The story How we approached it What it took The sales motion What I learned The workflow Do it yourself

BUILD LOG · BE THE ANSWER THE MACHINES QUOTE

Your buyers stopped searching the way you write for. They ask an answer engine now, and it either quotes you or it does not.

AEO and GEO are not SEO with new keywords. The machine answering your buyer only repeats what it can source, in the buyer's own words. So the strategy starts where it always should have: understanding the customer well enough that the answer engine has something true, and yours, to quote.

Steal this build ↓See the stack The people part →

7 figures

of churned win-back the same backward read surfaced

25% to 100%

share of the prompt budget working the positioning, was 25%

108 to 0

undifferentiated utility prompts, eliminated

2 to 56

Define prompts, the category ground nobody owned

THE STORY, FROM THE SEAT

Here is one I actually shipped. What happened, what I did, and the stack if you want to run it yourself.

Most positioning still writes for the search box and the human skimming a results page. But buyers changed how they look for you. They ask an answer engine, and it returns one synthesized answer that either includes you or does not.

If your positioning is not built to be quoted by the machine answering your buyer, you are not ranked low. You are absent, at the exact moment of intent. And you cannot keyword your way out of it: the engine repeats what it can source in the buyer's own language, so the work runs backward, from understanding the customer to writing the thing the machine will quote.

Keep Building,

Heath

FOUNDER, THE SALES OPERATOR

THE BUILD IN FIVE MOVES · THE GIANT METHOD

GBuyers ask an answer engine now and get one synthesized answer that either includes you or does not.
IGong transcripts for the buyer's language, Salesforce attribution read backward from the win, Claude for the prompt research and the rewrite.
AAI pulls the buyer's language and drafts the quotable positioning; the operator sets the prompt budget and locks the canonical story.
NThe positioning became a rep talk track, rehearsed until it sounded natural, so the human channel matched the machine one.
TThe backward read surfaced a churned win-back pool north of two million dollars, and Define prompts went from two to fifty-six.

RUN THE PLAY YOURSELFThis exact build is a playbook. Run it on your own data.

→

01 / HOW WE APPROACHED THE PROBLEM

Ran the loop.

The build started from a change in buyer behavior: buyers stopped searching the way the positioning was written for, and started asking an answer engine that returns one synthesized answer. If the machine cannot quote you, you are not ranked low, you are absent at the exact moment of intent. And you cannot keyword your way out, because the engine only repeats what it can source in the buyer's own language. So the work ran backward, from understanding the customer to writing the thing the machine will quote.

1
Start with the customer, not the keyword
The same inputs that build a real ICP, customer interviews, G2 reviews, call transcripts, and closed-won patterns, are what tell you the language buyers actually use. AEO is downstream of understanding the customer; you cannot be quoted on words your buyer does not say.
2
Read attribution backward
I ran the research on how customers actually find us, why they buy, and why they stay, reading the demo and attribution data from the win back to the origin instead of crediting the last click.
3
Rewrite positioning to be quoted, not ranked
I rewrote the positioning for how buyers search now, so the answer engine repeats it. Written for the answer, not just the page, in the customer's language rather than the category's.
4
Reallocate the prompt budget, quietly
Same prompt budget, reallocated from undifferentiated utility to three jobs: Defend the terms we own, Intercept the buyer's real question, and Define the category on our terms. And name the AI only where it earns the line, because leading with it makes you sound like every other tool.
5
Version the story like code
Positioning drifts the moment it lives in three places. So the canonical story got locked and versioned like software, one source the site, the listing, and the sales pitch all pull from, instead of three eras of the message fighting each other.

02 / WHAT IT TOOK CROSS-FUNCTIONALLY

Nobody ships this alone. Here is who had to move.

Getting cited by answer engines is a positioning decision executed in content and code. Neither lever sat inside sales.

MARKETING

The content and the positioning the engines read.

The prompt budget reallocation only works if the pages actually say the differentiated thing. The positioning shift and the content that carried it were marketing's build.

WEB AND ENGINEERING

Shipping the page changes.

Answer engines read what is live, not what is planned. The structural and copy changes had to ship on the site for any of the measurement to move.

03 / HOW WE TURNED IT INTO A SALES MOTION · THE PEOPLE PART

One story, everywhere the machines read. The team still has to tell it the same way.

Getting quoted by an answer engine means saying the same thing everywhere, and that includes the reps. The leadership half was turning the positioning into a talk track every rep uses, so the human channel matches the machine one.

THE FRAMEWORK BEHIND IT

SMART Prospecting on one narrative

The positioning defines the story; SMART carries it into the outreach. Show you understand the buyer, map the problem in the same words the site and the answer engines use, take action, so a prospect hears one consistent story whether they searched, asked an AI, or got an email.

01
Made the positioning a rep talk track
The same story the site tells, in the rep's own words, so nothing got lost between the page and the call.
02
Killed the freelance pitch
Coached the team off improvised messaging and onto the one narrative that was actually working.
03
Rehearsed the story until it was natural
Reps practiced the narrative so it sounded like them, not a script they were reading.

Carry the narrative into outreach →

04 / WHAT I LEARNED

Not proof. Just what the build taught me.

01
You cannot keyword your way out
The engine repeats what it can source in the buyer's own language. AEO is downstream of understanding the customer; you cannot be quoted on words your buyer does not say.
02
Absent is not the same as ranked low
The answer engine returns one synthesized answer that either includes you or does not. Missing it means missing the exact moment of intent.
03
Name the AI only where it earns the line
Leading with the AI makes you sound like every other tool. The prompt budget went to Defend, Intercept, and Define instead of undifferentiated utility.
04
Positioning drifts when it lives in three places
The canonical story got locked and versioned like software: one source the site, the listing, and the pitch all pull from, instead of three eras of the message fighting each other.

05 / THE WORKFLOW

The runnable version. Copy it into your stack.

ClaudeThe answer-engine prompt research and the positioning rewrite.

SalesforceThe demo and attribution data, read backward from the win.

GongCall transcripts for the buyer's own language.

NotionThe canonical positioning story, locked and versioned in one source.

[ SALES OPERATOR ]

Getting Found in the AI World

Heath Barnett · Sales Operator

G / Groundthe problem that started all of it, with the receipt and the cost

AEO

You are absent at the moment of intent.

Buyers ask an answer engine and get one synthesized answer that either includes you or does not.

Drag

You cannot keyword your way out.

The engine repeats what it can source in the buyer's own language, so the work runs backward from understanding the customer.

A / Assignthe build, block by block, and who owns each one. Open a step to see it run.

HUMAN + AI, IN THE LOOP

AI pulls the buyer's language, reads attribution backward, and drafts the quotable positioning; the operator sets the prompt budget and locks the canonical story.

Start with the customer

Gong01

Pull the buyer's own language

Details

Call transcripts. You cannot be quoted on words your buyer does not say.

Salesforce02

Read attribution backward from the win

Details

How customers actually find us, back to the origin, not the last click.

Write to be quoted

Claude03

Research the answer-engine prompts

Details

What the machine will source and repeat, in the buyer's language rather than the category's.

The operator04

Reallocate the prompt budget

Details

Same budget, three jobs: Defend the terms we own, Intercept the buyer's question, Define the category. Name the AI only where it earns the line.

Human

Lock it

Claude05

Rewrite positioning for the answer, not the page

Details

Written to be quoted, not ranked, in the customer's language.

AI + human

Notion06

Version the story like code

Details

One canonical source the site, the listing, and the pitch all pull from, instead of three eras fighting.

Human

N / Normalizethe motion that made it stick

A workflow without a motion is dead. This is how it became the way the team works.

Made the positioning a rep talk track

The same story the site tells, in the rep's own words, so nothing got lost between the page and the call.

Killed the freelance pitch

The team got coached off improvised messaging and onto the one narrative that was actually working.

Rehearsed the story until it was natural

Reps practiced the narrative so it sounded like them, not a script they were reading.

T / Tie backthe result it drove, and how they know

RESULT · 01

108 to 0

undifferentiated commodity-utility prompts, eliminated.

RESULT · 02

2 to 56 Define prompts

the category ground nobody owned.

RESULT · 03

One source, versioned

the story stops drifting.

OUTPUT · 01

A buyer-language corpus

The words the engine will source.

OUTPUT · 02

Positioning built to be quoted

Written for the answer, not the page.

OUTPUT · 03

One versioned story

Locked, not drifting across surfaces.

The Sales Operator

06 / HOW YOU DO IT TOO

Here is what I built. Here is how you build it.

The whole thing installs as one plugin. Or grab the skills a la carte. Everything on this page, runnable, on what you paste today.

✓ EVAL PASS · 4/4

THE PLAYBOOK · ONE INSTALL

Get found in the AI era (AEO/GEO)

4 skills chained into one runnable play. Installs as a single plugin, no copy-pasting each skill. It runs on what you paste; connect your stack to go live.

SEE THE FULL PLAYBOOK →

SEE IT RUN IN CLAUDE

EXAMPLE CHATCasey, Founder, running the loop in one sitting

Buyers are asking ChatGPT before they ever talk to us. First, what language do our actual customers use?

Step 1· Customer Research

I built the interview slate, a guide that won't put words in their mouth, and the way to turn transcripts into positioning and messaging inputs. We start from their words, not ours.

Receipt

Slate of 10, a non-leading guide, and the phrases pulled from real transcripts.

Turn that into positioning we all say the same way.

Step 2· Positioning Brief

I grounded a positioning statement and message house in the customer language we just captured, so the site and the reps sound like the buyer, not the brand deck.

Receipt

Positioning statement in customer words, value pillars, and reframes.

Now audit our site. Will an answer engine actually quote us?

Step 3· Seo Aeo Audit

I checked entity clarity, liftable answers, schema, and internal links on both the search and answer-engine tracks, then ranked the fixes by effort versus impact.

Receipt

12 fixes ranked; entity clarity and liftable answers are the top two.

Brief the content that gets us quoted.

Step 4· Content Brief

Each brief is built around a question buyers actually ask an engine, written so the piece can carry the liftable-answer format Step 3 just flagged.

Receipt

3 briefs built around buyer questions, each one written to be quotable.

THE SKILLS IT USES

Make a strategic doc land in one read.

strategic-doc-structure

Strip the AI tells and make it sound like you.

humanize-draft