Thanks, everyone, for coming here. We appreciate everyone making the time and coming through the Atlanta traffic. I'm John Kutay. Today we're doing a live recording of the What's New in Data podcast. We talk about latest trends in both data AI from a practitioner's perspective, even up to the industry trends.
Super excited for our guest panelists today. We have Brian Chambers, chief architect at Chick fil A. And we have KC Rakam, head of customer engineering over at Google Cloud. Thank you to both of you for joining. So I'll just go ahead and start off. Uh, Brian, introduce yourself to the audience. What do you work on?
What do you do? Yeah, so as you just heard I'm the chief architect or leader of the enterprise [00:01:00] architecture practice at Chick fil A. Been there about 20 years. And we are really responsible for figuring out is the business well supported by technology, and what are the things that we can do What are the technologies we need to make investments in?
Uh, what do we need to R& D and explore? So that ultimately we can have the right foundation for all the different engineering teams and teams working on data and all those things to build on top of. So, that's what I do. Thank you, Brian. Casey, same question to you. KC Rakam, been with Google Cloud seven years now.
So, I'm responsible for data and AI customer engineering at Google, predominantly focusing on our strategic retail. and consumer goods. So one of the key things that we work with our customers is helping them on board as well as make the use cases successful in terms of deployment on Google Cloud when it comes to data analytics and AI across the board.
Yeah, and the use cases and business value of all these data initiatives is super critical to prove out the value and ultimately make [00:02:00] sure that they're good investments both for the company and the, and the customers that they're serving. That being said, you know, we'll get into some, some cool real world implementations.
Brian Chambers, I came across your work. A lot of people on Twitter were pointing me to it. It's a very popular thing that you built at Chick fil A. I'll let you talk to the audience about what you've built there. Yeah, sure. And I think we're talking about at the edge in our restaurants, specifically.
So, I already used the buzzword edge computing. To me, that means putting the resources for compute as close to the users and the user experiences as is necessary to meet your objectives, but no closer. And so for us in a restaurant business, that's really busy. We were really focused on making sure that we had continuity for the solutions that we're ultimately going to.
Exists in the restaurant, and that led us into our version of Edge Compute, which if you're familiar with it it's, we have a stack of Intel NUCs that run in the back of every one of our restaurants. We run Kubernetes on top of that in the [00:03:00] form of K3s which is an open source project from Rancher.
Which is really great and kind of simplifies and cuts a lot of the things that are cloud related out of the project. Makes it a little bit easier to manage. It's a single binary. So we run that on top and then we have a bunch of applications ultimately that run inside of Kubernetes that let us do things that we want to do.
So a lot of our usage and the reason that we went down this road actually came out of starting to explore IoT. For our restaurant. So we kind of had a vision for connecting all of the different things that exist in the restaurant. That would be the fryers and the grills and the things that do the cooking.
The holding cabinets were things like how long is something in holding matter. And on and on. So we started to connect things and we knew pretty quickly, Hey, we don't want to make all of these devices intelligent. Like send them data from other things and have this giant mesh of things flowing everywhere.
We want to centralize the intelligence and the decision making in a place where We have kind of like a normal developer experience paradigm which, you know, trying [00:04:00] to mirror the cloud as much as possible, that ended up being this edge environment. So that became a place where, and that's what our applications basically that look like, that run in Kubernetes.
They're consuming telemetry data and events things from the different kitchen equipment, devices or other applications that run on tablets that team members use inside the restaurant. You know, crunching numbers doing things with the data that's collected and then making decisions that.
Usually for us right now, are more about displaying an action to somebody. Not really like automating an actual behavior that a human would have done in the past. But perhaps there's places we do that in the future. So that's a little bit about what we built. Oh wow, so you touched on an interesting point there, which a lot of companies are talking about, which is...
You know, empowering people with data, not replacing people with data, and essentially having that human in the loop that's more informed and has more insights. So, you know, my question to you is, you know, what's the value to the Chick fil A customer that they're getting from you centralizing all the data in the, in the cloud?
Yeah, I mean, hopefully it translates into things that we've always tried to do really well, which would be [00:05:00] provide really high quality food with really great service at a very rapid pace. Like, we don't want you waiting a long time, we want your food to be hot and awesome and you enjoy every bit of it, and we want the people who serve you to be able to do that with kindness.
Now, if you think about things that could detract from that, they could be things like... This is so stressful. I just don't have the energy to smile at you right now. And so we want to try and alleviate those things by empowering people. It could be that we have really big hold times because we focused on cooking the wrong thing you know, based on what's happening in the restaurant at any given moment.
And so we're trying to address that with better information based on reality, incoming demand, what's in work in progress, all those kinds of things. So that those people can focus on the right stuff. So really it comes down to, at the end of the day, all this technology is really about making sure that people, both the customers as well as the people who work there, have great experiences and ultimately Chick fil A gets to have a positive influence on them and on their life.
Absolutely. It's, it's such a story that a lot of people resonate with [00:06:00] and when people listened to the previous episode of What's New in Data that you joined and you were talking about. The implementation. It's always sort of mind blowing to them that, you know, there's these IOT devices in every single Chick fil A that they drive by or literally drive through to get their food.
And to see, you know, data come back and make the consumer's life better is one of those inspiring things that a lot of people are able to actually experience if you go to a Chick fil A from Brian's work. Yeah, I mean, I think what's really important is that These kind of things don't have to be about creating something that's never existed before and something completely fundamentally new.
It could just be about making the things that are already critical to your business better. And I think that's really important to remember. Data can be something that equips you to just do the essential stuff even better than you could do it in the past. It can create new experiences and new things that you couldn't do before.
And that's, you know, part of our reality and our hopes for the future. But just being able to do the stuff that's really important and essential even better is. It's totally worth it. It's a great investment. Absolutely. And, and Brian, you have incredible [00:07:00] depth on the, on the value you're bringing to Chick fil A.
And now I'm going to pass it to KC Rekam, who has incredible breadth working with all types of great customers at Google Cloud. KC, if you could talk about some of the amazing cutting edge use cases that you're working on there. Yeah, yeah, absolutely. So as you all probably are, are seeing, there is a huge explosion in especially around AI.
And leveraging on the foundational data, right? How can you build A. I. Power products? I think working with a lot of our customers, especially in the retail and C. P. G. We see more specifically around how can you leverage generative A. I. To build products? A. I. Power products. So there are three core foundational use case categories that we see a lot of customers actually building, right?
So one is consumer and customer experience. This is more focused on how to How to drive from a retail perspective, how to drive the search and the recommendations to eventually [00:08:00] convert them into like a purchase, right? I think that is more focused on like, you have the traditional products, you know, you have product attributes, but now with generative AI, you can actually create new set of product attributes that can help the search journey better for the users who are on the e commerce website, right?
That is one area that we're seeing a lot of explosion where, you know, that's translating into actually net new revenue for the customers. Along with that is AI powered chatbots, which is whether it's like as simple as, hey, I want to learn better about a particular product the the one of the retailer has on their website.
Or it could be where using natural language, you can actually say I have a project going on. You know, what do I what do I need? Right? The A. I. Chatbot with the generative. I now provides not only all the list of products that you can go by where to buy how to buy, but also how to put everything together.
So that's a that's a great value for a customer where now you're not only seeing [00:09:00] that the translation from a sale for retailer is much faster, but also for the customers. They're seeing a lot of value in terms of how they're engaging the mechanism of engaging. So that's more on the customer and consumer side.
The second is on the marketing automation, where we're seeing a huge automation, especially how the marketing campaigns are being driven anywhere from content creation, like, you know, creating new content for the marketing, you know, personalization more around like content personalization, specifically being able to see who the customer is or a user is and be able to cater that message directly to that marketing automation.
That is, you know, super . Key, especially, you know being able to recognize who the who the actual user is coming in and personalizing that. And then the second is actually the campaigns itself on the marketing automation where the way you can now with using generative AI, how can you [00:10:00] actually draft like blog posts to social social media posts all the way to from a customer service perspective where you can actually respond to an email with a response back, which is more context aware is, is super critical.
So that is where we're seeing a lot of adoption in the marketing automation side. And then the third area where we're seeing a lot of customers leveraging generative ai, especially AI, is around the developer experience around the code code creation, code assist. Explainability of a code, right?
These and then also being able to use natural language to create a new products. It's as simple as like, you know, report generation could, they used to take months. Now it can happen in minutes to an hours, right? So that's, you know where we're seeing those are the three areas where we're seeing a lot of adoption.
And this is where I think the customers are looking at and adopting the modern technologies, especially with generative AI. Right. Thank you [00:11:00] so much for that, KC. And you know, this all really comes back to this, this concept of the customer 360, knowing everything about a customer. But, you know, as we see working in data, that's a moving target, right?
It's always kind of changing, like all the information that we have about a customer and all the ways we can action it. So KC, I have a question for you, which is how does AI help us have a better view of the customer and then automate those actions that you were talking about? Yeah, I think any for any AI fundamentally data is is super key, right?
I think AI is only as good as you know, the data it is being trained on. So that is that is super critical. So a lot of times embarking on AI project fundamentally is making sure that you have all the foundational data in place before you can actually start you know, powering these I AI use cases.
I think I recently Read a, a tweet that the most the most popular programming language now is actually English [00:12:00] which is, which is very interesting if you think about it, because now the natural language has made it so easy. Like, you know, you think about like, code generation is like you know, I had become simple.
It doesn't mean that, you know, it's at 90 or a hundred percent accurate. It's not accurate. There is hallucination and all that, but still, you're now able to have a jumpstart on what you're trying to do, right? I think that is super critical in terms of when you're embarking on AI specific projects. It's like, you already have the foundation, but key thing is, like, make sure that, you know, you have the fundamental data in place.
Absolutely. And, you know, we're talking about things like text to SQL, SQL to text, raw data to embeddings. You know, at Striim, you know, the work we're doing around data streaming, we're working with a lot of customers on this, is something very innovative, which is capturing the raw data, automatically generating those embeddings, so that something like, you know, Vertex AI and Google Cloud's vector extensions can index that and turn that into a real time AI store, right?
Real time chat experience, things along those lines. And what we're seeing is [00:13:00] that... for even though A. I. Is is revolutionary in terms of its its value to the world in the economy. It's really evolutionary for data teams, right? You look at the skills required. You know, their extensions to existing databases existing data pipeline tools.
So that being said, you know, and I'll ask both of you this. You know, what's your advice to data teams that are looking at adopting A. I. Brian, I'll start with you. Yeah. It's definitely a rapid changing space, right? There's a lot of uncertainty. So I was listening to a podcast yesterday and I heard this story, which was a Warren Buffett story that I had never heard before.
And this story is really short, so don't worry. So what the case was, it's the 2008 7 8 economic collapse, Great Recession, and Warren Buffett's riding in the car with this guy. I can't remember who it was. It doesn't matter. But he asked him this this question, he's like, Warren, you know, what do you think [00:14:00] about all this stuff that's going on?
Like, are we ever going to get back to a good economy again? Are we ever going to recover from this? Like what's going to happen? And Warren responds to him and says, what was the most popular candy bar in 1934? Anybody know the answer by the way? Guess? Snickers. And then he says, what's the most popular candy bar in the world today?
And that's all he said. What's the point of that? The point is there's some things that don't change, and there's some things that change a lot and change really rapidly. And we can only control so much. And it's really easy to experience a lot of things. If you are in an uncertain situation, you can get analysis paralysis.
You can fear taking any kind of action. You can focus on everything as it changes and never get anything done. So my advice to people would be, like, what are the things that aren't going to change? What's still going to be true? And I think there's some really, like, logical ones. Like, the people that it takes to get things done.
[00:15:00] Like, if you're going to do AI, you're going to need to engage people who are in legal, who are in privacy. People who are really good at data, people who are really good at engineering, people who can do data science, like those things we know and those things aren't going to change. If you're going to do this, you've got to have a really great data foundation, you've got to be good at data quality, you've got to be good at cataloging, you've got to be good at management, you've got to understand the metrics about your data that you have, so spend your energy on those things.
APIs, you want to have AI do things in the real world, you're not going to achieve that magically, you're going to have to take the things that your business does either in your digital world or in the physical world. And put something in front of them that can make them happen in a reliable way. And that's not going to change and you can use that for all kinds of other things.
So, on and on again, I think the thing I would advise is, don't worry too much about, like, getting the right thing right now, because there's probably not a right thing and the landscape is probably going to change dramatically over the next three, six months. But some of those things that you can do can set you up for long term success no matter where the tech goes.
And when things stabilize and there's a right path, you're kind of positioned. [00:16:00] Don't do nothing in the meantime, like find ways to experientially learn get your hands dirty see how things work in the real world, but don't be too focused, I think, on the tech focus on the things that aren't going to change.
Absolutely. And, you know, speaking of Snickers, you know, they always say, if you're hungry, why wait? If you're a data team thinking about AI, why wait? Dive in, dive into it, right? That's right. Yeah. KC, same question. Yeah, I agree with Brian. It's, it's important to it's important to experiment, right? I think experimentation helps overall, like, you know, you can basically don't be afraid to experiment, right?
I think also the other important aspect from my perspective is, is AI, specifically AI projects. You need to have, you need to understand what is the business value AI is creating, right? That is super critical because I have seen working with a lot of customers, a lot of AI projects have failed just because they were doing [00:17:00] AI for the sake of doing because everybody's doing.
It's not. You got to understand what is a business value, how it's what is the, what is the value it's creating for the company? And then you have to also have, like, when you're getting into the actual practitioning side of it, you got to have a clear KPIs, you know what AI is going to drive and how it's going to help business value.
So I think that is super critical from my perspective. Yeah. The other other other areas are like, you know, experimentation and don't be afraid to experiment. Be update. Be updated to latest trends. What's going on? I mean, you can. You can master everything, but I think having enough knowledge around like you know what's going on is also important.
So that way you can connect the dots. There are a lot of tools available in the market. I think it's easy You know, to connect the dots from that perspective, but like the open towards, like, you know, learning what's going on in in in the industry. I think that is also like super critical. But end of the day, it comes down to a I projects is like the business value that they create.
Absolutely. [00:18:00] And that's an incredible point. And, you know, on on what's new in data, we often talk about best practices for data teams. We also like to think about the art of the challenge, art of the possible and challenge people to think big. And at one of our previous events, the one that we did in San Francisco, we had Bruno Aziza from from Google and Sanji Khan.
Ridhima Khan runs Dapper Labs. They do a lot of very innovative stuff. She has this amazing story about how they have an AI generated brand ambassador who has millions of followers on Instagram, millions of dollars in brand deals with the likes of BMW. And this is completely, you know AI generated.
Web3, social media influencer. And you think about the future where, you know, you can have these AI generated personalities start interacting with consumers. You go on Instagram now, I mean, her, uh, her tag is Mikayla with a Q. And it's just mind blowing how many people engage with her. So, that's, that's just a example of a modern experience that's, that, that blew my mind from that [00:19:00] live event.
I want to go to each of you, and it doesn't have to be what you're working on personally, but what's just like a, a, a modern experience that's, that's blown your minds? I think it's, it's, mine are like more tied to, I spend way too much time with my customers, so I have to, I have to translate back to I think A couple I've seen recently. One is especially I'll go back to my customer consumer. One of the category, right? Marketing automation. So I'm working with a retailer and they recently just built out what they call as a marketing AI powered marketing studio.
So basically they build this entire studio where the marketing team can run their entire campaign like, you know anywhere from like, you know customer lifetime value to personalization all using generative AI. Foundationally, but the core essence was like, you know, you get all the data in one place or centralized where you have accessibility.
Then you can run these these campaigns, right? So that was one like which I have not seen [00:20:00] any anywhere. It's super modern. I have not seen any of the marketing related campaigns like that. And then the second one was like working with another retail is embedding product. A I chat bot within the website itself where you can actually go in And then start using natural language.
You can actually start asking questions like, Hey, I have a project for summer coming up and it'll tell you all the products, you know, the descriptions about the products and then that is also linked to the customer experience. If something goes wrong, it'll tell you like, Hey, this product so and so you, this is how you need to use.
Uh, these are the ways that you can actually build stuff and all that. Right. So, so those are like, you know, very revolutionary from, I mean, retail is still behind, but you know, if you think about like the others, but like from a retail perspective, that is all super revolutionary. What they're trying to do is like, they're trying to engage the customer and then driving more customer lifetime value out of those the, the, the AI power products that they're [00:21:00] building.
Excellent. I was thinking about it. I don't have anything that I haven't written about before, so not a new answer, but who saw the Lex Friedman Mark Zuckerberg metaverse podcast, anybody? Yeah, that's what I'm going to go with. Yeah. Maybe explain it. Yeah. I'll explain it to you guys. So let me lead with and telegraph that I am not bullish about the idea of the metaverse currently.
I think there's yeah, well, another conversation for another time. But there Lex Freeman, who's a, has a popular podcast did an episode recently with Mark Zuckerberg if you don't know him, the CEO of meta and they did it in the metaverse, which took some work. So before they actually did this episode, they both had like these full body scans that they did.
They had some specific technology that they were equipped with that let. You know, in addition to the headset, [00:22:00] the VR headset, let their movements be picked up a little bit, a bit better. And so it, it isn't something that's readily available, but both of them, I mean, Zuck, I guess has seen it a lot of times, but Lex was like ogling over the experience and how real it looked and how responsive it was to facial expressions and on and on and on.
And so watching them kind of have this. face to face, but actually a thousand miles apart conversation was or thousands of miles apart conversation was pretty incredible. And makes me like a little bit more interested in. Where does this augmented reality and virtual reality stuff go?
Another one I just saw a commercial for during college football was the little AR lens that goes inside of a football helmet. I forget who's making it, but it allows people who can't, who are deaf, who can't hear to get plays, you know, read into their optic and then it goes away and they're able to [00:23:00] participate in the game.
So these things are small and early, but it's just fascinating to see how. The technology that maybe sounds sci fi. I've read it in a lot of books before. These kinds of things just get to magically exist. Some of it is coming to exist, like, in the real world. And I think that's pretty incredible for thinking about what the future of human experiences look like.
I like in person things, but what's the next generation gonna default to? What's gonna be their way of interacting? How do they wanna be served by people? So I think that's a fascinating one to, to keep an eye on and watch. And it kinda blows my mind to see it. Kind of becoming real enough that I was impressed instead of like it's cartoony characters and it's lame.
This was pretty impressive, so I'd encourage checking that episode out. I have no idea what number it is, but you can find it if you search Lex and Zuck metaverse. Absolutely, and you know, as we go into this, you know, potential metaverse future, where there could be AI generated avatars who have big social media followings, maybe even bigger than [00:24:00] most people, you know, we're gonna have to think about what it really means to be...
Be yourself to be humans that that individuality is going to be more important than ever Brian where can people follow along with your writing? Yeah, so a couple places. So We have a Chick fil a tech blog. You can search for that. It's on medium I write there quite a bit as well as you know, try and promote what others are doing Find me on LinkedIn, of course, and then I also do a blog that is not really Chick fil a specific.
It's just more Things that I'm thinking about as it relates to tech or things around tech, leadership, et cetera, which is called the Chamber of Tech Secrets Brianchambers. substack. com if you're interested in checking that out. Excellent. Thank you, Brian. And, and, and KC, where can people follow along with your work?
I'll give it simple, LinkedIn and Unreal. Yeah. That, that, that makes it, that makes it easier for sure. Yeah. And I'm John Kutay, host of What's New in Data. Thank you all for coming to What's New in Data live here in Atlanta. The live recording of this podcast will be up on all channels soon on [00:25:00] YouTube.
Brian Chambers, Chief Architect at Chick fil A. Thank you so much for joining us. KC Rakam, Head of Customer Engineering at Google Cloud. Thank you for joining us as well. Thanks for having us. It was great.
