Skip to content
Courses / Will AI Eat This Business? / Building at four, five, six

Building at four, five, six

5 min read · The honest verdict

Lesson video in production

The full lesson text below is complete — the video version lands with launch.

Building at four, five, six

If your idea scored in the middle bands, the verdict is build, and build deliberately. A four, five, or six is not a condemnation; it is a work order. The machine will do real work inside your value chain, your competitors will use it, and your margins will depend on which side of the split you place yourself. Every course in this catalog rated in the middle bands teaches some version of the following four moves. Here they are in one place.

Own the accountability layer

In any business where production automates, the durable human asset is being the one who is answerable. The social media marketing agency course in this catalog is the model: posts, graphics, and captions generate in minutes now, and the course says so without flinching. What the client actually retains is the human who records the baseline honestly, translates the numbers into the owner's language, resets expectations when a campaign underperforms, and justifies the next invoice. That layer cannot be generated because it is not content. It is responsibility with a face.

Look at your task inventory and find the rows where you carry consequences. Then design your offer so those rows are the product, and the generated stuff is the packaging. A biller who sells "claims submitted" is selling bytes. A biller who sells "your revenue cycle handled, your compliance mine to lose" is selling accountability, and prices accordingly.

Own the demand

The demand-gate lesson showed what happens to businesses whose customers arrive through a gate someone else controls and fails to police. The adaptation is to move your demand behind a gate you own: a direct client relationship, an email list, a referral loop, a phone number in a hundred neighbors' contacts.

The laundry pickup course teaches exactly this fork. The platform apps will find you customers fast, and the course uses them, but it treats platform dependence as the one AI-shaped risk in an otherwise deflection-heavy business, because an algorithm that feeds you customers can also reprice and re-rank them. The graduation is owning the relationship: the customer who texts you directly, refers the neighbor, and would not dream of re-shopping the route. Every middle-band business needs a version of that graduation, and it belongs in your plan from week one, not as an escape hatch for later.

Price on outcomes, not output

When output gets cheap, pricing on output is a countdown timer. The video editing course makes the point sharply: editors who charge by the video compete with everyone racing to the bottom of the gig platforms, and the floor under per-video pricing is whatever a machine plus a bargain hunter will accept. Editors who price on outcomes, retention, the creator's channel growing, the client not having to think about editing, hold a rate that reflects the result, not the render time.

This is the bytes-business survival skill in miniature. If part of your deliverable is now generatable, stop charging for that part as though it were scarce. Charge for the part the machine cannot produce: the judgment, the taste, the accountability, the relationship. Bundle shamelessly. The generated part is your cost savings, not your price list.

Use the machine openly

There is a foolish version of AI caution where a middle-band operator refuses the tools to prove their humanity. Do not. The low-exposure trades in this catalog demonstrate the correct posture: the cleaning operator with automated scheduling and routing pockets the difference, and no customer has ever asked whether their confirmation text was hand-typed. In the middle bands it goes further. The machine can draft your first pass, your product descriptions, your client reports, your listing copy. Your job is to be the editor, the taste, the warranty, and the name on the door.

The honest boundary is disclosure where the customer would care. If a client is paying for your judgment and your drafts are machine-assisted, fine, judgment is what they bought. If a client is paying for handmade work, "handmade" has to mean handmade. The moment AI becomes a way to secretly downgrade what you promised, you have not adapted. You have joined the slop.

Re-score on triggers, not on vibes

Exposure is a moving number, and the middle bands move the most. Put the assessment on a calendar and on triggers. Re-run the full scorecard every six to twelve months, and immediately when any of these fire:

Re-score triggers

A platform you depend on changes its content, disclosure, or monetization rules
A new model starts doing your core deliverable acceptably, not impressively, acceptably
Prices in your niche start sagging and the cheap providers are winning bids
A regulator or licensing body changes the rules that anchor your accountability layer
Your own customer mix shifts toward buyers who cannot tell your work from the cheap version

A re-score that comes back higher is not a failure. It is information arriving in time to matter: raise the accountability content of the offer, move more demand behind your own gate, or start the next business while this one still pays. The failure mode is drifting from a five to a seven without ever re-running the numbers, because you liked the answer the first time.

Those are the moves for the middle. What remains is you, a fresh idea, and the whole method run end to end.

Downloads for this lesson

Keep going — you're working through Will AI Eat This Business?.

All courses are free ↗