A guy called me a while back. He had built his own website using AI, he was proud of it, and he wanted help getting more customers. Fair ask. That is the call we want.
So before I pitched him anything I asked him to do something. Open the AI you used to build it, point it at your own site on your phone, and ask it to grade you out of ten on search, on mobile and on how easy the thing is to use.
He came back with three numbers. Three out of ten for SEO. Four out of ten for mobile. Five out of ten for user experience.
He did not need more customers. Customers were finding him. They were landing on a site that did not load right on a phone and could not tell them how to order, and they were leaving.
Why the AI built him a bad site
Not because the AI is bad. Because he asked it to build a website, and it built him a website.
He did not ask it to build a site that loads in under two seconds on a four year old Android on a bad connection in a parking lot. He did not ask it to put the ordering link above the fold because a hungry person decides in about eight seconds. He did not ask it to write the page so Google understands he serves wings in Sussex County. He did not know to ask, which is the entire point.
These tools answer the question you ask. They do not tell you that you asked the wrong question. A professional in any field is largely a person who knows which questions are the real ones, and that does not go away because the typing got faster.
The gap between making a thing and making a thing that works
I am not anti-AI about this. I teach restaurant owners to use it. I use it every day in my own place and in this business. The tools are genuinely good.
But there is a difference between a site that exists and a site that earns. One of them looks finished. The other one is finished, which means it was built by someone who has watched a hundred people try to order a burger on a phone and knows where they give up.
Same thing happens with menus, with photos, with ads. AI will make you something. Whether it makes you something that moves covers depends on whether anyone in the room knew what good looks like.
What he should have done, and what you can do
Two honest options, both fine.
Build it yourself with somebody who knows the field sitting next to you for a few hours. You keep the control and the savings, and you learn which questions to ask. That is most of what we do in a tune-up.
Or hand it to someone and spend your hours on the floor instead. Also fine. Nobody gives you a medal for building your own website at midnight.
What does not work is building it alone, assuming it is done because it looks done, and then spending money on ads to send more people to a four out of ten.
Run the test tonight
Pull up your own site on your phone, the way a stranger would. Ask any AI to grade it out of ten on search, on mobile and on how easy it is to order or book. Be specific, ask for the first thing it would fix.
Under a seven on any of them and you have your answer about where the money is leaking. It costs nothing and takes four minutes, and it is the same thing I would do on the first call anyway.
Next: What branding actually means when you own a restaurant