Connect With Causeway: How AI is Catering to the Restaurant Industry

Causeway Solutions

ConnectWithCauseway, AI

July 29, 2024

Connect With Causeway: How AI is Catering to the Restaurant Industry

Hungry for innovative ways to attract customers while controlling costs, restaurants are upgrading their processes and improving quality. AI is helping to redefine the industry – from automating processes to enhancing customer experience and making operational decisions. Tune in to learn how AI and data analytics are important ingredients for the restaurant industry’s evolution. Get a peek at how Causeway Solutions is using AI in day-to-day operations. Listen to the full episode.

Episode Highlights

  • One of the most interesting ways restaurants are exploring AI is by automating all sorts of processes from order taking to even cooking using AI. This can be seen at a restaurant in California called CaliExpress. They call themselves the first fully automated restaurant. They've developed the technology so that AI manages everything from ordering to cooking and even to serving. And their plan is to license this technology to other restaurants and make money that way.
  • One of the key strengths of AI and where it shines is data analysis. AI can generate insightful information about your data, except there's one small caveat and that is you have to guide it. AI is an army with no general. It's incredibly capable, but it just can't guide itself. What I mean by that is you have to be very specific in your requests when you want it to do something for you or else it won't really know where to go. But if you lead it properly, it can be incredibly powerful.
  • I used ChatGPT 4.0 for a test. We provided data from our monthly surveys, positioning ChatGPT as a restaurant consultant who was to look at this data, generate insights, and then also generate prescriptions for the restaurant to take action based on these insights. There were some good and bad results.
  • One of the interesting things that AI noticed about dining preferences is that a significant proportion of respondents prefer local restaurants over chain restaurants. This preference could indicate a trend towards supporting local businesses. The prescription it gave was that restaurants, especially local ones, should highlight their unique offerings and improve their community involvement. And chain restaurants might consider localizing aspects of their menu. This is a pretty reasonable prescription, and I think it could be valuable for a restaurant to know that this is happening.
  • Another interesting insight was on vacation planning. A segment of respondents is considering long distance or international vacations and the prescription was restaurants and tourist heavy areas should prepare for potential increases in foot traffic. Now this is also a useful insight and prescription. It is a bit obvious. So that points to some of the shortcomings, but nonetheless, it is a good insight and a useful prescription.
  • And then the last good, useful insight was about recent vacations, how a portion of respondents have taken a vacation recently, which could influence their dining habits. And the prescription it provided was that restaurants could capitalize on post vacation dining by offering special deals or themed menus that resonate with popular vacation destinations.
  • The real time saving is in the grunt work. Instead of having to sort the cross tabs and process the data yourself and get a good grasp of it, you can just lead AI along and have it do that. This took me about 30 to 60 seconds.
  • If it helps you to get beyond all that grunt work, it gives you that much more time to provide insights and applications. That's really what we want to get out of data. We really want to get answers.
  • We did an updated version of that survey back in April where we asked the same dining option questions to compare this year from last year. The customers who prefer chain restaurants were more likely than those who preferred local restaurants to say that they plan on making a major purchase like a car, a new appliance or a new technology within the next month.
  • There was a similar correlation with people who said that they dined frequently at fast food places, specifically, that they were also far more likely than the general population to make a major purchase in the following month, with even a quarter of them saying that they plan to purchase new tech alone the next month.
  • Those are some insights that only a person familiar with the questions and the topics can provide rather than just glancing at the answers or the cross tabs or just running it through AI. So those hours and hours of work still led to something good that ChatGPT didn't pick up right off the bat.
  • We are using AI at Causeway now. One of the general projects we have is natural language querying, which is essentially the ability to query data using English sentences as opposed to SQL or Python. The true value lies in enabling non-technical people to be able to generate insights themselves and interact with the data directly. That's never been the case before. You've always had to know SQL or Python to be able to interact with data directly. So non-technical people having direct access to data would be a major breakthrough.

Transcript: How AI is Catering to the Restaurant Industry

Podcast Episode 11, published May 14, 2024


Thérèse Mulvey, Vice President of Strategy:
Hello and welcome to Connect with Causeway. I'm your host, Thérèse Mulvey, Vice President of Strategy at Causeway Solutions. Today we are going to be talking about the hot topic of the day, AI, as well as one of my favorite pastimes, eating out. I'm confident when I say that we all love foods, but restaurants have not had an easy decade. From COVID to rising food prices and staffing issues, the restaurant industry is not for the faint of heart right now. So with some research in hand, and AI riding shotgun, we will see what we can learn today.

To help us with our topic, we're really excited to welcome another one of our Causeway Solutions colleagues, Nick Schroeter. Nick is our Marketing Analyst in our Digital Marketing Department, and he has become our AI expert. He is also a highly dedicated and very much appreciated friend of the Connect with Causeway podcast. And if that's not enough, he is our first Gen Z guest, which warms my heart. So welcome to Connect with Causeway, Nick.

Nick Schroeter, Marketing Analyst
Hi, Thérèse and Lauren; thanks for having me on. I'm excited to be the first Gen Z guest and to discuss AI with you.

Thérèse:
No pressure there, Nick, but you are representing your entire generation.

Nick:
I'm prepared.

Thérèse:
We're also joined by our millennial and Manager of Strategic Partnerships. Lauren Kornick. Welcome back, Lauren.

Lauren Kornick, Manager, Strategic Partnerships
Hey, Therese, I'm proud to represent your third favorite generation.

Thérèse:
I'm so happy that everybody has a place. All right, let's start by looking at the state of the restaurant industry. Lauren, what did you find and what does it look like?

Lauren:
We mentioned the restaurant industry when we were talking with Hailey Russell about vacationers. We had food vacationers and Hailey said there were 72,000 fewer restaurants than there were in 2019 compared to 2022. So pre-COVID versus post-COVID.

But by 2023, there's a bit of good news. The year-over-year traffic growth improved for restaurants for the last quarter of the year, but that growth was only for fast casual dining establishments. You know the type. When compared with full service fancy restaurants, that category either stayed stagnant or even decreased for that 2023 time period.

On top of fast casual dining, local restaurants seem to be the way to go. Almost two thirds of all diners said when they do choose to go eat out, they chose local over chain restaurants. So, there's some good news and some not so good news for the industry as a whole. But restaurants are still looking for ways to optimize their value in the marketplace to see growth post pandemic. They're also fighting with things that seem to be out of their control, especially inflation nowadays.

Thérèse:
Yeah, I did some research, too, because I'm always so frustrated when I go into a restaurant, and I see so many empty tables. I just don't understand the staffing issue. But as you mentioned, Lauren, COVID was a huge influence. Three years later there's still a 3% shortage for jobs in the category and that actually represents half a million jobs. That's where you've got the restaurants who have availability, but if they don't have anyone to wait on you, it's not going to be much fun.

Surprisingly, the pandemic not only influenced workers' perceptions, but it also motivated them to seek change. Especially in an industry that’s so demanding both physically and mentally, a lot of people who were in restaurants, once they could go back to work, chose not to go back to work in that industry.

And then as we begin to look at both restaurant trends and the topic of AI, I wanted to see what experts are saying. What are the crucial areas for operators to look at. Restaurants have always been a tough business, but particularly now, what do they need to look at? The three things they talked about was optimizing service, clearly there's opportunity there for AI, attracting skilled personnel, and monitoring performance. One in three are definitely a clear opportunity for AI integration for this category.

Lauren:
Yeah, and that's a clear opportunity for AI integration right here.

Thérèse:
Exactly. So, let's get to our AI expert. What kind of AI are we talking about, Nick, when it comes to this industry as well as AI in general, and how prevalent is it?

Nick:
It's quite an interesting example as restaurants are not what you would usually think AI could be applied to. One of the most interesting ways restaurants are exploring AI is by automating all sorts of processes from order taking to even cooking using AI. This can be seen at a restaurant in California called CaliExpress. They call themselves the first fully automated restaurant, which is a pretty big claim. They've developed the technology so that AI manages everything from ordering to cooking and even to serving. And their plan is to license this technology to other restaurants and make money that way. It's quite an interesting idea and I'm curious to see if other restaurants will actually use this technology.

Thérèse:
Well, that's a great example and I must say that I would never have thought of that. And the big question is can I get AI to cook my dinner at home? That might be something we explore in our next episode. So that's an interesting way of looking at how restaurants might use AI for automation. How are we using AI at Causeway, Nick?

Nick:
Very differently. One of the key strengths of AI and where it shines in my opinion, is data analysis. I use it for data analysis weekly, if not daily. I know a number of other people who do as well. AI can generate insightful information about your data, except there's one small caveat and that is you have to guide it. That's where I came up with this phrase that AI is an army with no general. It's incredibly capable, but it just can't guide itself. What I mean by that is you have to be very specific in your requests when you want it to do something for you or else it won't really know where to go. But if you lead it properly, it can be incredibly powerful.

Lauren:
What we're wondering today is if the restaurant industry can use AI, beyond just the automation example that you gave, Nick. Can they utilize the data analysis aspect of AI as well?

Thérèse:
I saw an eMarketer article. In general, data analysis and market research have the greatest acceptance right now for using AI, which makes a lot of sense. But the automation example that Nick started with, I never even thought of that piece. Obviously, that is an important piece. But I do think if you go back to what the experts were saying that restaurants need to do, I can also see where AI would be a really great help in terms of them understanding whatever data they can gather on their consumers, on their business, and having AI find the insights for them.

We did a test on this using our monthly survey. We have asked about restaurants in the past and we asked about them again. Nick used an AI tool to look at last year’s research on restaurants to determine if what we had seen from an insight perspective could be understood by AI. Nick, what did you find? Are we all out of a job here?

Nick:
Not yet.

Thérèse:
Okay, good to know.

Nick:
But it's a fair question. A little background about the test. I used ChatGPT 4.0, which is the paid version. I provided the cross tabs data to ChatGPT, positioning ChatGPT as a restaurant consultant who was to look at this data, generate insights, and then also generate prescriptions for the restaurant to take action based on these insights. There were some good and bad results.

I'll start with the good results. One of the interesting things that AI noticed about dining preferences is that a significant proportion of respondents prefer local restaurants over chain restaurants. This preference could indicate a trend towards supporting local businesses. The prescription it gave was that restaurants, especially local ones, should highlight their unique offerings and improve their community involvement. And chain restaurants might consider localizing aspects of their menu. This is a pretty reasonable prescription, and I think it could be valuable for a restaurant to know that this is happening.

Another interesting insight was on vacation planning. A segment of respondents is considering long distance or international vacations and the prescription was restaurants and tourist heavy areas should prepare for potential increases in foot traffic. Now this is also a useful insight and prescription. It is a bit obvious. So that points to some of the shortcomings, but nonetheless, it is a good insight and a useful prescription.

And then the last good, useful insight was about recent vacations, how a portion of respondents have taken a vacation recently, which could influence their dining habits. And the prescription it provided was that restaurants could capitalize on post vacation dining by offering special deals or themed menus that resonate with popular vacation destinations.

And again, I thought it was a good prescription. We all know at least where I come from in New York, everybody goes on vacation in the same place. Is that the same for you in Connecticut, Therese?

Thérèse:
Yes, we have a lot of beaches that we go to and we're pretty happy with them.

Nick:
So, I think that's a pretty decent idea for a restaurant to use this insight and the prescription was reasonable as well.

Thérèse:
It makes sense. So, let me ask the big question, what was the time saving, given that Lauren and I know what it took to look at all this information last year?

Nick:
The real time saving is in the grunt work. Instead of having to sort the cross tabs and process the data yourself and get a good grasp of it, you can just lead AI along and have it do that. This took me about 30 seconds, 30 to 60 seconds to do.

Thérèse:
Oh my gosh.

Nick:
... I'm no data expert, but if I had to look at the cross tabs, I would say it would take me at least a few hours to process it and I would probably miss things. The AI doesn't miss like we do.

Thérèse:
Wow, 30 seconds? That seems too good to be true. And where were you 20 years ago when I was knee-deep in this stuff? All right, you might as well bring us to the bad news, Nick.

Nick:
You're right, it is a little too good to be true. One of the insights it provided was a distribution among respondents who are spending less, more, or the same amount. Basically, saying that everybody spends a different amount at a restaurant. And the prescription it offered was that restaurants should cater to varying economic conditions by offering diverse menu options. Now the problem with this one is the insight is true that there are a lot of different spending patterns among respondents, but the prescription is completely absurd. Can you imagine going to a restaurant that had a $500 tasting menu and a $12 burger on the same menu? It would be absurd to say the least.

Thérèse:
I'm not even going to a $500 tasting menu.

Nick:
Yeah, neither am I, but it would be a weird restaurant that has both of them. This is a great example of AI not knowing how to lead itself or having that human context. It really is the army with no general.

Thérèse:
I do love your saying.

Lauren:
Yeah, the AI restaurant aspect of it all aside, it's still a pretty good starting point. Some of those insights that the AI found from the survey match the ones that Thérèse and I and the other analysts made on our own. We had the customers who prefer local versus chain restaurants or even the food vacationers we mentioned previously. There is a match here, so it's a really good starting point.

Thérèse:
Yes, it's a great example of how to use it. I also think understanding how to use it is enhanced when you also understand how to not use it. I really appreciate the example of where it didn't work. You do have to be careful, and you have to stay involved. But if it helps you to get beyond all that grunt work, it gives you that much more time to provide insights and applications. That's really what we want to get out of data. We really want to get answers. We don't just want to crunch numbers even though sometimes it seems like we do.

Nick:
Absolutely.

Lauren:
Yeah. And there'll still be some insights that we'll get on our end that ChatGPT might not pick up. The robots aren't taking over.

We did an updated version of that survey back in April where we asked the same dining option questions to compare this year from last year. Overall, there wasn't much change in opinion from this time to the previous time. And rerunning the updated survey through ChatGPT resulted in the same insights that Nick just read out. But from a human touch, we're able to look beyond those answers and gather some additional insights that we didn't look at last time or that ChatGPT didn't even pick up this time. There are some other interesting behaviors we found.

The customers who prefer chain restaurants were more likely than those who preferred local restaurants to say that they plan on making a major purchase like a car, a new appliance or a new technology within the next month. There was a similar correlation with people who said that they dined frequently at fast food places, specifically, that they were also far more likely than the general population to make a major purchase in the following month, with even a quarter of them saying that they plan to purchase new tech alone the next month.

Those are some and insights that only a person familiar with the questions and the topics we were searching for or knowing the specifics of what to search for can provide rather than just glancing at the answers or the cross tabs or just running it through AI. So those hours and hours of work still led to something good that ChatGPT didn't pick off right off the bat.

With that in mind, Nick, how do you think restaurants are going to use AI in the data analysis way?

Nick:
Quite honestly, I don't think they will. There are massive costs associated with large language models, LLMs as they're referred to, which is what ChatGPT and most AI chatbots are. And local small business restaurants are a low margin business. There's just not enough free capital available to do that. And on top of that, how much data can a local restaurant really collect? It's really not such a massive amount that it requires an AI to use it, if they even collect any at all. On the flip side, I would say big chain corporate restaurants, if you can even call them that, like McDonald's, Burger King, Shake Shack, they will certainly be using AI because they collect a massive amount of data, they do enormous amounts of marketing and have huge budgets for these things. So, I see the big guys using it and the little guys probably not.

Thérèse:
That's really interesting, Nick. I think restaurants are a particularly good category to talk about because they are so different. I also agree with Lauren, which segues back to the comment, your comment of army with no general, but what does this mean for the world of AI? What does it mean for how people should think about AI using it in daily life? I know you were my bridge into AI. I didn't think I would ever use it, even though I already use it with Siri and voice commands. I have started using ChatGPT for help with some things that have saved me a great amount of time, I didn't even do that three months ago and now I probably use it a few times a week. What do you think this means about using AI in your daily life? What's the latest that people are talking about when it comes to this topic?

Nick:
As we've seen, AI is incredibly powerful, but it lacks that human context necessary to guide itself. One of the cool things that is being talked about a lot now is AGI, which is artificial general intelligence. The promise is that it will have the ability to lead itself. And that's the point where it takes over the whole world, Lauren, like you were mentioning before. And that is about five years away according to OpenAI CEO, Sam Altman. Since he's closer to this technology than anybody else, he's probably got the best guess.

But in terms of what people should be thinking about, I think while AI is so new and it's very accessible at the moment. ChatGPT, the older models being free and 4.0 being a very small cost, it's worth trying out, using it and getting to know it as it will only become more popular. The better acquainted with it you are, the better off you'll be. And not only that, it can help you now. It's not as powerful as it will be, but it helps me all the time already. My one recommendation would be just to try it out and experiment with it.

Lauren:
And it's ramping up in multiple fields and even in fields and industries that I don't think you'd expect. We're talking about restaurants, but meanwhile our friends over at Canopy Advertising just held a panel with our team at Causeway, along with Angela from RevOpp AI at the Argentum Senior Living Executive Conference. They were discussing similar applications, but more specifically in the industry of senior living. Restaurants is one thing, and senior living is a whole other topic, and these are definitely not industries we'd automatically think about utilizing AI in. Canopy also did an episode on their podcast on Canopy IQ with Angela Urbaczewski of RevOpp AI and from the Argentum panel. They were talking about AI in this episode about was the idea that AI is not a storyteller, it's only really helping you to tell the story.

Thérèse:
Yeah, I did think it was a really interesting point that Angela made, that AI is working for you, not instead of you. It's really what we've been saying, just a different way of saying it. So it's making parts of your job quicker and easier, but you still really need to be the manager or the thinker.

In this situation at Causeway, hypothetically, combining processes with AI, it could help facilitate our audience creation. We get actionable insights to follow up on ourselves that can cut initial research time down. The next time we start our process of audience creation, like customers likely to prefer a local chain restaurant, we have that foundation to start with to use AI to cut some of the research time in trying to expand that audience. Provided we give it the required guidance, there's a lot of opportunity. The interesting thing I thought about the Canopy story was that AI actually gave the client, because it allowed them to use their data better, it gave them more time with their residents. AI actually increased the personal part, which is also something you'd never think about.

Lauren:
Nick, I know we talked about, "Oh, we're utilizing AI at Causeway," and without giving away any top secret plans, can you give us many more insight on how we're specifically using it or what's our thought process in using it?

Nick:
We are using AI at Causeway now. One of the general projects we have is natural language querying, which is essentially the ability to query data using English sentences as opposed to SQL or Python. This is incredibly powerful. It would cut down time to generate insights, which would be awesome, but I don't think that's the true value in it because we have some amazingly talented SQL experts here at Causeway who are very quick and smart. I think the true value lies in enabling non-technical people to be able to generate those insights themselves and interact with the data directly. That's never been the case before. You've always had to know SQL or Python to be able to interact with data directly. So non-technical people having direct access to data would be a major breakthrough.

Thérèse:
That sounds good to me 'cause I'm one of those non-technical people that wants to interact with data. I'm sure it's also something we really need to be careful of too. So as always, there's two sides.

Lauren:
Therese, I think you'd be an amazing general to the AI Army in that case.

Thérèse:
Oh, thank you, Lauren. You know how that makes me so happy. Well, I want to thank you both for your time and your insight. I decided to do a little experiment on my own and turn to AI to write something for me to end this podcast. I told AI what we were going to be talking about, a few of the key things I needed in my wrap up. And I'm going to read you what it came up with, and then I have a couple more comments.

Thank you for joining us on today's episode of Connect with Causeway. A special thanks to Nick and Lauren for their insights and contributions. Today's discussion not only highlighted the current state of the restaurant industry and the unique challenges it faces, but also the exciting opportunities AI presents. From automating processes to enhancing customer experience and making operational decisions, AI is proving to be a valuable ally in navigating these complex times we're living in. As we've explored AI functions best with clear direction, there comes the general, it is a tool that extends our capabilities, but does not replace the nuanced understanding and strategic thinking that only humans can provide.

See, my fear is that in five years, the AI bots are all going to be laughing about the fact that we thought that.

Whether you're in the business of food marketing or any other field, remember the future of AI is about collaboration, not replacement. See, there they go. Making us comfortable again. Join us the next time as we continue to dive deep into the trends and technologies shaping our world. Safe travels through your digital landscapes, and until next time, keep connecting with Causeway.

Lauren:
That was so good.

Thérèse:
I know. Keep connecting with Causeway. Why didn't we think of that? And maybe our next podcast will really just be three virtual people talking with better grammar.

Nick:
Yeah.

Thérèse:
Anyway, thank you both from this human's brain with no AI help.

Lauren:
And the two human guests.

Thérèse:
For my two human guests, I really enjoy interacting with humans, and I do like using AI every now and then. For all the humans listening, as well as any AI people listening, be sure to send us your feedback. Please subscribe to our podcast and tune in again for our next episode.

To learn more, visit Causeway Solutions to get started!