// Composite narrative — details drawn from real deployments. Names fictional. Feelings exact.
Maria had been running the dining room for eleven years. She was good at it — genuinely, measurably good. Her regulars asked for her sections. Her team did not turn over the way other teams did. Her numbers, when she got to read them, were usually fine.
She got to read them on Monday.
Friday was different. Friday was Maria plus sixty-two covers plus a kitchen running four minutes behind plus the bar without a lead because someone called out plus Table 7 where she could tell something was off but she had not made it there yet. Friday was everything simultaneously and the knowledge that she was always one conversation away from losing the room.
She had learned, over eleven years, to triage. To know what to sacrifice. To accept that some tables would slip while she held others. She had learned to read the Monday report and understand what Friday had cost, and to carry that understanding into the next Friday, and to try to do better.
She was very good at trying to do better.
The First Service
The first Friday with superGM.ai was uncomfortable in the way that good things are sometimes uncomfortable. The dispatch queue was accurate in ways she had not prepared herself for. Table 11 — she would not have gotten there for another six minutes. The system had it flagged at seven-thirty-two. She got there at seven-thirty-six. The guest was still deciding. She caught it.
Table 3 was a VIP she had served before. The system had his profile loaded before she reached the host stand — the bourbon he preferred, the corner table he liked, the fact that he had left a complaint eight months ago about a slow first course. She handled the first course personally. He noticed. He did not say anything. He came back the following Tuesday.
The kitchen had the filet situation flagged at six-forty-seven. She approved the emergency order at six-fifty-one. None of her guests heard the words “we are out of that tonight.”
At ten-fifteen, she sat down with the service summary. Six interventions. Five recovered. One yield event. The yield event was a pricing adjustment she had approved before service — a full Saturday with a convention in town, three corporate tables, the crowd that does not check prices. The adjustment had been clean. Nobody had mentioned it.
What She Said
We asked her, three weeks in, what had changed. She thought about it for a long time before answering.
“I used to spend a lot of energy on what I could not see,” she said. “Like, I always knew there were things happening in the room that I was not aware of. Tables that were turning in a direction I had not caught yet. I just had to accept that. I had to trust that if I moved fast enough and stayed present enough, I would catch most of it.”
“Now I do not have to trust that. I know what is happening. Not everything — I still have to use my judgment. But the things I used to miss because I was somewhere else? I do not miss those anymore. The system gets to them before I would have. It tells me when it needs me. Everything else — it just handles.”
She paused.
“It is a very strange feeling to realize how much of what I was carrying I did not have to be carrying. I thought that was just the job.”
What She Did Not Say
She did not say it was easy. The calibration took two weeks. There were signals the system weighted incorrectly for her specific room. There was a table configuration her dining room used that produced false positives for crowd contagion until we adjusted the propagation model.
She did not say it replaced her. Her judgment — the eleven years of reading a room, reading a person, knowing when a situation needed warmth versus efficiency versus a comped dessert versus a genuine conversation — none of that is in the system. None of it will be. That is hers.
What she did say, eventually, when we asked what she would tell another GM who was considering it:
“I would tell her that she has been working harder than she should have to. And I would tell her that is not her fault. And I would tell her that there is now something built for her, finally, that actually works for the job she actually has. Not the job someone imagined she had.”
That is what we built it for. That specific sentence.