Riding the Waves of AI Innovation, with Brian Yamada, Chief Innovation Officer at VML

Transcript

Our transcripts are generated by AI. Please excuse any typos and if you have any specific questions please email info@digitalshelfinstitute.org.

Peter Crosby (00:00):

Welcome to unpacking the Digital Shelf where we explore brand manufacturing in the digital age. Hey everyone. Peter Crosby here from the Digital Shelf Institute. Brian Yamada, chief Innovation Officer at VML says, this era of AI innovation could be best described using the title of the Oscar winner for best picture in 2023. Everything Everywhere, all at Once. So Brian joined Lauren Livak Gilbert and me to take some of the craziness out of it all by digging into the three waves of ai, innovation, content, experience, and agents, how you should think about them and how to work through them within your organization and with your partners. Brian Yuma, welcome to the podcast. We're so happy to have you here.

Brian Yamada (00:56):

Yeah, thank you for having me. It's a pleasure to be here.

Peter Crosby (00:59):

So as the Chief Innovation Officer at VML, you have a front row seat really to what is happening in the world of commerce, how brands are adapting to all of that's coming at them, which is pretty enormous right now. I'd love to just pick your brain about some of the big themes you are seeing right now from your seat.

Brian Yamada (01:17):

Yeah, absolutely. I mean, probably not a surprise to you or anybody that AI is everywhere, but maybe to break that down a little further,

(01:30):

We like to think about three types of disruption that are coming through because of that or opportunities, depends on which side of the coin you're trying to look there, but the first wave is very much the generative like AI to create. So how we can use AI to develop localized content, test a range, different types of images or calls to action or copy treatments on PDPs or media that are driving conversion. But we're not necessarily trying to change the experience layer yet. So if you're on an owned website, you're still a person coming to a website and the images may look more like Boston or Kansas City or what have you, but you wouldn't necessarily know the dimensions of all of what happened under the surface there. The next year is where AI will let us recreate the experience layer. So again, to use the website as the context, instead of just going to a static page, you're coming to experience something that's much more dynamic. You may have a conversational agent that's welcoming you, it's helping you shop. You have, we may be looking at things like the analytics of the page to understand where's somebody pausing or we can track through analytics things actually angry clicks. Where is somebody frustrated and what is an angry click?

(02:59):

Exactly. How do you know? Yes, exactly the way that somebody is moving through, they're able to actually sense that type of interaction. I mean, we've done experiences, actually 10 years ago we did something that was sensing, we did a car configurator based upon people's emotion where we would show them different images using an AI engine that would sense what you reacted positively to and then build the Bentley. In that particular case in the uk, that was actually most suited to your taste.

Peter Crosby (03:29):

So when I recently went to my natural gas provider to understand why my bill went up 30%, they would be able to tell that I am angry as F,

Brian Yamada (03:42):

Okay, good. If they're using the appropriate responsive platform, the same way that we should be able to recognize that in a voice if you're having a phone conversation, those types of assisted tools will hopefully help us make that experience interaction different or better. And it can be the website, it could be the call center, it could be retail environments. Those opportunities exist across the board. The last third layer of AI that we talk about is agentic. So if you think about things like operator that came out from open ai, that changes the dynamic pretty completely. So in the first area, we're changing the content, it's kind of back of house and the experience here, we're changing the interactions that happen in agentic. Like you ask your agent a question to go find something for you instead of going to the website at all. Right? So it's helping you keep an eye out to find whatever the pair of shoes that your favorite football player or tennis player wears, or how to keep an eye on airline prices to different areas. So we like to say that dynamic will completely change the shop and buy

Peter Crosby (04:55):

Experience. So when you talk about them as waves, are they all started right now, but the content one is fast and furious and the experience one is a little slower and the agents are kind rolling in a nice surf thing. How do they pile up?

Brian Yamada (05:13):

It's a great question because they're all happening at the same time at different rates of speed. We like to say if it feels like it's crazy right now, it's because it is. We've never been in an era where there is a faster rate of change. So every model you interact with is only going to get better. That is the worst version that model will ever be from this point forward. So that aspect of exponential change is really happening across each of those tiers. The generative is the most mature of that, but we're still just kind of at the outset of generative video music, 3D CAD gaming. You think about all of the types of things that could get generated. We're still kind of in the early version of that experiential right behind and AgTech. The thing to keep an eye on is more when does that start to scale? Because it definitely will change things, but we just have to keep an eye on consumer behavior to see when that will hit

Lauren Livak Gilbert (06:14):

In the experiential element. That sounds to me very much like personalization or hyper-personalized experiences for people. From a brand perspective, how far do you think that will go? Because you can get super personalized to someone and you might get data about them at a certain moment in time and then continue keeping that data about them for the rest of their lives, and that might change or they might not be comfortable with that data. Can you talk about how far down the road you think that will actually go and consumers will actually want?

Brian Yamada (06:47):

Yeah, I think the right questions to think about where does personalization actually matter to our audiences, our customers? So I 100% agree that we'll have the capability to fine tune a wider range of things than we do now. You think about the concept of personalization scale. I think we've said those words for a long time, and now we actually have the means to make that much more tangible. But I think we do a lot of types of communication and testing where we think that it's personalized, but I'm not sure if our audience is always in on that joke. So it is important, I think, to test our way and some of the testing is performance. Some of the testing may be more satisfaction. Where are we making them happier? Where are we improving their ease? Where do we recognize that, oh, as you came to the website, your gas bill went up tremendously. Let me give you some information on how you might, right, exactly. How do we help be as proactive as possible and avoid getting over to that creepy line? It is going to be, I think more challenging for marketers. There's both more tools to figure out what we can do with all those things, but what's the Spider-Man line with great power becomes great responsibility. We have to think about how to use those in ways that are adding value to our audiences, not just doing more but trying to do better.

Lauren Livak Gilbert (08:14):

And when you think about the number of products that now a consumer's going to interact with, you could have a personalized experience, but if you're using an agent, they're going to tell you one product or they're going to tell you one specific thing or maybe only recommended three products versus if you were scrolling through an Amazon or going to a website, you could see the whole portfolio of products. So a lot of brands are having a moment where they're like, wow, we had competition before and now there's even more competition because there's less recommendation. So what is your advice to brands? How should they be thinking about this change?

Brian Yamada (08:48):

Sure. I think it depends on your category, and I always like to remind, just because we talk about machine to machine and this ENT era, it doesn't mean the humans go away. There's still a person on the other side of that that is ultimately going to make sometimes a rational decision around what will other people think when I'm wearing this clothing or driving that car. So even as we enter the agentic age, brands will still matter, and arguably in some categories it will matter more because there's a vast difference if you ask your agent to find you the best mid-size SUV within this price range, or if you say, tell me the difference between the Mazda line or the Ford lines, and if you're starting out with brand as part of that prompt or context to your agent, then you're going to get a vastly different set of choices back. So it's actually I think more important, not less for there to be differentiation as you're going in.

Lauren Livak Gilbert (09:51):

I love that, especially I just bought a car, so this resonates very well with me, especially when you go in and you're shopping for a car and you're talking to a sales person and you're just trying to figure out all the details. I just went on chat GBT, and I was like, give me every spec that I need, and then I just brought that printed out to the dealership and I could have that conversation and it compared it across the board for me. So I didn't even have to talk to a human. I literally went in and said, this is the car I want.

Brian Yamada (10:20):

Yes, and I think that's both the opportunity and the challenge that marketers will have is to try to figure out where to automate to make things easy, where to provide agentic assistance in ways that are providing more value, but how to balance that with a human touch. It really depends on the person on the situation. Sometimes people want to talk to a live person, some people don't. And for the exact same scenario, the exact same problem or thing. And I think that's where, again, back to your personalization question, I think that's where knowing the difference and helping somebody get to the right type of experience is going to be so important and all of this, we just have, I think more tools, more elements at our disposal to find the right way to make a customer happy and get them to convert.

Peter Crosby (11:12):

That's the opportunity for me, and I've been thinking recently about that is the future of commerce sort of a game of niches that the more you can answer the mission of as many sort of the more you can combine product content for a particular kind of use case or mission with the consumer data and the behavioral predictive data, then you're more likely to win that sale, also win it as quickly as possible. In some ways it is like, I mean, football a game of inches, this may be a game of niches and it's kind of going to be an elbow fight, I think, to have the right stuff to get into that conversation. Is that right? Do you think about it that way?

Brian Yamada (12:16):

I think that question has existed for a while. I mean, think about categories like craft beer, how many different labels are there now, and even if you like beer, there's absolutely no way to keep up with all of them. So how you stand out, how you differentiate across that. But then there's still times that people are just, they don't want the challenge of choice and they just ask for a Coors Light or a Bud Light or a Stella or whatever is comfortable to them. So I don't think that's necessarily a new problem. It will be interesting to see how far up the supply chain does personalization go and can you build more personalized product in addition to more targeted marketing? But there it becomes more, I think of a scale and a profit and loss question. Where will you find where is it worth the highest return? Right.

Peter Crosby (13:12):

Yeah. Fascinating. So in this world, I mean part of what when you mentioned somebody is on a shopping journey where they're actually mentioning the brand is different than the one where they're kind of just doing a generalized mission. So in this world where either could be happening, how do brands stand out from the pack? How do they try and grab the attention to the consumer? How are they trying to win those moments?

Brian Yamada (13:40):

Sure. So I think it starts with brand building differentiation. You need to establish demand upfront. So ideally your brand is there at the forefront, and then when we describe what the agent year will become, so if you're old like me and lived through the first iteration of search and SEO, when search started to become the way that people would shop for brands and for products and categories, we had to understand how the search algorithm worked and decode, how do we position content and metadata in the right way that that would be discoverable and that brought up SEO. What we think we're heading into is this sort of machine to machine where we'll need to create brand agents and those brand agents will interact with consumer agents to determine what's this person looking for and tap into what we would describe as a brand API that will serve the content and brands will need to choose what, what experiences do you want to distribute?

(14:42):

Which things do you want to say? I need to understand more about your person before I'm going to show you a price to see if I should give you an offer or I need to opt in as an example. Or there may be some experiences like showing you a 3D model of the car to see if your kids or your bike or your dog or whatever are going to fit well in that backseat by coming to our site or coming to the showroom. Have you, right, so the old Hollywood phrase was like, have our people talk to your people. I think as agentic does start to rise, it will be like our agents talking to consumer agents that are looking for solutions or looking for product.

Lauren Livak Gilbert (15:20):

That's fascinating. Well, specifically that's blowing my mind. Yeah. Well, and also the fact that I had take a measuring tape out to see if my dog's crate would fit in the back of the trunk, so I would love a 3D model, so I didn't have to do that. So would that be, you would take your brand style guides and all of the claims about your products potentially, and you'd train an LLM to understand your brand and then that would talk to maybe one on the retailer site and you would have a conversation about presenting those products out to a consumer. Did I get that right?

Brian Yamada (15:54):

The agent to agent piece is more like when a consumer is talking to agents saying, Hey, help me understand the dimensions of the back of this car. The agent would go back to the brands to understand, help me see the dimensions and might say, show me a picture, show me, generate an image of, or take a picture of your dog crate and we can understand the dimensions. Then show that to you in the environment or what have you. Right? The agents will actually interact with one another, which is a little bit different than when the generating brand content or brand copy. Those types of elements would be more where we're training on things like brand guidelines and understanding the aspects of how much copy typically converts well on A PDP to generate the recommended version of that page or a banner ad or test different calls to action. Both are possible, it's just different types of tools. I think back to the SEO comparison, if that was linear algebra, the hard part about this moment is it feels more like calculus, like advanced calculus, figuring out which models you need to use and which cases and how to begin to manage and structure all of this in a way that you can govern it all safely.

Peter Crosby (17:11):

I never took calculus for a very particular reset, so oh my God, that's why I struggled to understand all

Brian Yamada (17:19):

Of this. You could have a math agent helping you with that today. I could. I could.

Lauren Livak Gilbert (17:23):

The agent to agent model. To me, what comes to mind is how a brand actually makes that happen in an organization because it involves supply chain, getting the right specs and then the marketing team and the sales team and maybe the category team. How are you consulting with brands or working with brands on how to actually be able to support things like an agent to agent model internally? Yes,

Brian Yamada (17:48):

That's a great question. Yeah, it definitely requires examining things like your processes, how does that work now, who is interacting with what data sets at what point? And then there's absolutely an aspect of needing to rethink your technical and your data architecture in a way that you have to understand what brand API really gets down to is what data sets do you want to distribute? What aspects do you want to make available? Are you looking to publish? So if the web was branded as publisher, but you controlled that space, you now have a much more dynamic environment where let's say you're in travel, you might start with a conversation with your agent to try to identify which beaches you might want to go to or what hiking trails you'd like to take your dog to. But at some point you might want to see the map, you might want to see images of those places. You might want an itinerary or timeline, and your agent actually is going to help you shape those in different ways. So you've got to that. I think as brands, you've got to ready to understand the availability of your data and then start to make choices around which aspects do you want to publish or not

Peter Crosby (19:14):

At your clients, who's sort of the loudest voice in the room around who are the people that are really sort of grabbing this and becoming, maybe it's not consistent, but becoming the change makers, the sort of forcing of conversations and things like that. Who's sort of most invested in figuring out how to make this work across your

Brian Yamada (19:38):

Clients? Again, maybe to go back to those areas of change, the CMOs are probably the most involved in generative as you think about the owning of content and publishing media. Then the experience layers typically are owned by the chief experience officers to try to understand what are our owned environments looking like and where do we start to affect that change. I think the hardest is that last tier because it kind of takes them both of those plus the CIO or CTO to understand your existing tech and data infrastructure to establish, and that again, the hard part about all of this happening at the same time. It is the movie from a couple years ago, everything everywhere all at once. We are very much in that moment, and it does force a lot of new questions. Does force you to reexamine or rethink is your data, is your tech stack, are your processes AI ready?

Peter Crosby (20:42):

Yeah, and I mean so much of when I think of how SEO works, whether you call it GEO or generative engine optimization or agent to the early advantage is real when it comes to these optimization strategies. And I wonder how you counsel your clients in terms of urgently to take that third wave on in order to get sort of the spoils that will come from those who are there first. How do you think about that?

Brian Yamada (21:16):

Yeah, I mean, you're right to say we definitely, there's always a first mover advantage

(21:24):

And I think there will be a race to own. You're the go-to agent and I think there will be many agents. I don't think you're going to live in a world where you have one. Microsoft arguably is trying to race Google to be your co-pilot agent, your work agent that's helping you manage email your calendar, like work you. You'll have different agents that are helping with social, you different agents that are helping with content. You keep an eye on what shows you should be paying attention to maybe because of what your friends are watching, what you want to be able to keep up with in the conversation. So there will be a race to try to see which agents do start to get scale across each of those different areas. And I think for us, wave one, keep in mind we still have opportunities to evolve your generative and your content strategies like how and where generative applies to solving for making things more relevant to the audiences you have today. And you have an opportunity to rethink your experience layer before you even get to Egen X. So we're definitely starting a lot of those conversations now. We're looking to see where there's the biggest opportunities in value and especially on the do you need a brand agent and a brand API, yet we're not seeing the scale of AgTech happening on the consumer end, so we need to have that scale there before you're working that the same way that you're going to work search engine optimization today.

Peter Crosby (23:02):

Yeah, Adobe put out a stat talking about how in the shopping season the use of agents or chatbots in holiday shopping rose 6000% over the prior year. Yeah, I'm sure. From zero, exactly. On a pretty small base.

Brian Yamada (23:23):

Yes. And that's the thing is there will be tremendous growth around that, but you're right to say it is still starting from a relative small scale and behavior won't change overnight. Habits get longformed and things like shopping are still fun right there. People like to brow, so it doesn't mean that the old things go away. I think that's part of the problem of everything everywhere, all at once time we're living in is none of the other stuff has fully disappeared. You still have to figure out how to make choices across all of those areas too.

Peter Crosby (23:56):

Yeah. So Brian, just to close out, I was lucky enough last week to go to the launch event of Future 100 at your offices in New York and it was awesome and overwhelming and a hundred options for things you could do to win this little game of shopping, which was amazing. Such great work and great people. I was wondering, as you think about what's in that incredible volume or the things that are on your mind or your customer's mind, what are you most excited about that you just giddily want to share with our audience that they ought to maybe start thinking about or you can do the top three, whatever it is that comes to your mind. Sure. Yeah.

 

Brian Yamada (24:48):

I mean, on one hand we live in this era where it feels like everything's moving a million miles an hour all at one time. But on the flip side of that, everything is we now have so many possibilities that we can unlock for our customers or our potential customers, and that can be with things like tech for accessibility, letting everybody shop regardless of their ability to see or hear. It may be making your retail experience different for different range of people. I had love the categories of tech around accessibility tech for good tech that actually is solving people's problems and the tech that ironic, the tech that makes us feel more human. Those to me are the things that are most interesting,

Peter Crosby (25:42):

Particularly

Lauren Livak Gilbert (25:43):

The human factor. Oh, I was just going to say the human factor doesn't go away. And I think to your point before, it makes it even more impactful when you bring the human and the heart and the emotion into it because we are all humans at the end of the day, and we're social creatures and we have emotions and we can't lose track of that in any of the work that we're doing.

Peter Crosby (26:03):

Yeah. Speak for yourself Lauren. Do you want to a robot?

Brian Yamada (26:09):

No, there were plenty of robots at CS is here.

Peter Crosby (26:12):

Yeah, I bet. Then

Lauren Livak Gilbert (26:13):

NRFI shook a hand of a robot and he said hello. I was like, oh, hello, hello

Brian Yamada (26:20):

From me. There were robot dogs, lots of robot companions. We saw a robot planter this year where it would actually help bring your plant to life. Yes, lots and lots of early exploration in that category too. So yeah, the robots are coming

Peter Crosby (26:42):

Between now and then when robots take over the world. In the meantime, we try to merge the human with the digital and hope it creates a better experience for all. And Brian, we're so grateful for you and VML to come on our podcast and share your views. We're all just trying to figure it out.

Brian Yamada (27:04):

Yeah, there's a lot going on, so thank you for having me.

Peter Crosby (27:07):

I appreciate it. Thanks again to Brian for calming The Crazy Clock is ticking down to the Digital Shelf Summit where VML will be joining us for a Raucously Smart panel of folks from Kroger, the world of retail media and brands to talk about the future of omnichannel experiences. And that's happening just in the first 90 minutes. Join us in April by registering at digitalshelfsummit.com. Thanks for being part of our community.