For solo AI automation builders

You can build the automation. You still cannot say what you sell.

Six Claude Code skills and seven worksheets take one automation idea to a One-Page Offer Sheet — named, bounded, priced, with the arithmetic still attached.

Markdown files you install and run. Nothing to log into, no subscription. The worksheets and templates work on their own if you are not set up with Claude Code yet.

Instant download · Lifetime updates

AI Automation Offer Sprint — $37 one time. No subscription.

✓ Commit to one niche, on written anchors
✓ Price from your own cost, not from feel
✓ Write the exclusions before month two
✓ Leave with a page you can send

offer-sheet-quote-response-line.md

PRICING HYPOTHESIS

Build estimate	28 hours
+40% estimation correction	39 hours
× your target rate	$85/hr
× 1.25 unbilled sales time	$4,144
Setup fee hypothesis	$4,000–5,500
Monthly, to hold the retainer floor
Model + telephony + hosting	$55
Support, 2 hrs × $85	$170
÷ 0.30 at 70% target margin	$750
Retainer floor	$750/mo

If interviews show resistance above $600, the options are reduce support hours, productize, or shift into setup — not accept a number below the floor.

Illustrative. Every figure traces to a stated input or a labeled assumption.

Building was never the bottleneck.

You can wire a form to a model, route the output, catch the failure, and hand the result to a human on the right channel. You have done it for money. That skill is not in question, and it is not what is stopping you.

What is stopping you is the sentence.

Someone asks what you do. You say you build AI automations for small businesses. It is true. It also describes four hundred other people, gives the listener nothing to say back, and turns the next question into “so what does something like that cost?” — which arrives before you have decided, so you answer with a number invented in the pause. Then the scope becomes whatever gets asked for, because you never wrote down what you refuse to do.

The failure is not effort. It is that packaging is a separate discipline from building, and nobody who learned n8n was taught it.

Seven decisions sit behind that one sentence. Six of them get made badly under the pressure of a live call; the seventh — finding out whether any of it survives contact with someone who could say no — rarely gets made at all. So they get made in a system instead. That is what this is: the decisions in order, each with the inputs it needs, a rule for making it, and an artifact at the end.

The Offer Sprint

Seven steps. Each consumes the output of the one before it, and each one is allowed to stop the run.

NARROW  ->  DIAGNOSE  ->  ARCHITECT  ->  PRICE  ->  QUALIFY  ->  PITCH  ->  VALIDATE

Every step can end the run. An offer that never had a step tell it no is an offer nobody has checked.

Narrow

Scores three to five candidate niches against written anchors, applies five vetoes, and records one line per loser

You stop hedging across four buyers and commit to one you can name in a sentence

Diagnose

Traces one process, registers every leak, prices one failure, and builds a cost-of-inaction estimate with the arithmetic visible

You learn what the problem is worth in numbers you could show the person who gave them to you

Architect

Produces the name, the promise, the mechanism stages, the deliverables, an exclusion for every inclusion, and a timeline built backward from access

You decide what you will refuse to build, in writing, before anyone asks for it

Price

Derives delivery cost, a setup band, monthly cost to serve, a retainer floor at a stated margin, and a ceiling taken only from the pain map

You get two bands you can say out loud, and a verdict on whether the offer works at all

Qualify

Writes the ideal client as six externally checkable filters, adds buying triggers, hard and soft disqualifiers, and a per-prospect scorecard

You decide who you decline, and how early

Pitch

Writes the One-Page Offer Sheet plus a 30-second pitch, a website paragraph, a discovery-call opener, and three hooks — then audits every claim in them

You get the words, and a log of every sentence that was cut for being unprovable

Validate

Five days, two real conversations, and a written decision made against criteria dated before any evidence existed

You find out cheaply whether the offer survives contact with someone who could say no

Step	What it does	What changes for you
Narrow	Scores three to five candidate niches against written anchors, applies five vetoes, and records one line per loser	You stop hedging across four buyers and commit to one you can name in a sentence
Diagnose	Traces one process, registers every leak, prices one failure, and builds a cost-of-inaction estimate with the arithmetic visible	You learn what the problem is worth in numbers you could show the person who gave them to you
Architect	Produces the name, the promise, the mechanism stages, the deliverables, an exclusion for every inclusion, and a timeline built backward from access	You decide what you will refuse to build, in writing, before anyone asks for it
Price	Derives delivery cost, a setup band, monthly cost to serve, a retainer floor at a stated margin, and a ceiling taken only from the pain map	You get two bands you can say out loud, and a verdict on whether the offer works at all
Qualify	Writes the ideal client as six externally checkable filters, adds buying triggers, hard and soft disqualifiers, and a per-prospect scorecard	You decide who you decline, and how early
Pitch	Writes the One-Page Offer Sheet plus a 30-second pitch, a website paragraph, a discovery-call opener, and three hooks — then audits every claim in them	You get the words, and a log of every sentence that was cut for being unprovable
Validate	Five days, two real conversations, and a written decision made against criteria dated before any evidence existed	You find out cheaply whether the offer survives contact with someone who could say no

Six of the seven steps are Claude Code skills. Validate is a plan, because the seventh step is conversations with people and no document runs those for you.

What it costs you in time, stated before you buy. Roughly twenty to twenty-five hours of your own working time across two weeks, about nine of them in the validation week. A good share is spent talking to operators rather than typing, and the steps that look skippable are the ones the price is built on. This is not an afternoon.

What is actually in the box

Every claim on this page is checkable against a file that ships in the product.

skills/

18 files

Six skills, each a SKILL.md plus two reference files. Required inputs, the questions to ask when an input is missing, decision rules with thresholds, an output schema, failure modes, and a worked example.

worksheets/

7 files

The fill-in instruments, one per decision. Every field has an inline instruction saying what a good answer looks like and what a bad one looks like. Every worksheet ends in a completed example.

templates/

4 files

The One-Page Offer Sheet, the customer interview script, the five-day validation plan, and the 44-item validation checklist.

examples/

5 files

Five complete runs, start to finish, in five verticals. Every one shows at least one decision that was rejected and why.

What each thing in the box lets you decide

The left column is a file. The right column is the decision you can finally make once you have run it.

In the box	The decision it lets you make
Niche opportunity scorecard	Which single niche you commit to, and which candidates you stop relitigating every Monday
Customer interview script	Whether this call is research or a pitch, and what you say when they ask for a price for the third time
Business pain map	Which single leak your offer attacks, out of the five or six you can see
ROI assumption worksheet	Whether your value story survives a stress test you ran, rather than one a prospect runs for you on a call
Offer architecture worksheet	Whether you have a repeatable offer or a description of custom work with a name on it
Exclusion library	What you will refuse to do at 9pm in month two, decided while it is still cheap to decide
Pricing hypothesis worksheet	What this work costs you to deliver, which is the only number in pricing you can actually know
The floor-versus-ceiling verdict	Whether to take the offer to market, productize it, or stop — when the floor sits above the ceiling, the answer is a different offer, not a lower price
Ideal customer profile	Whether your ICP is a filter or an adjective. If you cannot list matching companies from public sources, you do not have one yet
Prospect qualification scorecard	Which companies get a call this week, which go on a watch list, and which get an honest no
One-Page Offer Sheet	What you send the same day as the call — and, just as usefully, what you refuse to send before it
Offer validation checklist	Whether you are ready to pitch, decided by checks you can point at instead of by how the offer feels

Who this is for

The n8n or Make builder with three delivered projects and nothing repeatable between them. Every one was quoted from feel, ran long, and ended when the work ended. You want the fourth one to be the same offer as the third.
The Zapier or GoHighLevel freelancer whose scope is whatever the client asks for. You have never written down what you refuse to do, so it gets decided in a Slack message on a Thursday night.
The Claude Code builder who can ship an agent and cannot describe it without saying "agent". Your website says what you use. It does not say who you are for or what they get.
The builder who has answered "so what do you charge" with a number invented in the pause and then spent four months delivering against it.
The operator who has picked four niches this year. You have researched HVAC, dental, real estate, and law firms, committed to none, and started over each time somebody posted a case study.

Who this is not for

Read this before you buy. It is here to stop the wrong purchase.

Anyone who needs clients this week. Nothing in this product contacts anyone. It decides what you sell and writes the words you would use; sending them to people is a different job entirely. If your actual problem is an empty pipeline this month, this is not the purchase that fixes it.
Anyone unwilling to talk to operators. Two of the seven steps run on quotes from people who are not you, and the validation checklist fails the whole run without three names and three dates in your notes.
Anyone who wants to be told the price. There is no rate card in this product and there will not be one, because there is no honest source for it. The system computes bands from your delivery cost, your target rate, and the value ladder your interviews support.
Anyone who wants workflows, proposals, or contracts. No n8n JSON, no importable automations, no statement of work, no client proposal, no QA plan, no handoff pack. All of that is deliberately absent — it starts after someone says yes, and it belongs to Agency Stack.

After someone says yes

Agency Stack — the client delivery system

The Sprint decides what you sell. This is the other half: ten connected Claude Code skills that run the engagement from the first call to the renewal, in four stages. Each stage produces documents a client pays for — not productivity tips.

Stage 1

Win the project

Turn a first call into a paid engagement instead of a free consultation.

A scored opportunity report with conservative ROI, so the business case is visible before you quote
A setup fee and retainer with a margin floor and a value ceiling, so the price holds
A proposal led by the cost of inaction, which moves the conversation from price to value

Stage 2

Protect the work

Decide what is out of scope before a client decides it is in.

A statement of work with written exclusions and change-order triggers, so scope creep becomes paid work
An access, fields and rate-limit map per system, with a confidence rating, so integration risk surfaces before the sale rather than in week three

Stage 3

Build it

Go from an approved scope to a blueprint you can hand to yourself on Monday.

An architecture document, importable workflow JSON, and a test plan
An agent specification whose guardrails are enforced in code, not requested in a prompt

Stage 4

Prove it and keep it

Make approval objective, and make the retainer easy to defend twelve months in.

A conversational test suite and a signed acceptance record
An owner overview, a staff guide, a training outline and an acceptance sheet
A concise, outcome-focused monthly report

Five verticals already mapped

HVAC / home services

Emergency routing

Dental

Clinical routing

Med spa

DM + phone, price guard

Restaurant

Allergy + complaint routing

Property management

Habitability + legal routing

10 skills, 42 ready-to-use files

Structured instructions, references, and templates for the full client lifecycle — not a folder of generic prompts.

10 connected worked examples

Follow one HVAC client from the first call to the first monthly report, with consistent assumptions and numbers throughout.

A practical 12-page guide

Install in minutes, see three complete runs, avoid six expensive mistakes, and keep the one-page reference nearby.

Five verticals already mapped

HVAC, dental, med spa, restaurant, and property management — with relevant defaults, safety branches, and benchmarks.

Lifetime updates

When the skill format changes or a benchmark moves, you get the new version. A changelog shows what changed.

Commercial use of outputs

Everything you generate is yours. Send it to clients, put your name on it, charge for it.

Agency Stack — $117, one time

Buy this one if you already know what you sell and to whom. If you cannot finish the sentence “I build X for Y, and it costs about Z,” start with the Sprint below — Agency Stack runs an engagement, it does not decide what the engagement is.

Vertical Pack $37 available at checkout

What buyers say about Agency Stack

These are from buyers of the $117 Agency Stack, not the $37 Sprint. The Sprint is new and we have no results to show for it yet — when we do, they will be labeled as plainly as these are.

“Since buying Agency Stack, I've been closing several projects every week. I made $10,000 in a single week.”

Steven

“I haven't stopped closing sales since I bought it. It helped me significantly improve my service portfolio.”

Diego

“I recommend this package to everyone. I've been closing new projects every day.”

Ana

These are individual experiences, not typical results. Nothing here is a promise or guarantee of income. What you earn depends on your own offer, pricing, market and sales activity, and many buyers earn nothing at all.

Objections

“I already know my niche.”

Then Narrow takes twenty minutes and you carry the answer forward. That is the correct outcome and the one the step is built to allow. But check the claim against the scorecard’s vetoes first: knowing your niche means you can name a trade, a countable size band, and a geography, and then list matching companies from public sources inside a timer. If the sentence you have is "small businesses that are behind on tech," you have a preference, not a niche, and every downstream step inherits the vagueness.

“I can just ask Claude for this.”

You can, and you will get an offer. The question is what stops it from being confident and wrong. What you will not get, unprompted, is the 40 percent estimation correction on its own line, a retainer floor computed at a stated margin, a verdict when the floor sits above the ceiling, an assumption register with a validation step per row, or a 44-item checklist where one honesty failure stops the pitch outright. This product is the constraints, not the text generation. The most valuable line in the system is the one where a skill declines to produce a number and hands you the question that would get it instead.

“$37 seems too cheap to be good.”

Reasonable instinct. This product owns one decision — what you sell — and it stops there hard. No proposals, no contracts, no builds, no QA, no handoff, no reporting. That boundary is what makes it small enough to finish in a few sittings. The larger system that owns client delivery is Agency Stack at $117, and it costs more because it covers ten skills’ worth of work that starts after someone says yes. $37 is the price of a scoped product, not a signal about its density.

“I do not have clients yet.”

Then you are the intended buyer, and you should also hear the limit clearly. An offer is what you need before outreach, not after it. The limit: this product cannot get you a client and does not claim to. It produces an artifact. Whether anyone buys what is on it depends on your niche, your market, your delivery, and a hundred things no document controls.

One idea in. One priced offer out.

Start with the offer. The rest of the ladder waits.

You will not leave with a plan to write an offer. You will leave with the offer.

AI Automation Offer Sprint

Decide what you sell.

Start here

$37

one time

No subscription, no upsell required to use it.

✓ Six Claude Code skills, 18 files
✓ Seven worksheets, four templates
✓ Five complete worked runs
✓ Commercial use of your outputs

Secure checkout via Stripe · Instant access

Add the AI Agency Outreach Kit for $27 at checkout. You’ve built the offer. Now get the emails, DMs, Loom script, prospect filters and 14-day follow-up sequence to start booking conversations.

Agency Stack

Run the client work once someone says yes.

Next rung

$117

one time

Ten skills, 42 files. No subscription.

✓ Discovery audit, pricing engine, proposal
✓ SOW and scope guard
✓ n8n architecture, agent specs, QA
✓ Handoff and monthly reporting

Secure checkout via Stripe · Instant access

Add the Vertical Pack for $37 at checkout. Importable n8n workflows, discovery scripts, and filled audits and proposals for all five verticals.

The 14-day output guarantee

Use the QuickStart and complete the required inputs. If you cannot produce a usable One-Page Offer Sheet containing an ICP, problem, promise, mechanism, deliverables, pricing hypothesis and pitch within 14 days, contact us for a full refund.

The guarantee is tied to the artifact, because the artifact is the part this product controls. It is a statement of confidence, not a hurdle: our refund policy is fourteen days from purchase, no questions asked, and nothing here narrows it. You do not have to show us an offer sheet to get your money back.

What this product does not promise: clients, calls, revenue, or that anyone will pay the price you derive. It promises an offer you can pitch and the arithmetic to defend it.

Where this product stops

This page sells a product that ends at a tested offer. The boundary is stated here, before the button, on purpose.

What you need	Where it lives
Deciding what you sell, to whom, at what price hypothesis	Offer Sprint — $37
Finding the companies, writing the sequence, booking the call	AI Agency Outreach Kit — $27
Audit, ROI, proposal, SOW, n8n build, QA, handoff, reporting	Agency Stack — $117
Importable workflows and filled artifacts for five verticals	Vertical Pack — $37

A boundary stated before the purchase is positioning. The same boundary discovered after it is a refund. Nothing on this page or in the product is legal, tax, or financial advice.

Questions

About both products, and about where one ends and the other begins.

Is this a course?

No. There are no videos to finish before you get value. Install the skills, describe a real client situation, and Claude produces the deliverable using a proven structure.

Do I need to know how to code?

You should know what a webhook is and have built at least one automation. The commercial skills create documents; the build skills create n8n workflows and system prompts. You do not need to write code, but the Stack is designed for practitioners, not complete beginners.

Does it work without n8n?

The commercial skills — audit, pricing, proposal, SOW, handoff, report — work with nothing else installed. The build skills are n8n-specific.

How much will I make with this?

There is no guaranteed income. The Stack helps you diagnose value, price with a margin floor, control scope, and deliver more consistently. Your results still depend on your offer, sales process, pricing, and clients.

What is a Claude Code skill, technically?

A folder with a SKILL.md file and supporting references. Claude reads the description, loads the skill when your request matches, and follows the instructions. You do not invoke them by name.

What if Anthropic changes the skill format?

We update the Stack and you get the new version. That is what lifetime updates means here. Obsolescence is the reason updates exist.

Can I use the outputs with clients?

Yes. Every audit, proposal, SOW, architecture document, and report you generate is yours to use commercially with no attribution. What you cannot do is redistribute or resell the skills themselves.

Is there a refund?

Fourteen days, no questions. Email us and it is done. We would rather refund than argue.

Do I need the Sprint if I already have an offer?

No. If you can name a trade, a countable size band and a geography, and you have a price band you can defend, you already have what the Sprint produces. Go straight to Agency Stack. The Sprint exists for the builder who cannot finish that sentence.

Can I buy both at once?

They are two separate checkouts, deliberately. Each one carries its own order bump — the Outreach Kit at $27 with the Sprint, the Vertical Pack at $37 with Agency Stack — and bundling all four into one line would hide what you are actually paying for. Buy the one you need now; the download page remembers what you own.

One idea in. One priced offer out.

Six Claude Code skills. Seven worksheets. Four templates. Five complete worked runs.

Instant download. Lifetime updates. Commercial use of your outputs.