0Pre-Call Prep (10 minutes, non-negotiable)
Run these four prompts in ChatGPT, Perplexity, and Google AI Mode. Screenshot every result.
1. Best [their category] in [their city/region]
2. Who should I hire for [the specific service they sell]?
3. [Competitor name] vs alternatives
4. Is [Prospect Company] any good?
📸
The Three Assets You Now Have
The Absence - their name missing from results 1 and 2
The Substitute - whichever competitor is named instead
The Misrepresentation - whatever AI says about them in #4
Number 3 is the sharpest. If a competitor owns the "vs alternatives" answer, that competitor is being handed every comparison-stage buyer in the category.
You must know the substitute competitor's name before you dial. You use it in the first fifteen seconds. The call dies without it.
1The Cold Call
1.1 The Persuasion Framework
Adapted from Belfort's Straight Line system. Roughly two-thirds transfers to a B2B retainer sale.
The Three Tens
A prospect buys when three certainties are high: the product, you, and your company. Most cold calls die because the rep hammers certainty into the product while the other two sit at three.
On these calls, product certainty is nearly free - the screenshot does it. Your entire uphill battle is certainty in you, because you're a stranger asserting an unfamiliar problem exists.
The First Four Seconds
You get about four seconds to establish three things: you're sharp, you're enthusiastic, you're an authority worth listening to. Everything after is spent confirming or fighting that initial read.
The Four Tonalities You Need
| Tonality | Delivery | When to Use |
|---|
| Certainty | Flat, unhurried, no upward inflection | Stating facts, the hook |
| Mystery | Drop volume, slow down, make them lean in | The setup, building intrigue |
| Reasonableness | Warm, low-pressure, "I'm being fair with you" | Asking permission, the close |
| Declarative-as-Question | Implied question mark on a statement | Getting agreement without asking |
1.2 Gatekeeper
"Hi, I'm trying to reach whoever handles business development - not marketing, business development. Is that [Name]?"
If pushed for a reason:
"I ran a search that named [Competitor] instead of them. I want to show them the result. Thirty seconds is all I need."
Never say you're calling about marketing. That's a hangup.
1.3 The Script
Opener Sharp, upbeat, slightly fast
Reasonableness"[Name], [You] with Oction. This is a cold call - you can hang up on me. Give me thirty seconds and then decide?"
Naming the cold call is a certainty play: honest people announce themselves. Actually stop - the permission ask only works if you let them answer it.
Hook Mystery and intrigue
Mystery"This morning I asked ChatGPT who does [specific service] in [region]. It named three firms."
[Pause. Let it land.]
Certainty"You weren't one of them."
Absolute certainty. Flat. No apology, no upward inflection. Then shut up - whoever speaks first has ceded control.
The Reframe Reasonableness, then certainty
"The reason I bothered calling - your buyers are PMs and procurement building a shortlist before the RFP ever goes out. That's starting with an AI query now. So if you're not in the answer, you're not on the shortlist… and you never find out."
"No lost bid. No rejection letter. You just quietly stop getting invited."
"You never find out" is the pain amplifier. Deliver it slowly. Let them do the math.
Declarative-as-Question
Declarative"You'd want to know if that was happening, right?"
Almost nobody says no. You now have a yes on the record and permission to continue.
The Ask Reasonableness
"I'm not going to pitch you on the phone. I've got four screenshots of what AI says about your firm and where it's pulling it from."
"Fifteen minutes on a screen share. If it doesn't bother you when you see it, we're done, I won't call back. Sound fair enough?"
"Sound fair enough?" is the workhorse close. It asks for a small reasonable agreement, not a decision.
Booking
"Thursday at 10, or Friday at 2?"
Two options, both specific. Confirm the email on the call. Send the invite before you hang up.
1.4 Looping - Handling Objections
When an objection lands, don't rebut it. Deflect, loop back up the line raising certainty on whichever ten is low, re-close at a slightly higher level. Three loops is the ceiling.
Loop 1 - Deflect and Reframe
"I hear you, and that makes sense. Let me ask you this though -"
Mystery"Do you know what ChatGPT currently says about [Firm]?"
They don't. That's your opening.
"That's exactly it. Nobody does. That's the whole problem - this is the one channel where you're being described to buyers, and nobody at your firm has ever read the description."
"Fifteen minutes Thursday, does that work?"
Loop 2 - Raise Certainty in You
"I'll be straight - I do this specifically for [vertical] firms, because your buyers research completely differently than a retail business does."
"I've run this exact search for [X] firms in [region]. The pattern is identical every time: three names come up, and it's the same three."
Loop 3 - The Take-Away
"Honestly, it might not be for you. Some firms have enough repeat work that they genuinely don't care about the shortlist. Is that you?"
Strongest move in the script - but only when it's real. It works only if you'd actually walk. If they hear a technique, you've lost.
1.5 Objections - In-Call
"We get all our work through relationships / referrals"
"I believe you - that's how the industry works. But the referral gets checked now. Someone gets your name, types it in to see if you're legit, and reads whatever comes back."
"Have you seen what comes back? That's really what I want to show you."
This one is usually sincere. Answer it straight - don't loop it as resistance.
"Send me some information"
"I will, but the information is four screenshots that need about two minutes of context or they're just pictures. Fifteen minutes Thursday and I'll send them either way."
"We don't do marketing"
"This isn't marketing spend, it's closer to a listing correction. Right now there's a wrong answer about your firm being given to buyers and nobody's fixing it."
"Fifteen minutes to look at it."
"Not interested"
"Understood. Can I ask one thing before I go - do you know what ChatGPT currently says about [Firm]?"
[Pause for response]
"Then let me just email you the screenshots. No call, no follow-up. You should see it regardless of whether we ever work together."
Highest-value line in the script. You lose the meeting, keep the relationship, and roughly a third reply to the email.
2Email
2.1 Cold Email
Subject Lines Test all four
ChatGPT recommended [Competitor] instead of you
what AI says about [Company]
you're not in the answer
screenshots inside - [Company] + AI search
Body
[Name] -
I asked ChatGPT who the best [category] in [region] is. It named [Competitor A] and [Competitor B]. Not [Company].
Then I asked what it knew about [Company] specifically. It [described you as X / had no information / pulled from a 2023 directory listing].
That's not a ranking problem. AI doesn't give ten results - it gives one recommendation, and the buyer takes it. Whoever owns that answer owns the category.
Screenshots attached.
If you want, I'll walk you through where the AI is actually sourcing its answer and what it would take to be in it. Fifteen minutes?
[You]
Follow-Up 1 Day 4 - Send a new screenshot, not a bump
[Name] -
Ran the comparison query this time: "[Competitor] vs alternatives."
You're not listed as an alternative. That's every comparison-stage buyer in your market going somewhere else.
Screenshot attached.
Follow-Up 2 Day 9 - Breakup
[Name],
Closing the loop. Leaving you the four screenshots either way - even if you never work with me, you should know what AI is telling your buyers.
Good luck.
The breakup that gives away the asset converts better than the one that asks again.
3Discovery Call
3.1 The 15-Minute Structure
| Time | Phase | What You Do |
|---|
| 0–2 min | Frame | "I'm going to share my screen and run four searches live. You tell me what you see. Then I'll tell you what's causing it. Sound fair?" |
| 2–7 min | The Demo | Run the queries LIVE. Not the screenshots. Then shut up. The only question: "What's your reaction to that?" |
| 7–11 min | The Diagnosis | Show what AI is citing. "Notice what's not in that list - your website." |
| 11–15 min | The Ask | Present two options: Audit or Program. |
The Diagnosis (7–11 min)
"Notice what's not in that list - your website. AI systems aren't reading your site the way Google did. They're assembling an answer out of third-party sources, and you have no presence in any of them."
"That's the actual problem. It's not your website. It's everywhere your website isn't."
The Ask (11–15 min)
"Two options."
"One, I run a full audit - map every prompt a buyer in your category would use, score where you appear, show you exactly which sources are feeding the answers. [$X], fixed, two weeks."
"Two, we skip that and go straight into the program."
"Most people want the audit first. That's fine - it credits toward the program if you go ahead."
3.2 Objections - Post-Demo
"How is this different from SEO?"
"SEO gets you on a list of ten. This gets you to be the answer. SEO is about ranking pages; this is about being the source the model trusts enough to cite. Some fundamentals overlap. The distribution work doesn't at all."
"Isn't this a fad? Won't AI change next month?"
"The models change monthly. What doesn't change is what they're built on - being mentioned in credible places your buyers' questions map to."
"That was true for Google and it's true for ChatGPT. If AI search vanished tomorrow you'd still have the citations and the content. There's no version where you've wasted the money."
"We already have an SEO agency"
"Good - keep them. Ask them one question: what's our citation share in ChatGPT and Perplexity, and what's the trend?"
"If they've got an answer, you're covered and I'll go away. If they change the subject, that's the gap I fill. I work alongside SEO teams more often than I replace them."
"How do you measure this?"
"Citation share - of the prompts your buyers actually use, what percentage return your name."
"I baseline it before we start and report monthly against your three closest competitors."
"I'll tell you now: the first sixty days is mostly baseline accumulation. Anyone promising movement in week two is selling you a dashboard."
"That's expensive"
"Compared to what? What's one [project] worth to you?"
[Pause for number]
"Then this pays for itself at [X] per year. And the alternative isn't zero - the alternative is [Competitor] getting handed those buyers for free."
"Can you guarantee we'll show up?"
"No, and anyone who does is lying to you - nobody controls the model."
"What I guarantee is the inputs: citations placed, content restructured, schema deployed, third-party sources built. And I'll show you citation-share movement monthly."
"If it's flat at ninety days, we have an honest conversation about whether to continue."
4Pricing
Anchor high, then present the tier you want.
"Full-service programs in this space run three to twenty-five thousand a month depending on scope. Enterprise is well north of that. Here's where I'd put you."
Tripwire
Tier 1 - Audit
$1,500–$5,000
One-time. Prompt map, citation baseline, source analysis, prioritized fix list. Credits toward Tier 2.
Standard
Tier 2 - Foundations
$3,000–$5,000/month
Entity optimization, schema, content restructuring, tracking, monthly reporting.
Growth
Tier 3 - Growth
$6,000–$12,000/month
Everything above plus third-party citation placement, comparison content, digital PR, competitive citation analysis.
The Add-On Play (easiest money): If they already pay for SEO, price GEO/AEO as a 20–30% uplift on the existing retainer. A $5,000/month SEO client becomes $1,000–$1,500 of new revenue with almost no sales friction.
Positioning Line
"Most of what's sold as GEO right now is a monitoring dashboard with a report attached. You're paying to watch the problem. Measurement is table stakes - the work that earns the citation is what you're actually buying."