Skip to content

Before The Build: Map the Money, Manage the Risk

Two concepts, in order: make money, and keep it. Most builders only ever work out one of them.

Free, nothing gated, and no email required. Everything you type stays in this browser and is never sent anywhere, which is architectural rather than a promise: there is no server call on this page to send it with.

CLARITY ON WHAT YOU ARE ACTUALLY BUILDING, BEFORE YOU BUILD IT.

Clients Don't Want AI Systems, They Want Profits.

Cody Landry, in the pose of someone about to do something the Declaration of Independence does not cover
How people look at me when I tell them I map AI to revenue.

A client telling you they need a website is not a brief. It is a symptom. We would never build the website without first asking what got them to that sentence, because half the time the honest answer is that they do not need a website. They need sales, and a website was the only lever they knew how to name.

An AI request arrives the same way. "We need a chatbot" or "we need an AI agent" is somebody self prescribing, the same way "we need a website" always was. Nobody asked what problem it is actually solving, including the person asking for it, and nobody asked whether it moves a number they actually get paid on. Clients do not want an AI system. They want the profit it was supposed to produce, and the system is just the thing they knew how to ask for.

So before either module below, ask the question the client did not: what is this actually supposed to move, and for whom. The pricing module only prices what you tell it to. If what you tell it to price is the wrong thing, a defensible number on the wrong system is still the wrong system, just priced with more confidence.

Map the money before you take the order. That is the whole reason this comes first.

How to work through this

Run it against a real AI idea if you have one on your desk right now, a client's or your own. Your own is a better first run than a hypothetical, because every answer is checkable and you will find at least one open gap.

The pricing module turns a handful of numbers into a fast answer: worth pursuing, or not yet. Nothing is scored here in the sense of a grade, it is arithmetic, and it is the gate before you spend more time on this idea.

The foundation module has what to ask, how to actually check it, and what a gap answer means for the build about to happen. Mark each one as you go. Its output is a list of what is settled and what is not, not a number, and it is what turns the price above into something you can defend. What you get at the end is a text summary of everything you entered, both modules together. Copy it into your notes or your proposal. That is the point of doing it here rather than reading it.

  1. 1. Pricing: make the money, value based and mapped to revenue

    Before you spend an hour securing this or proving out its compliance, find out whether it is worth building at all. This prices what you are about to build against what it is actually worth to the client, not the hours it takes you to build it.

    Go to module 1

  2. 2. Foundation: keep the money, what to answer before you build

    You just found out this is worth pricing. Before that number is something you can defend to a client, run every idea through the same seven questions, in the same order, before you write the first prompt. That consistency is what makes the output usable across completely different projects.

    Go to module 2

1. Pricing: make the money, value based and mapped to revenue

Before you spend an hour securing this or proving out its compliance, find out whether it is worth building at all. This prices what you are about to build against what it is actually worth to the client, not the hours it takes you to build it.

Hours times your rate prices your time. This model prices a share of what the system is worth to the client instead, which is a different number and usually a bigger one, because a client about to get real value from something will pay for the value rather than for your afternoon.

If you already ran the worth it calculator on the main AI Unleashed page, you have the exact numbers this needs: hours a week, coverage, loaded hourly cost, margin impact, running cost. Put the same ones in below rather than working them out twice. If you have not run it yet, this module asks for that same handful of inputs itself, because pricing a system and deciding whether to build it lean on the same arithmetic, just aimed at a different question.

What to charge for it

Value first, then a fee. The value comes from the same math as the worth it calculator on the main page, and the fee comes from a share of that value rather than from your hours.

QUESTION 1 OF 5

Across everyone involved, not per person. The same number as the worth it calculator, if you already ran it.

This band is Cody's own starting framework, not a rate card. Argue with it if your market or the client's risk tolerance says otherwise. Nothing you enter into this calculator leaves this browser either.

Open the worth it calculator on the main page

2. Foundation: keep the money, what to answer before you build

You just found out this is worth pricing. Before that number is something you can defend to a client, run every idea through the same seven questions, in the same order, before you write the first prompt. That consistency is what makes the output usable across completely different projects.

It's not what you make, it's what you keep. This module is the keep half: the audit trail, the reverse path, the kill switch, and the compliance exposure you did not name out loud. Skip it and there is nothing left to keep.

These seven are what Cody calls filter prompts: the same diagnostic questions, run on every AI build idea before design starts, so the output is consistent no matter how different two projects are. They come from two places in Hollow Buttons, the book Cody is writing on the gap between software that looks finished and software that is: the four question Hollow Test, and the longer Enterprise Pre-Flight Questionnaire built for bigger engagements. Both were written to check a feature somebody already shipped. Here they are turned around, asked before the first prompt is written rather than after the thing exists.

Work them in order, on a real client idea if you have one and on your own next build if you do not. Each one is a question you can actually ask out loud in a scoping call, with the sub checks that back it up and what a gap answer costs once the build is underway. An open gap here is not a failure. It is the cheapest place in the whole project to find one.

STEP 1 OF 7

2.1 Can you audit it?

Before you build, ask whether every decision the AI will make on the client's behalf can be traced to who or what triggered it, and when, from a screen the client can open themselves, without a developer reading logs for an evening.

  • Name the exact action the AI takes unsupervised: it sends an email, updates a record, quotes a price. Ask whose name is on the log line when it is wrong.
  • Check whether the planned log captures the actor, the input, the decision, and both the before and after state, or just a line that says 'processed.'
  • Decide up front how long that log is kept, as a deliberate choice rather than a platform default you never looked at.

What it means: An AI system with no real audit trail is the exact gap Hollow Buttons calls the difference between a security sticker and a deadbolt. 'It has a login' is one claim. 'Here is who can log in, and here is the record of every attempt' is another. Design the log before you design the feature, because it is much more expensive to add after the AI has already made six months of undocumented decisions.

These seven questions are Cody's own filter, adapted from the Hollow Test and the Enterprise Pre-Flight Questionnaire in Hollow Buttons, the book he is writing on the gap between software that looks finished and software that is, reframed here for an AI system that does not exist yet rather than a feature that already shipped. Nothing you type here leaves this browser.

Open the AI hygiene checklist

Where this sits in what you sell

This exercise does not sell itself from a standing start either. Somebody has to have a reason to believe you can ask the right questions before they will pay you to design something, and the way that reason gets built is by giving away the cheap version of the same judgement first.

This is not theory and it is not a formula from somewhere else. It is the ladder this business actually runs, and each rung below is a real thing on this site rather than an illustration.

  1. 1. Free, and it can tell you no

    A tool that runs a version of the foundation module without a conversation. The AI hygiene checklist covers the same ground: secrets handling, access control, logging, error paths, rollback, human review gates, and what happens when the model or vendor changes underneath you. Run it on the build you are about to scope.

    Run the AI hygiene checklist

  2. 2. Low priced, and complete on its own

    A $47 kit somebody can buy and use without ever talking to you: forty vendor vetting questions, a scoring sheet and nine contract clauses, for the meeting where somebody else wants their signature on an AI build. It asks the same kind of questions as the foundation module above, ownership, access, logging, rollback, exit, credentials, what happens when the vendor changes, pointed at a system somebody is trying to sell them rather than one they are designing themselves. It has to be genuinely useful alone, because a product sold as a lead magnet is a lead magnet with a price on it.

    See Before You Sign

  3. 3. The exercise you just priced

    Your judgement, applied to a system that does not exist yet, with a fee that names what it is worth rather than how long it took. This is the first rung where the client is buying judgement rather than a document, and everything above it exists to make this one an easy yes.

  4. 4. The system foundation and pricing were about

    The actual build: the agent, the automation, the integration, done to the standard the foundation module just walked through. Catalyst is where that gets built, scoped and priced separately, and quoted after this exercise, never bundled into it.

    See how Catalyst is scoped

The rung people skip is the second one, and skipping it is why the real build is a hard sell. Free proves you can ask the right questions. Paid proves somebody will pay for that judgement. Those are two different pieces of evidence, and the second one is the one that makes the actual build easy to quote.

Take it with you

Everything you entered, as text. Copy it into your notes, your proposal, or the covering email. Nothing here has been sent anywhere and closing this tab loses it.

The prompt and agent kit

A scoping prompt to run against whatever LLM you build with, and two agent outlines, an adversarial security agent and a compliance challenge agent, to argue with what that prompt produces. Free, same as everything else here.

Download the prompt and agent kit (PDF)

This exercise is educational. The foundation module is a set of questions to run before you build, and answering all seven does not certify the system as secure, compliant, or safe: it means those seven ordinary, well documented failure modes were considered on the day you ran it. The pricing module prices what a system might be worth, using numbers you typed, and it is not a quote, a guarantee anybody will pay it, or financial advice. Neither module makes an earnings claim: whether anybody buys the build, and what you make from it, depends on you, your market and your reputation, none of which are on this page.

Run it on your own next build first.

Every question above is one you can ask today, about something you are actually about to build, without asking anybody for permission. That is the first run, it takes less than an hour, and it is worth doing before you run it in front of a client.

You will find at least one open gap. Everybody does, and finding it before the first prompt is written is considerably cheaper than finding it after the AI has been making decisions for six months.

Find out what is actually broken before you spend another dollar on AI.

Two weeks, fixed fee, fixed start date. A map of what you run, a ranked list of where AI actually holds with a dollar figure on each, and a fixed price plan to build it. You own the document either way.

Read it. If you do not agree we found at least three things worth fixing, tell us and we refund the fee. You keep the document.

Get the audit

$5,000. TWO WEEKS. THE FEE COMES OFF THE BUILD IF YOU START WITHIN SIXTY DAYS.

WHO THIS IS FOR

A company of roughly ten to two hundred people, already running a CRM or operational system, with revenue coming through it and at least one process held together by hand. That includes agencies, whether the system is for your own operation or for a client's. Pre revenue, or want a chatbot on a marketing site? Wrong firm, and we will say so on the first reply.

$5,000
FIXED FEE. IT COMES OFF THE BUILD IF YOU START WITHIN SIXTY DAYS.
2
WEEKS, ON A FIXED START DATE.