Absolutely. I want to predict actually what I want next. It could get into the mind and predict where we're next. Speaking about who's next, we've got our amazing guest, Indy. Indy Cho, welcome.
Good morning. Hi, I'm Indy. I'm the AVP of data products and data science at a small-known retailer that people may have heard about. It's called Costco. I was going to wear my Kirkland T-shirt today and I couldn't find it. I think it got swiped by one of my kids. That's how hot it is.
You've got good taste. Okay. So, hey, Indy, with over 20 years of experience focusing on data and AI, you've done it in so many different industries. Now in commerce, talk about how new opportunities arise. How do you think companies should view the power of data, gen AI? How is it creating new opportunities, driving operational efficiency, getting to innovation?
Yeah, I think it's a great question. The first 20-plus years of my career were all in aerospace and defense. There's a lot of manufacturing, a lot of maintenance and support. There's some pretty clear use cases there. It's all about driving as much productivity as possible. I'll just give you a really quick example. aerospace and defense. I'd before switch over to the retail opportunities, but over aerospace and defense. I mean, one of the biggest opportunities that we were looking at was what was called fault isolation modules, right? And Ray, you fly around the world a lot, you know this, right? Everybody hates delays. But did you know that when that little fault light comes on, there's gotta be some technician who's opening up a book. And by the way, these books are about 3000 pages of volume and there's about 18 to 20 volumes stacked in the back and they gotta follow this breadcrumb trail of what to do in order to go fix that problem. We said, hey, you know what? This is something that we can absolutely throw machine learning and AI out at the beginning and say, how do we optimize this, right? How do we get to the best answer as quickly as possible? And so when I think about that and that translating over into where we are now, I think those are still problems that we live with on a day-to-day basis. There's always gonna be process efficiencies, whether it's in supply chain and retail, whether it's merchandising, whether it's in inventory allocation, sales forecasting, all those kinds of mechanical processes. But I think the opportunity is gonna be, I think it's gonna really democratize a lot of decision-making, right? And it's gonna allow people to kind of work at their own pace. You know, the days of kind of centralized decision-making are over. I think one of the things that makes Costco really successful is that it is very federated. I mean, you got a lot of things out there in the world that corporate buyers will buy, but our different regions have a lot of autonomy as well. And I think it's gonna, and that's something we wanna continue to promote. And I think having good information available is gonna be really critical for making better decisions to connect with the members locally.
Wow, that's fantastic. But from a retailer's perspective, what specific challenge do you foresee when you're implementing AI? And how will this challenge be addressed now and in the future?
You know, I think from a retailer's perspective, I think the question doesn't start with the technology. The question has to start with the ethos, right? And for example, you know, we talk about prices quite a lot in retail, right? Consumers are very price sensitive, and every retailer does what's called price comping, right? Which is they go out and determine, you know, hey, there are two like products, and what, you know, what are my competitors selling it at? And they try to get the best price. And a lot of companies, of course, are driven by profit, right? So there's dynamic algorithms out there that perform dynamic pricing models. And so you go onto a website and you just don't know if you're getting the best price at the moment. All you know is that you're getting the best price against your competitor, right? And then they're kind of manipulating all those prices. And we have a very different philosophy around how we do pricing because our ethos is very clear, right? We have this idea that we're gonna provide the best member. We're not gonna do all sorts of weird price manipulation and everything. And so I think it kind of starts with the ethos of where you want to be. I mean, could we, you know, mark something up an extra penny or an extra 10 cents just to get some additional profit? Absolutely, we could do that, right? But I think what it leads to then is the, is kind of the, we lose trust with our members when we do that, right? So I think the ethos becomes really the guidepost for what we're gonna do with AI, right? If you don't have that to begin with, I think AI is just gonna be something that's gonna be really disruptive in a negative way for retailers because it's gonna disrupt trust with the members and consumers.
That's a really good point on trust, Indy. I noticed recently that my hotel site, my airline site, if I don't flush the cookies, the price keeps going up every time I visit, right? And that's a very, very different ethos than what you're talking about. And it's amazing. Like when you start with a trust and a member-focused model, you end up in a very different kind of scenario. This is also the same place as when you think about when you apply AI and when you actually apply the human touch, where do you see that working together, where you're bringing these experiences together, having the AI experience, the human experience, and bringing that physical digital world together?
Yeah, you know, I think the, it's an interesting question because I think the future isn't really gonna be about AI, at least not for a long while, or we just kind of completely let the reins go. I think it's really gonna be about curation, right? I think curation is gonna be a new big buzzword that we're gonna hear a lot about because it just doesn't get it right all the time. And when you think about what we're doing today, and we see some amazing work today with Adobe Firefly. I think they're the kind of first movers out there in the market, really kind of talking about how you can use Gen AI to create any sort of graphical image that you want, right, to generate marketing materials. You could create amazing personalized marketing materials, but at the same time, I think we as humans kind of can differentiate what we see in front of us as being synthetic versus being somewhat curated, right? It's the same reason why we, you can ask Gen AI to create a mashup of a Taylor Swift and Kanye song or something, whatever it is that you want. It's gonna be seamless thing, but you also know that that's probably never gonna happen, right? Like that mashup would never, ever happen. And so I think the future is gonna be a lot about curation because the content creation will be limitless. I think that's what Gen AI is gonna give us, right? Access to information and access to content creation is gonna be limitless. It's gonna be about our ability to curate. So when we think about the images that show up in our e-commerce sites, whether it's advertising that we want to put into inline messaging as we kind of interact with members, you have perhaps like in transit when the product is in shipping and you're getting those notifications about when they arrive, or even customer service calls. If they have to call because maybe they're missing product, a missing item, that never happens, of course, with us, right? Never, never at all. Never, never at all. But those experiences, when they come online and they want to interact with a chatbot agent, do they know that they're actually interacting with a chatbot, right? An advanced AI or if they're interacting with a human, I think these are gonna be blended experiences for a really long period of time while we get things right.
Yeah, I think I like that. I think curation, I do believe that it's gonna be a really big important. And I think that's what we'll see more. I totally agree with you on that, Indy. But I recently read at Costco, an article that Costco uses data to connect with customers around food safety. Now, as a retailer, you have lots and loads of data, customer data, but do you feel today that retailers are fully leveraging data to enrich customer lives and provide better experience, or just some of them just scratching the surface?
You know, I'll give you my opinion on other retailers. And I think it's a little bit of the Wild West, right? I think it's creating a lot of distraction, a lot of noise. I think they're over-utilizing the data. They're not doing enough curation of it. And I think it kind of goes back to that conversation around ethos. It goes back to the trust. What's the goal? What's the objective? You know, just continuously hammering my inbox with ads that are completely not relevant to me, not contextual, it's not valuable. And over time, it's just eroding trust. And guess what? The click-through, the open rates, the click-through rates, conversion rates for me are gonna go down significantly. So I think that there's just a lot of almost, I would call it overzealousness for how they use consumer data. I think a lot of it is gonna get back to figuring out how they want to, what kind of relationship do they want to have with me as a consumer, right? I think a lot, we're thinking a lot about that as well. There's this kind of rule that we talk about, which is, you know, the don't be creepy rule, right? I think everybody's probably had that experience where you might be watching, you know, one of the ad-driven video services, like a Hulu or whatever, and all of a sudden you see some ad about a conversation, you've never Google searched, but you see an ad for something that you might talk about, and you go, gosh darn it, what service picked that up? Was it my Alexa? Was it my Google search chat? Something fed that algorithm, and that gets creepy. And we want to make sure that we're not being creepy about stuff.
Yeah, you don't want Big Brother watching you.
Exactly.
Really, really good point there. Yeah, so there's that fine line between personalization and creepy. But as we get to the digital area, the other piece that's really about, when do we actually have that human touch? When do we actually improve that connection, right? As I said before, as you heard, right, we can fully automate something. We can augment the experience of a machine with a human to make sure we get the exceptions and patterns and understand why we break things or get things to level of precision. We can augment the humans with a machine to improve that decision-making, improve that experience, or we can actually apply the human touch. For you, when you make those decisions