Imagine you run a busy restaurant and you have hired the most talented, expensive chef in the city. People come to the door all day. Half of them are not hungry. They just want to know if you are open, or where the bathroom is, or whether you take cards. And every single time, you send the question all the way back to the kitchen and pull the chef off the line to answer it. He always answers. He is overqualified for it. He is also slowly going broke on your behalf.
That was us, quietly, for a while. We had built a smart system that leaned on a powerful and pricey thinking engine, and we were routing absolutely everything to it, including the questions a doorman could answer in his sleep. Are we awake? Yes. What is running right now? Here is the list. Each of those little answers took a few seconds and cost real money, and there were a lot of them.
The obvious move, when you build something with a clever brain at the center, is to make the brain cleverer. Better instructions. A bigger engine. Feed it more of your own information. Pay a little more next month. We did the opposite, and it felt almost insultingly simple. We put a doorman at the front. Not a genius. The plainest, most boring helper imaginable, the kind that just recognizes a familiar question and hands back the answer it already knows. And every time we taught the doorman one more question, the whole place got faster, cheaper, and steadier. The genius in the back got less and less to do. Everyone out front got served quicker. The unclever thing turned out to be the clever thing.
When the realization fully landed, it was a little embarrassing. Asking the expensive brain takes a few seconds and costs a small coin every time. It can get tired, it can be busy, it can occasionally make something up, it can even give you a slightly different answer tomorrow. The doorman is instant, free, and says the exact same true thing forever. The real question was never "should this be answered by the genius." It was "does this question even have more than one right answer?" Are you open has one answer. Where do I log in has one answer. Spending the chef on those is paying a wine expert to read you the menu out loud.
So one weekend we wired the doorman into the front of everything. Now, the moment a question arrives, the plain fast helper looks at it first. If it is one of those one-answer questions, it replies in a blink, for nothing, and the genius never even hears about it. And here is the lovely part: even on a day when the expensive brain is overloaded and unavailable, the people asking the simple questions still get a real answer instantly, because that answer was sitting in our records the whole time. The system got more dependable as the smart engine got less involved, which is the exact opposite of what everyone selling you intelligence promises.
If you run a business that has started bolting clever automation onto things, sit with this. Most of what your customers ask has exactly one correct answer. What are your hours. Has my order shipped. Where do I sign in. What is your refund policy. None of those are hard problems. They are facts you already have, with a polite sentence wrapped around them. Send each one to an expensive brain and you pay a little coin, wait a few seconds, and accept a small chance it invents a return window you do not actually offer.
The trap is treating "we use smart automation" as the goal. The goal is that the customer gets the right answer instantly, free, every time, with no chance of being wrong. If a doorman does that, let the doorman do that. Save the genius for the moments that truly call for judgment, the messy, ambiguous, this-one-is-different moments where being clever actually earns its keep. Putting your most flexible, most expensive mind on the front desk to recite your opening hours is hiring a surgeon to read the specials. They can. They are wildly overqualified. You will go broke doing it.
Look at every place automation touches your business and ask the same small question of each one: does this have exactly one right answer? If yes, just keep the answer at the door. Save the brilliant, costly thinking for when judgment is the whole point. Your costs flatten, your speed jumps, your mistakes drop, and your customers get served faster. The only one who comes out behind is the genius in the back, and honestly, the genius was exhausted anyway.