number, ICP mismatch
  Fix: list hygiene, rotate caller ID, test 7-9am and 4-6pm windows

Symptom: Connect > 15% but conversation rate < 30%
  Likely cause: opener is failing the pattern interrupt
  Fix: A/B test opener variations, record and review hangup points

Symptom: Conversation > 60% but meeting-set < 15%
  Likely cause: weak value tease or no clear next step
  Fix: rewrite the close, add specific calendar offer, reduce friction

Symptom: Meeting-set healthy but held rate < 60%
  Likely cause: weak confirmation cadence, prospect not bought in
  Fix: add same-day calendar invite + day-before reminder + value pre-read

Sample Scripts

Sample 1: SaaS Sales Rep Calling a CTO

Context: AE at observability platform, calling Maria Chen, CTO at a 250-person Series B fintech. Research signals: hiring 3 SREs (LinkedIn job post 2 weeks old), recent post about "alert fatigue is killing my team", company raised $40M Series B four months ago.

The call:

Rep: "Hi Maria, this is Jordan from Sentinel. I'll be straight — this is a cold call. Saw your post on alert fatigue two weeks back, that's actually why I picked up the phone. Got 30 seconds?"

Maria: "Uh, sure, go ahead."

Rep: "Appreciate it. Most fintechs at Series B with the SRE hiring you're doing tell us they spend the first 90 days drowning in noise from the legacy alerting setup before they fix it. We helped Plaid cut on-call pages 60% in their first quarter on us. Curious if alert noise is on your near-term list, or if you've already got someone solving it?"

Maria: "We're using Datadog and PagerDuty — it's working, but yeah, the noise is real."

Rep: "That's the most common stack we replace alongside, not against. We sit on top of Datadog and dedupe before PagerDuty fires. Worth 15 minutes next week to show you what that looks like with your alert volumes? I can do Tuesday at 11 or Thursday at 2."

Maria: "Thursday 2 works."

Rep: "Great — sending the invite now with two questions for you to skim beforehand so we don't waste your time. Talk Thursday."

Why it works: A-tier signal (her own post), CTO-tone (specific tools, peer-named), permission opener, complement-then-pivot on Datadog, two specific times instead of "what works for you?"

Sample 2: Recruiter Calling an Engineering Manager

Context: Recruiter from a boutique exec search firm, calling David Park, Engineering Manager at a mid-stage AI infra startup. Research signals: 3.5 years tenure (long for the role), no recent LinkedIn activity, his old VP just left for a competitor.

The call:

Recruiter: "Hi David, this is Sam from Northpoint Search. Quick one — did I catch you at a terrible time?"

David: "I have a couple minutes, what's up?"

Recruiter: "I'll be transparent, I'm a recruiter — but I'm not pitching a job today. I work with senior eng managers in AI infra and I'm doing a market read. You've been at [Company] three and a half years which is impressive, especially given some of the recent moves on the team. Two questions, then I'll get out of your way: one, are you in build mode or coast mode right now? Two, if I called you in six months with the right director-level role, would that be useful or annoying?"

David: "Honestly, I'm in coast mode. Things have been weird since [VP] left."

Recruiter: "Heard. That's the third call this month I've had like that — there's a lot of motion in AI infra right now. Look, I won't push anything on you today. Can I send you one or two role specs over the next month, just for calibration, no commitment?"

David: "Yeah, that's fine."

Recruiter: "What's the best personal email — I don't want anything tied to your work account."

Why it works: disarms the recruiter-call defenses immediately ("I'm not pitching a job today"), references the VP departure as a real signal, low-commitment ask (just calibration emails), gets the personal email which is the real prize for a recruiter.

Sample 3: Founder Doing Founder-Led Sales to a Head of Product

Context: Founder/CEO of a 12-person product analytics startup, calling Priya Rao, Head of Product at a Series C consumer app. Research signals: she gave a podcast interview 3 weeks ago about "product teams flying blind on activation metrics", company has 4M MAU.

The call:

Founder: "Hi Priya, this is Alex Kim, I'm the founder of Lumen Analytics. Calling personally — I heard your podcast with [Host] three weeks back and I literally pulled over to take notes when you described activation metrics as 'a Rorschach test'. That's the reason I picked up the phone today. I'll keep it under two minutes — worth hearing?"

Priya: "[Laughs] Sure, go for it."

Founder: "Thanks. We built Lumen specifically because every product team I worked with before this had the same problem you described — every PM had a different definition of activation, and the data team couldn't agree with any of them. We work with about 40 product teams now, and the wedge is dead simple: we let your PMs propose a definition, your data team approves it, and then everyone's looking at the same number forever. The reason I'm calling you specifically: with 4 million MAU and the segmentation you talked about, this is exactly the scale where the disagreement gets expensive. Two questions: one, does that resonate at all? Two, would 25 minutes with me — not a sales rep, me — be useful next week?"

Priya: "Yeah, that's a real problem for us. Send me a calendar link, I'll find time."

Founder: "Will do — and one favor: when you book it, tell me the one definition of activation your team most disagrees on. I'll come ready to talk through how we'd resolve it. Talk soon."

Why it works: founder-to-Head-of-Product asymmetry, A-tier signal quoted directly, specific stage relevance (4M MAU), real two-minute promise, asymmetric-value ask (25 min with the founder, not a rep), homework request that primes the demo to be customized.

Output

The agent produces:

• Prospect research summary: 9-point checklist filled in, with signal strength noted
• Three ranked pain hypotheses with confidence percentages
• Three to five opener variations chosen from the 8-pattern library, customized
• Full call script with branches for the most likely first response
• 17-second voicemail customized to the signal
• Objection handling cheatsheet for the top 5 objections this prospect is likely to raise
• 14-day, 11-touch sequence with channel, day, and per-touch message
• Tone calibration notes specific to the persona
• Compliance checklist for the relevant region (TCPA / GDPR / CAN-SPAM)
• Metrics targets so the rep knows what success looks like

Common Scenarios

"I have a list of 50 prospects, write me a sequence"

Provide the list with title, company, and any signals. The agent generates a templated sequence with per-prospect personalization tokens.

"My connect rate is 4%, what's wrong?"

The agent runs the diagnostic logic in Step 9, asks about list source, dialing windows, and caller ID setup, and recommends the highest-leverage fix.

"Write me an opener I can A/B test against my current one"

Provide your current opener. The agent diagnoses what's working and produces 2-3 alternatives from different patterns to test.

"I'm a founder, I've never done sales — give me a script I won't sound stupid reading"

The agent uses the founder-to-buyer pattern, keeps language plain, builds in natural pauses, and writes for spoken delivery (short sentences, no jargon).

"We just got told to stop calling them, what do I do?"

Stop. Add to DNC immediately on all channels. The agent walks through the suppression checklist and the polite "closing the loop" email if appropriate.

Tips for Best Results

• Provide one specific signal per prospect — even a single LinkedIn post beats no research
• Tell the agent what tone you want: "peer-technical" vs "warm-conversational" produces very different scripts
• Share what's worked and not worked — the agent will steer away from patterns you've already exhausted
• Mention your current connect/conversation/meeting rates so the agent can target the weakest link
• For founder-led sales, say so explicitly — the asymmetry changes the whole opener pattern
• Specify region (US / EU / UK / APAC) so compliance guidance is correct
• If you have a recorded call that flopped, share the transcript — the agent can pinpoint the exact second the prospect disengaged

When NOT to use

• Inbound leads who already requested a demo — they want a discovery call, not a cold opener; the pattern interrupt is wrong for warm intent.
• Existing customers — use a customer success or expansion playbook, not cold-call patterns.
• Highly regulated outreach (healthcare, financial advice, political) — these need legal review beyond TCPA/GDPR basics; consult counsel.
• B2C cold calling to consumer mobile numbers in the US — TCPA exposure is severe; this skill is built for B2B.
• Markets where cold calling is culturally inappropriate (e.g., much of Japan, Germany for senior execs) — use referral or warm-intro motion instead.
• When you have no research signal at all — generic scripts produce generic results; do the 9-point research first or expect 1-2% connect-to-meeting rates.
• High-ACV enterprise deals (>$500K) — cold calling can open the door, but those motions need ABM, multi-threading, and exec-sponsor engagement well beyond a script.
• When the prospect has already opted out — adding them back to a sequence is a compliance violation regardless of how good the script is.