Is your business AI-proof? The would-you-do-it-anyway test

Most checklists for an AI-proof business only look at the business: its moats, its data, its license. ZeroCrew thinks that misses the biggest risk. The same AI that could copy your product also lets ten new founders launch it this weekend, so the business that lasts is the one whose founder keeps going for years, and people keep going at work they love.

The would-you-do-it-anyway test is ZeroCrew's check for an AI-proof business: seven yes-or-no questions in two layers. The first layer checks you, the founder, because your energy is the part no model can supply. The second layer checks the business for moats AI can't copy. ZeroCrew adapted four of the questions from a popular one-point-per-yes checklist shared among founders and added the founder layer.

What makes a business AI-proof?

An AI-proof business is one that AI makes stronger instead of replacing. ZeroCrew's view is that two things decide it: a founder who would keep doing the work without needing the money, and at least a few moats a general model can't copy, like a community, physical work, a license, or knowledge from every job.

AI's reach is the reason ZeroCrew's test checks the founder first. The latest revision of "GPTs are GPTs", by researchers at OpenAI, OpenResearch, and the University of Pennsylvania, estimates that about 15% of all US worker tasks could be done significantly faster with a language model at the same quality, and 47% to 56% with software built on top. ZeroCrew's reading: when that much work gets faster, more people can build the same product, so staying power matters more.

What is the would-you-do-it-anyway test?

The would-you-do-it-anyway test is a seven-question check ZeroCrew built for founders who want to know if their business will last as AI gets better. Three questions are about you and four are about the business. Answer each one yes or no, honestly, about the business you run today, not the one you plan.

#	Question	Layer	What a yes means
1	Would you still do it if you were rich?	You	You'd keep going after competitors quit
2	Does your human energy change the result?	You	Customers feel the difference a caring person makes
3	Do people come for you?	You	Your taste, voice, or characters are the product
4	Do people come for each other?	Business	Community, events, or both sides of a marketplace
5	Does it touch the real world?	Business	Someone shows up, or a license or contract is needed
6	Do you get smarter with every customer?	Business	You learn what a general model never sees
7	Does AI make you cheaper, not obsolete?	Business	Customers pay for the result, and AI does more of the work

Would you still do it if you were rich?

Question one is the gate of the test: would you still do this work if money no longer mattered to you? ZeroCrew puts it first because AI lowers the cost of starting almost any business, so the founders who last are the ones who would keep going anyway. A yes on question one makes every other answer count.

Passion and staying power are linked in the research. In a study of 129 entrepreneurs published in Entrepreneurship Theory and Practice (2015), Melissa Cardon and Colleen Kirk found that passion for inventing and passion for founding helped explain why confident founders persist. ZeroCrew's reading for the AI era: confidence gets cheaper when a model does the work, and love for the work is what keeps a founder at it.

To answer question one honestly, picture a year with no money worries and ask which parts of your business you'd still do on a Monday morning. ZeroCrew suggests writing those parts down, because that list is the core you protect. The parts you'd drop are the work AI agents can take, which the last section of this article covers.

Does your human energy change the result?

Question two asks whether a customer can feel the difference when a person who cares does the work. ZeroCrew counts a yes when the effort itself is part of what people buy: a coach who notices, a chef who cooks for the room, a teacher who stays late, or a craftsperson whose hours show in the piece.

People value work more when they believe effort went into it. In "The effort heuristic" (Journal of Experimental Social Psychology, 2004), Justin Kruger and colleagues found that people rated the same poem as better, and worth more money, when told the poet spent 18 hours on it rather than 4. ZeroCrew's takeaway: when AI output is instant, visible human effort becomes a reason to choose you.

Do people come for you?

Question three asks whether customers choose you for your taste, your voice, or the characters you create, rather than for a task anyone could do. ZeroCrew counts a yes when people would notice and care if someone else did the same work in your name. An audience that follows a person is hard for AI to take.

The label "made by a human" changes how people judge work. In two studies published in Cognitive Research: Principles and Implications (2023), Lucas Bellaiche and colleagues labeled AI-made paintings at random as human-made or AI-made, and people liked the "human" ones more and rated them more beautiful, profound, and valuable. ZeroCrew applies the same idea to founders: a person people follow is worth more than output they could get anywhere.

Which business moats can't AI copy?

Four moats make up the second layer of the test: people who come for each other, work in the real world, knowledge that grows with every customer, and a price tied to the result. ZeroCrew adapted these from a popular one-point-per-yes checklist and turned its last question around, so AI counts as help, not as a threat.

• Do people come for each other? Communities, events, and marketplaces get better as more people join. A model can copy the software, not the members.
• Does it touch the real world? Someone has to show up, build, or fix something, or you hold a license, approval, or contract that no prompt can get you.
• Do you get smarter with every customer? Every job teaches you something a general model never sees: your customers' edge cases, prices, and patterns.
• Does AI make you cheaper, not obsolete? If customers pay for the outcome, AI doing more of the work lowers your costs instead of replacing you.

Selling the result tends to turn AI into an advantage. In "Generative AI at Work" (NBER, 2023), Erik Brynjolfsson, Danielle Li, and Lindsey Raymond studied 5,179 customer support agents and found that an AI assistant raised issues resolved per hour by 14% on average and by 34% for newer agents. ZeroCrew's point for founders: when customers pay for the result, faster work lowers what each result costs you to deliver.

How do you score the would-you-do-it-anyway test?

Score the test in two steps: question one first, then one point for each other yes. ZeroCrew doesn't treat every yes as equal, because a moat around a business you'd quit protects nothing. If question one is a no, rethink before you defend. If question one is a yes, the other six tell you what to build next.

1. No on question one: find the part of the work you'd do anyway, and rebuild the business around it.
2. Yes, plus 5 or 6 points: build big. Your business is likely harder to copy, so put AI to work on everything else.
3. Yes, plus 3 or 4 points: add one moat. A community or knowledge from every job is often the fastest to build.
4. Yes, plus 0 to 2 points: you're the moat for now. Grow an audience that comes for you, then add a second moat.

Which jobs and businesses does AI replace first?

AI tends to replace tasks before it replaces whole businesses, starting with work that's done on a screen, follows clear rules, and has no relationship attached. ZeroCrew's directory of AI agents by job shows where that's already happening: answering calls, bookkeeping, sales outreach, and first drafts of content all have agents founders can use today.

Jobs that score well on the would-you-do-it-anyway test usually involve hands, care, presence, or a license: trades, care work, coaching, live events, and craft. ZeroCrew's advice for anyone searching for AI-proof jobs or careers is the same as for founders: pick work you'd do anyway, then make sure it has at least one moat a model can't copy.

How does ZeroCrew help you keep the part you love?

The most AI-proof move is to give AI agents the work you'd drop if you were rich and keep the work you'd do anyway. ZeroCrew helps founders do that: the directory lists the AI agents and tools ZeroCrew researched for every job, and AI agent setup puts them to work for you.

Three jobs founders often hand to AI agents show how picking a tool works on ZeroCrew. Each job page ranks the tools ZeroCrew researched, by search demand or in three ways side by side, so you can compare them and pick the one that fits the tools you already use.

• Bookkeeping: Pilot, Puzzle, or Digits automate your bookkeeping and close your books. Pick Pilot if you want human bookkeepers to do the books, Puzzle if you want to run them yourself and add its bookkeepers later, and Digits if you want AI to automate the month-end close. See the bookkeeping page.
• Phone calls: Smith.ai, Synthflow, Goodcall, Rosie, or Dialzara answer your phone and take bookings. Pick Smith.ai if you want live people backing up the AI, and Synthflow if you want to build your own voice agent. See AI receptionists.
• Routine work across apps: Zapier, n8n, or Make connect your apps and run tasks on a schedule or a trigger. Pick Zapier for the most ready-made connections, and n8n if you want to host it yourself. See automation.

If you'd rather not set the agents up yourself, ZeroCrew does it for you. After a free discovery call, ZeroCrew picks the best agents for each job, connects them to the tools your business already uses, and keeps them up to date as new models come out, starting at $1,000 a month as of November 2026. You keep the work you'd do anyway.