For the owner who wishes they could clone themselves

Almost every time I post an AI tool I built, somebody tells me it already exists.

Sometimes they're right. Price tracking existed, kind of. Labor reports existed for sure. Software that flags an invoice going up has been sold to restaurants for years, by companies with booths at trade shows and reps who follow up.

Many of these tools existed in some form or another. What changed with AI is a massive reduction in time and friction.

You're spread thin enough that you ignore everything that isn't an emergency. That's a rational way to run a restaurant. The walk-in going down is an emergency. A vendor moving a case price four dollars is a line item you'll look at when things calm down . . . and things don't calm down.

So the task stays on the list. Not for a week. For years, in some cases, and the money it would have saved leaks the entire time, in amounts small enough that no single week makes you act.

For a long time I kept telling myself that some day I’d be disciplined enough to address the list. Then I did the math on it.

A job that takes five hours only gets done if you have five hours. But you have the hours you have.

Which means the number that decides whether a problem in your restaurant gets solved is how long the fix takes. How much it costs you to leave it rarely enters into it. (Partly because you weren’t able to put a number on it.) How much you care doesn't enter into it. Calculate a rough idea of the time it’ll take to figure it out and you’re suddenly muttering “Fuck it, I’ll deal with it later.” How many times have you said that?

Move a task from five hours to ten minutes and it crosses a line into “doable.” It stops competing with the walk-in. It stops needing an open week that never arrives. It gets done on Tuesday morning while you're doing something else.

Ruben Hassid has a framing for this I keep coming back to, from a conversation on Scott Clary's podcast. He describes AI as being at its best when it takes a rough idea or a mediocre first draft and works it into something professional.

In a restaurant that’s the entire opportunity, because so much of what you would fix is sitting right at that level. Rough idea … mediocre first draft… yeah, I had projects like that that had been sitting for a while.

Your best work gets to stay your best work. What gets addressed is everything you were never going to get to, or much improved versions of your previous mediocre solutions.

Every invoice gets price checked. Every morning I know what last night cost me in labor against what it made. Prime Costs get reported. Overhead expenses are tracked. Every review gets a response.*

I didn't add hours to do that. I removed the hours it used to take (when I did it), and the work moved onto the done list without me sitting down to it. No guest or employee noticed anything change.

That's the closest thing to a clone I've found. There was no transformation. A handful of jobs got cheap enough to actually happen.

*Here's one for you. I'm pretty good about responding to Google reviews and I always have been but I made a decision at some point a few years ago that I didn't care about Yelp reviews because I fucking hate Yelp. When I made the tool that drafted review responses for my approval, I decided to catch up on Yelp. We had 279+ reviews per store to catch up on. That wasn't even a time thing. It was just something that I didn't value and I didn't want to address. It didn't make it less important. Now it gets addressed, and that's good for the businesses.

Fewer emergencies, because most emergencies are a small problem that had a long time to grow. (Side note: Clean your walk-in condenser coils. Mine was down for two days because I didn’t do the scheduled maintenance and this just reminded me of that.)

The vendor price increase you catch in week one is a conversation. The same increase caught much later is a margin problem with a history, and now you're having a harder conversation from a worse position while three other things are on fire.

Speed compounds as a direct corollary to calm. When the cheap version of a task takes ten minutes, you do it early, and doing it early is what keeps it from becoming the thing that ruins a Thursday.

I've said that every problem in restaurants can be solved, and people argue with me about it. Well, I’ll double down: Not only did most of those problems have solutions already, but they were just waiting on somebody who had the time to make it happen. There has never been a better time to make time.