I’ve run the floor.
Now I build the AI behind it.

A friend of mine, Jay, taught me this in a casino. Every server who came by, whatever they had done, he thanked by name. It made an impression every single time. I have done it ever since, and I noticed something else: people started remembering my name too.

As a Royal Genie I looked after the guests who paid the most to be on the ship, and I never brought them anywhere unannounced. Before we walked in, the head of department or the floor supervisor already knew who was coming. The first thing my guests heard was “Welcome,” followed by their name.

It pleased them every time. The attention to detail is the whole thing.

When I owned my restaurant, I asked my servers to start every table by noticing one thing. Nice shoes. I love your hair. That ring is pretty cool. It hands the guest something to talk about, and often it turns out to be something you both like.

On the ship I did the same with jokes, usually stupid ones. A guest would say, “I did that once, but it was a long time ago, maybe fifteen years.” I would say, “Oh, so when you were eighteen.” They were clearly around fifty.

The small jokes are the ones that leave a mark. Everyone loves a sense of humor.

The first time I got this wrong, I was working in my dad’s restaurant. I was carrying a large tray of water and spilled all of it over a guest. I was embarrassed. I could not apologize enough, and nothing I said helped.

Years later, on the ship, a man screamed at me because too few of the crew spoke his language. There was nothing reasonable in it. This time I let him finish. I left a long, deliberate pause. Then I told him his feelings were understood, and I used my four magic words: I’m on your side.

On a good day, listening works the same way. I listen for what people like and what they don’t, because that is where the small gift comes from, or the note with the inside joke.

At butler school, one of the lessons that stayed with me was how to word things. Sometimes you have to tell a guest that something costs money, without making them feel you are taking advantage, and without implying they didn’t already know. My teacher’s phrase was that it is “of a chargeable nature.”

I don’t talk like that, so I don’t say that. I say, “I’ll also find out what costs are associated with it.” The same information, in my own voice.

The phrase isn’t the point. With the right language you can steer almost any difficult conversation, and the care you took is what people remember.

My guests on the ship paid between $15,000 and $70,000 for a week. They wanted to feel like kings, and who was I to say otherwise. So I never told them no. I told them I would look into it, even when I was sure it couldn’t be done. The reality is that most things can be bought.

One guest wanted a very expensive bottle that wasn’t on board. I found one in Miami. We were in the middle of the sea, on our way to the Bahamas. Rather than call it impossible, I priced it: the helicopter, customs, everything it would take to get that bottle to the ship. It came to around $30,000.

I knew they would decline. But they were the one who decided. They didn’t feel rejected, and they knew I had tried.

I’m Chinese, and I was born and raised in Jamaica. The accent does not match the face. I turned that into a game: something on me if you can guess my nationality before the end of your stay.

I played it at the front desk of a luxury hotel in North Carolina. When someone got it right, I honored the bet with a comped drink or an amenity.

It cost me very little. They remembered me, and they tipped better.

The fastest way I know to bond with a stranger is to know something about their home. I would talk to Polish guests about pierogi, Russian guests about their Easter traditions, Swedish guests about the stinky canned fish they sell.

Then I ask my favorite questions. What’s it like over there? Do you miss it? Did you love it growing up?

People will talk about where they are from for hours. Somewhere in those hours is everything you need to know about what they like and what they don’t. Culture, food and travel never fail.

Even when you are doing everything right, people have bad days. They are tired, and they keep talking to you out of courtesy.

I took one habit from the book What Every Body Is Saying and I use it all the time: I watch where the toes point. If I am in a long conversation and their feet have slowly turned away from me, I take the hint before they have to give one. “I can imagine you might be tired.” Or, “Anyway, let me get out of your way.”

They appreciate it.

On the ship, a couple’s friends pulled me aside during an afternoon cocktail hour. The two of them had been together fifteen years, they said, and had not had a proper date night, just the two of them, in longer than anyone could remember. Kids, work, life.

The couple had not asked me for anything. I spent the next two hours quietly learning from their friends what they loved, and what they would choose if nobody was watching.

By evening their suite was set. Candlelight. A menu built from what I had learned: not what they ought to eat, but what they actually wanted. No crew hovering. I appeared when I was needed and vanished when I wasn’t.

When they saw it, her eyes filled and he reached for her hand. When I came back to clear the plates, they were present with each other in a way they had not been all week. That dinner was not on any menu.

A restaurant owner came to me with a problem. Wednesdays were dead, and the guests who did come in, mostly women in their twenties and early thirties, ordered beer and vodka sodas. He had tried happy hours and two-for-ones. Nothing moved.

I had watched the same thing for years in fine dining and on the ship. Guests would hover over the wine list, hesitate, and order something safe. They didn’t dislike wine. They didn’t feel confident ordering it.

So we built Wine Wednesday, and it was not a discount night. Three wines each week: one old world, one new world, one wild card. Ten minutes from me at the start of the night on how to say the names, why each one tasted the way it did and what to eat with it. Then tastings at the bar, no pressure.

Within two weeks, Wednesday revenue was up 250%. Groups came for it, brought friends and started ordering bottles instead of glasses.

Things go wrong on every shift. A cooler runs warm. A table waits too long. Someone calls out. None of that is what sinks an operation.

What sinks it is the handoff at the door, the “I thought you had it,” the issue everyone assumed was resolved. Problems get buried under the pace of service, and they come back bigger.

So I ask three things of any operation I run. Every event is captured. Every issue has one person responsible for it. Every case is closed with proof that it was dealt with. Visibility, ownership, resolution.

“Fridge issue in kitchen” tells the next manager nothing. Compare it with this: walk-in cooler reading 48°F at the start of the PM shift; maintenance called; temperature-sensitive items moved to the backup unit; a health risk if it isn’t fixed before tomorrow’s service.

The second one answers four questions. What happened? Where? Who was involved? What was the impact? Anyone who reads it can act, including someone who was off that day.

I ask the team to write it before they move on to the next thing. A note reconstructed three hours later has lost the details that made it useful.

The weakest word in operations is “handled.” It closes the conversation and tells you nothing.

A real ending says what was done, who did it and how you know it worked. Maintenance replaced the thermostat. The unit is holding at 38°F this morning. The kitchen team confirmed it. Now it can be closed.

That last note matters more than any other. If the same thing happens six months from now, the next manager will know what was tried and whether it held. And waiting is no excuse for silence. “Parts arrive Thursday, technician booked Friday” is a healthy case. Three days of nothing is a forgotten one.

I use one test at the end of a shift. Could the incoming manager sit down, read for three minutes and know exactly where things stand, without me in the room?

If not, the handoff isn’t finished. What I say on my way out will be gone by the first rush. What I wrote will still be there next week, for the next manager and the one after.

It takes five to ten minutes at each end of a shift. Before service: what is open, what is overdue, what has no owner. After: everything major is written down, every open issue has a current note, and nothing lives only in someone’s head.

A guest with a problem should not have to wait while someone finds a manager. By the time permission arrives, the moment to fix it has passed.

So in my service playbook, every team member has the authority to act on the spot: replace a wrong or late dish, offer a drink, move the table, send a dessert, write a note by hand. No approval needed.

It only works if people know exactly what they are allowed to do, and know I will back the decision they made.

The first time something goes wrong, it is an incident. When it shows up three times, across different shifts and different people, it is no longer anyone’s mistake. It is a gap in the process.

The wrong question is “why did this happen again?” The useful one is “what do we need to change so it stops happening?” Usually the answer is a procedure nobody has written down yet.

I wrote more than twenty-five procedures for my last opening. A procedure earns its place when it answers a problem that kept coming back.

If people think that reporting a problem will get them in trouble, they stop reporting. You don’t end up with fewer problems. You end up blind.

Two things fix that. First, record the good as well as the bad: the compliment, the recovery, the server who handled a hard table well. A record that only holds failures feels like a file being built against you.

Second, act on what people tell you. When someone flags a broken door and sees it fixed by Thursday, they flag the next thing. When their note disappears into a void, they stop. Follow-through is what builds the habit.

Every operation has someone who marks everything urgent. After a week, nobody can tell what actually needs attention tonight.

I keep four levels and ask people to use them honestly. Low is minor or already dealt with. Medium needs attention soon. High is hurting service, safety or the guest right now. Critical is a genuine emergency.

The other half is noticing small things early. Equipment running differently than usual. A stock level that looks thin. A shift that is always stretched. None of them is a problem yet, which is exactly when they are cheapest to fix. The best operations catch it before it reaches the guest.

When I write a closing procedure for cash, the count is only part of it. The step that matters is the control: the net cash on every register’s report, added together, has to equal the net cash on the system report. If it doesn’t, you stop there.

Then the procedure says what to do about a gap. A small variance: recount once and note the reason. A larger one: recount twice, document it and tell the manager on duty.

It ends with a short checklist someone can follow at midnight when they are tired. I test every procedure the same way. Could a new hire run it alone, and would it catch its own mistake?

Reviews and surveys tell you about last week. The floor tells you about tonight. Slow tickets, dishes running out, complaints, recoveries: operational problems show up there first, before they reach your ratings or your revenue.

So I track the early numbers. Not only the average ticket time but the slowest one in ten. How many items ran out during service. How many hours a manager actually spent on the floor.

At my last opening I went a step further and built a score from the shift record itself, a read on guest satisfaction designed to move before the survey does. Then I bring it to the table. Replace “how did the week go?” with the record, and the conversation goes from opinion to specifics.