Will AI Eat This Business?
The method this site uses to judge AI risk in any business idea: exposure versus deflection, what a machine replaces versus what it only accelerates, and the honest cases where the answer is do not build it.
Every business idea you are considering carries an AI question, and most of what you read about it is either panic or sales copy. This guide teaches the assessment method we run on every course before we publish it: five questions that sort what a machine replaces from what it only accelerates, a zero-to-ten exposure score, and the adaptation moves that make a five or a six still worth building. By the end you can run the whole thing on any idea in an afternoon, including the honest verdicts: three business types scored so high we removed them from this catalog entirely, and we show you exactly why.
The promise here is candor, not comfort. Every figure in these lessons is sourced to a live page you can check yourself, every rating we cite is the real number published in that course's own frontmatter, and the removals are disclosed by name. You will see where we think the machine wins, where we think the fear is misdirected, and what we did when the answer came back ugly. No scores are hidden and nothing is dressed up as safer than it is.
This guide is not a course on using AI in your business, and it is not a catalog recommendation engine. It teaches one skill: judging exposure for any idea, yours included. When the assessment points you toward a business, the rated course for that business is where the how-to lives. When it points you away, the honest answer is to pick another idea, and this guide will tell you that plainly.
The method
Replace or accelerate
The distinction that does most of the work: a machine that produces your deliverable is a competitor, a machine that does your admin is a subsidy. How to run a task inventory.
What the customer is actually paying for
The deliverable test: bytes, physical outcomes, and accountability. Three kinds of value, three very different exposure profiles, and how to strip a service bundle apart.
Where the blame lands
Deflection sources: liability, licensure, physical presence, and local relationships. Why the businesses machines cannot hold responsible are the businesses machines cannot replace.
The five questions
The full assessment assembled: five questions, a scoring rubric, and the zero-to-ten exposure scale with the bands that tell you what to do with the number.
The evidence
Rated two through six: the catalog as evidence
Every rated course in the catalog, band by band, with the actual reasoning behind each score. Thirty-four worked examples you can check against the method.
Where the demand comes from
The platform pattern: channels that police machine-generated supply keep a human lane open, channels that do not get eaten. Etsy, Google search, YouTube, and Amazon KDP as the evidence.
The entry floor
AI eats the entry lane first, not the top. What the hiring data shows, why the floor is what a beginner actually buys, and the copywriting case that set this catalog's ceiling.
The honest verdict
The honest no
The full disclosure: forty-one planned courses became thirty-four. Three removals for AI exposure, four for capital intensity, named with reasons. What a score of seven or higher means.
Building at four, five, six
The adaptation playbook for the middle bands: own the accountability layer, own the demand, price on outcomes, and use the machine openly. Plus the re-score triggers that tell you when to re-run the assessment.