WEBVTT - Danny Meyer Talks AI in Restaurants

0:00:02.759 --> 0:00:09.600
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news. AI's use cases, whether

0:00:09.680 --> 0:00:12.880
<v Speaker 1>it's sending way beyond the office. Pressy Taste AI, for example,

0:00:12.880 --> 0:00:15.040
<v Speaker 1>a company bringing a power of artificial intelligence into the

0:00:15.080 --> 0:00:19.360
<v Speaker 1>restaurant space, helping manage crews even predict demand. We're joined

0:00:19.400 --> 0:00:21.680
<v Speaker 1>now by the CEO of that company in goos Stalk,

0:00:21.720 --> 0:00:24.360
<v Speaker 1>as well as Union Square Hospitality Group founder Danny Matt,

0:00:24.400 --> 0:00:27.080
<v Speaker 1>an investor in Press Taste. Danny, I start with you

0:00:27.120 --> 0:00:30.440
<v Speaker 1>because inlightened hospitality investments is where you've made this investment,

0:00:30.680 --> 0:00:34.159
<v Speaker 1>but you've actually got a plethora of AI investments. How

0:00:34.280 --> 0:00:36.560
<v Speaker 1>and at what point did you decide the artificial intelligence

0:00:36.680 --> 0:00:38.360
<v Speaker 1>was going to change up your game of restaurant.

0:00:38.440 --> 0:00:42.320
<v Speaker 2>Here we're facing, as we always have in our industry,

0:00:43.560 --> 0:00:46.840
<v Speaker 2>really a business model where we have pretty thin margins.

0:00:46.880 --> 0:00:50.839
<v Speaker 2>In the restaurant industry is notoriously competitive. But I think

0:00:50.880 --> 0:00:53.040
<v Speaker 2>what a lot of people may not realize is that

0:00:53.479 --> 0:00:55.800
<v Speaker 2>by the time you're done paying your rent and the

0:00:55.840 --> 0:00:57.840
<v Speaker 2>talent on the team, and the food costs and the

0:00:57.880 --> 0:01:02.120
<v Speaker 2>insurance and the florists and all the expenses that only

0:01:02.120 --> 0:01:02.800
<v Speaker 2>go up over.

0:01:02.720 --> 0:01:04.240
<v Speaker 3>Time, you're not lived with much.

0:01:04.280 --> 0:01:08.760
<v Speaker 2>And so we're constantly needing to find ways that we

0:01:08.800 --> 0:01:13.319
<v Speaker 2>can become more productive, and machine learning, it turns out,

0:01:13.680 --> 0:01:16.760
<v Speaker 2>is really really good at a number of things. It

0:01:16.840 --> 0:01:19.800
<v Speaker 2>helps us to not only have much better data with

0:01:19.840 --> 0:01:22.640
<v Speaker 2>which we can take much better care of our guests.

0:01:23.319 --> 0:01:25.600
<v Speaker 2>Buy the proper amount of food because we never want

0:01:25.640 --> 0:01:27.680
<v Speaker 2>to run out of food, and on the other hand,

0:01:27.720 --> 0:01:30.319
<v Speaker 2>we don't want too much food, have the right amount

0:01:30.319 --> 0:01:32.880
<v Speaker 2>of labor and talent on the floor, never want to

0:01:32.920 --> 0:01:35.200
<v Speaker 2>have too little, you never want to have too much.

0:01:35.880 --> 0:01:38.880
<v Speaker 2>We don't necessarily have the ability to think about that

0:01:39.400 --> 0:01:42.760
<v Speaker 2>and be very very present to offer great hospitality for

0:01:42.840 --> 0:01:45.880
<v Speaker 2>our guests. And so the more that we can find

0:01:46.080 --> 0:01:52.520
<v Speaker 2>outstanding applications for machine learning AI, what it does is

0:01:52.800 --> 0:01:54.760
<v Speaker 2>we get to be a lot smarter, but we also

0:01:54.800 --> 0:01:57.480
<v Speaker 2>get to be better at hospitality.

0:01:57.560 --> 0:01:59.640
<v Speaker 1>Which is precisely where you want to be. Getting a

0:02:00.480 --> 0:02:04.560
<v Speaker 1>sticking to that kitchen knitting. Meanwhile, Ango, you have been

0:02:04.720 --> 0:02:06.800
<v Speaker 1>applying AI in the real world for a long time,

0:02:06.840 --> 0:02:08.680
<v Speaker 1>whether it be in academia and we moved across to

0:02:08.840 --> 0:02:11.760
<v Speaker 1>really wanted to change up the back of kitchen. What

0:02:11.960 --> 0:02:15.000
<v Speaker 1>drew you there? And how quickly have you managed to scale?

0:02:16.280 --> 0:02:16.560
<v Speaker 2>Well?

0:02:16.600 --> 0:02:19.520
<v Speaker 4>We have I have been in machine learning. I did

0:02:19.560 --> 0:02:22.840
<v Speaker 4>my pH in machine learning, when I still need to

0:02:22.840 --> 0:02:26.080
<v Speaker 4>explain everyone what that is. But now I'm you know,

0:02:26.520 --> 0:02:30.680
<v Speaker 4>since a couple of months, almost like just two years,

0:02:31.000 --> 0:02:37.239
<v Speaker 4>where CHET, Chipte or Google AI other CHET agents came

0:02:37.280 --> 0:02:41.880
<v Speaker 4>to market, it is now a big, big hype. Everyone

0:02:42.280 --> 0:02:44.799
<v Speaker 4>you know has AI capabilities now in.

0:02:44.760 --> 0:02:49.200
<v Speaker 5>A chatbot, but the chetbot, you know, they have their limitations.

0:02:49.440 --> 0:02:52.600
<v Speaker 5>You cannot have a drive your car, you cannot have

0:02:52.760 --> 0:02:57.640
<v Speaker 5>it prepared your meal. You need specialized AI capabilities such

0:02:57.680 --> 0:03:01.600
<v Speaker 5>as you know, using computer vision to identify what is

0:03:01.639 --> 0:03:06.040
<v Speaker 5>going on machine learning predictions to that breaks down. Okay,

0:03:06.040 --> 0:03:08.920
<v Speaker 5>what does this weather forecast and this POS forecast and

0:03:08.960 --> 0:03:09.639
<v Speaker 5>this trade.

0:03:09.440 --> 0:03:12.960
<v Speaker 4>Area information mean for my kitchen right now in those

0:03:13.000 --> 0:03:17.000
<v Speaker 4>fifteen million increments, so my crew can do and deliver

0:03:17.120 --> 0:03:21.720
<v Speaker 4>the best to optimize those razor thin margins that exist

0:03:21.800 --> 0:03:22.679
<v Speaker 4>in the restaurant AI.

0:03:23.919 --> 0:03:27.440
<v Speaker 6>For this specialized AI, that's what prec taste AI brings

0:03:27.440 --> 0:03:30.920
<v Speaker 6>to the table. And it's basically also the missing puzzle

0:03:30.960 --> 0:03:36.160
<v Speaker 6>piece why software couldn't come into the kitchens. Kitchens are

0:03:36.200 --> 0:03:38.960
<v Speaker 6>often still run like in the fifties, and AI is

0:03:39.000 --> 0:03:42.840
<v Speaker 6>the missing puzzle piece that can help to bring true

0:03:42.920 --> 0:03:45.160
<v Speaker 6>help to those kitchen operations.

0:03:45.280 --> 0:03:47.840
<v Speaker 1>Danny, you've got some interesting puzzle pieces from investments. You've

0:03:47.840 --> 0:03:51.400
<v Speaker 1>got seven shifts, seven rooms, precy Taste, AI Converse, now

0:03:51.920 --> 0:03:55.480
<v Speaker 1>your amount who's often gone forward breaking boundaries when it

0:03:55.520 --> 0:03:57.720
<v Speaker 1>comes to tipping culture, when it comes to talking to

0:03:57.800 --> 0:04:02.960
<v Speaker 1>your own workplace, how they reacted to artificial intelligence with

0:04:03.120 --> 0:04:04.880
<v Speaker 1>fair with excitement.

0:04:05.400 --> 0:04:10.280
<v Speaker 2>Oh, with excitement. Everybody who we hire is somebody who

0:04:10.360 --> 0:04:12.800
<v Speaker 2>is a learn it all, not a know it all,

0:04:13.400 --> 0:04:16.200
<v Speaker 2>and we're all looking for ways we can learn to

0:04:16.240 --> 0:04:18.839
<v Speaker 2>be better. If you look at the different companies that

0:04:18.880 --> 0:04:23.120
<v Speaker 2>you're showing right here, each one of these is providing

0:04:23.160 --> 0:04:26.080
<v Speaker 2>the opportunity to make us better at the thing we

0:04:26.080 --> 0:04:29.000
<v Speaker 2>were doing anyway. So there's nothing there that we were

0:04:29.040 --> 0:04:33.240
<v Speaker 2>not doing before. It's just much quicker, much smarter, much

0:04:33.279 --> 0:04:37.000
<v Speaker 2>more productive. And everybody likes it because once you learn it.

0:04:37.040 --> 0:04:39.600
<v Speaker 3>By the way, when you get a new toy, it's

0:04:40.240 --> 0:04:44.760
<v Speaker 3>if I get a new update to my Apple phone,

0:04:45.839 --> 0:04:48.920
<v Speaker 3>my iPhone, it takes me a day to get used

0:04:48.960 --> 0:04:50.640
<v Speaker 3>to it, but once I get used to it, I.

0:04:50.760 --> 0:04:53.680
<v Speaker 2>Get exactly why they made those changes. So I don't

0:04:53.720 --> 0:04:55.479
<v Speaker 2>want to say that the first day we put something

0:04:55.520 --> 0:04:58.960
<v Speaker 2>in the practice, everybody goes so hallelujah. We have a

0:04:58.960 --> 0:05:02.800
<v Speaker 2>new reservation syst we have a new scheduling system, we

0:05:02.880 --> 0:05:05.920
<v Speaker 2>have a new way that we can talk to our guests,

0:05:06.400 --> 0:05:08.200
<v Speaker 2>a new way we can get to know our guests,

0:05:08.720 --> 0:05:10.839
<v Speaker 2>and in the case of Presi Tastes, a new way

0:05:10.880 --> 0:05:14.600
<v Speaker 2>that we can actually deploy our talent to be as

0:05:14.640 --> 0:05:18.520
<v Speaker 2>productive as possible per task. That's the thing I want

0:05:18.520 --> 0:05:20.719
<v Speaker 2>to be clear about Prezie Tastes, which is really cool,

0:05:21.320 --> 0:05:24.599
<v Speaker 2>is that it knows all of our recipes. It knows

0:05:24.640 --> 0:05:28.279
<v Speaker 2>exactly what we sold, what we're selling. It can actually

0:05:28.360 --> 0:05:31.520
<v Speaker 2>see how busy our restaurant is, and it can give

0:05:31.560 --> 0:05:35.640
<v Speaker 2>the kitchen in real time the predictions about how much

0:05:35.680 --> 0:05:38.800
<v Speaker 2>food you should be preparing to replace what just got sold,

0:05:39.240 --> 0:05:42.279
<v Speaker 2>how much food you should not be preparing, who should

0:05:42.360 --> 0:05:45.200
<v Speaker 2>do it. And it's the kind of thing that managers

0:05:45.560 --> 0:05:48.919
<v Speaker 2>have done an okay job of year after year. But

0:05:49.040 --> 0:05:53.440
<v Speaker 2>now those managers instead are taking their time to taste

0:05:53.440 --> 0:05:56.920
<v Speaker 2>the food and make sure it's perfect before a sauce

0:05:56.960 --> 0:05:59.400
<v Speaker 2>goes out, or they're taking the time to be with

0:05:59.480 --> 0:06:01.080
<v Speaker 2>our guests.

0:06:01.360 --> 0:06:03.960
<v Speaker 1>So I appreciate having you both. Pressie Taco Ingo Stalk

0:06:04.040 --> 0:06:06.800
<v Speaker 1>Union Square Hospitality Group Danny mar This is Bloomberg.

0:06:08.440 --> 0:06:08.680
<v Speaker 6>Yeah, I