AI Transforms Creativity, Responsibility, and Ecosystem Readiness at HP

Jeff Mosier:

Hi, I'm Jeff Mosier with the Emphasis Knowledge Institute. I'm here with Morgan Chemij, the Senior Director of Global Marketplaces for HP. We're here to talk about AI and marketing. Morgan, thanks for joining us today.

Morgan Chemij:

Thank you for having me. I appreciate the opportunity.

Jeff Mosier:

Absolutely. Well, tell us a little bit about how AI is changing how HP operates marketing and runs your marketing organization.

Morgan Chemij:

I think it's changing at every single facet of consumer touchpoints. It's changing the creative that we generate and how we generate that creative. It's changing the targeting and how we go about targeting consumers. It's changing to some degree how we adjust our media mix and where we place our dollars. So it really is impacting virtually every facet of what we do, and it's doing it quickly, which I think is the other facet. It's moving at a much faster pace than I think anybody could have expected or potentially any other type of marketing that could have predated it.

Jeff Mosier:

So this rapid evolution, what challenges does that present to you with it coming so fast?

Morgan Chemij:

I think there's quite a few. One of the things that I have under my purview is around the content generation piece. One of the biggest challenges that we have is AI is incredibly powerful. Our ability to go ahead and create multiple different locations off of a single photo shoot with a single modeling green screen. We can change the background. It can be a coffee shop in Bogota. It can be a street scene in Taipei like this, no problems. And so we're moving very quickly in terms of how we adopt that. But the next layer to that comes down to what if we don't need a model anymore? What if we can go ahead and just generate the human purely using images that exist in the open wild or photos that we've taken for our own purposes, and we can take five or six different personas and effectively amalgamate them and create one new persona. I'm like, "Is that okay? Could we do that? Should we do that? How do we let consumers that are looking at ads know that we're doing that?"

And so, I think that in and of itself is one of the larger challenges that we're facing with how quickly there's this push to move towards AI. And I think a lot of it's exacerbated by what's happening right now in the macroeconomy. And I think be it a consumer electronics manufacturer or a textile manufacturer or your local grocery store, I think everybody is feeling the pressure in terms of how do we maintain our pace and do it for less? And I think a lot of entities are looking at AI as that great solution that can enable that.

I think the other piece is people see AI as a panacea. While I think there's some truth that AI can do some things better, faster, more effectively, I don't think it can do all things. And frankly, a lot of the AI, it's only as good as what you put into it. So I think there's this garbage in, garbage out component. And if you're not feeding in good audiences or good baseline creative or good test results or good copy, the byproduct that you get from AI is not necessarily going to be that much better than what you had if you were doing it manually or via some other mechanism. And in some instances, it could be worse.

Jeff Mosier:

Yeah, that kind of brings up the question of data as you mentioned. AI is only going to be good as the data that you're feeding into it. What are the biggest data challenges you're facing right now?

Morgan Chemij:

I think there's some challenges in terms of what's the right way? There's legally permissible ways to use data, but what's the right way by the customer to go ahead and use the data? I think there are still... HPB and HP, when you think about it, we've got your print usage. To some degree, we have your PC usage. There's a variety of different data points that exist that we can use, whether it's to improve your customer service experience. If we know a problem's going to happen before you know a problem's going to happen, and we can help proactively solve that, or whether it's obviously for targeting capabilities, capitalizing on customer lifetime value by understanding the ecosystem of products that are in your household, et cetera.

So I think there's a lot of internal challenges. I don't think the problem is as significant with the data itself. I think there's good, clean data out there. There are good ways to leverage data that are perceived as permissible by most entities, authorities that keep an eye on this kind of stuff. And I think consumers are comfortable with a certain level of data usage and targeting. But I think for the future, we can do a lot more than maybe we could have done historically. And I think there's just a fine line between what's the right way to do it and what might be a little bit of a gray area. And I think as a brand, we want to make sure that we stay out of that gray area.

And so, I think we're evolving quickly, and I think we're having the right internal discussions. But yeah, the data and how we use it and how AI factors into it is a much broader conversation. I will say on the whole, we are more effective in terms of what we do now than potentially we would've been a year ago with the advent of AI and our ability to go ahead and analyze some of this data in a different degree and leveraging clean rooms to go ahead and understand all the different data points in a way that's ethically reasonable.

Jeff Mosier:

So how do you all approach responsible AI? What is the kind of philosophy behind that?

Morgan Chemij:

We obviously have an internal team that takes a look at all of these different elements. And so, before we even route somebody in through procurement, there's a predefined set of questions that we'll go ahead and ask potential vendors to populate for us. And based upon those responses, identify whether or not it makes sense to continue the dialogue. And then, when it comes all the way far enough along to open a PO, if that's the route we choose to go, there's a secondary set and procurement gets a little bit more involved. So we have a lot of checks and balances internally to protect our consumers and our brand.

Jeff Mosier:

What's been some of the biggest surprises you've had from AI as far as their capabilities, what they can do?

Morgan Chemij:

I don't know if it's been surprises in that capacity. I think there's still a lot of surprises in terms of what we don't know, frankly. And I think the other piece is, surprises with the actual adoption rate. So if you take a step back, think about when Spotify came onto the scene, I don't know, more than a decade ago. Spotify came onto the scene and it took Spotify 1600 days to get to a hundred million users. And fast-forward quite a bit, and Instagram obviously widely utilized, comes to the scene and it takes Instagram, I want to say about 900 days, about three years. So we've gone from five years with Spotify that made music at your fingertips and simplified your life as it pertained to music and music sharing. And then we've got Instagram that comes and in a very short period of time, three years, for sharing images and then what's happening in your life.

And then, if you skip forward to ChatGPT, ChatGPT took 60 days to get to 100 million users. That adoption is unheard of. I don't know. If you think way back when, how long did it take the printing press to see widespread adoption? So I think actually the adoption and this becoming integrated, not just from going and downloading the ChatGPT app, but just even on your Android or on Siri, how pervasive this technology has become in our everyday lives. You and I wouldn't have had this conversation two years ago. And even a year ago we could have it, but it wouldn't be as informed as it is today.

And I think there's a lot that's still unknown. And I think that's the other thing that's interesting to me and challenging. If you think about the worlds of perplexity or ChatGPT or Gemini, there's these platforms that have seen pretty widespread adoption for generative engine optimization, Geo as it's called. People use them day in and day out, and the amount of money that's been invested in these platforms is significant. Yet they, for the most part, are not turning a profit in any capacity. So I think we're very early in a much broader journey. And I think obviously a lot of these platforms in terms of even monetization or what does the monetization roadmap look like? How are they going to recoup all the money that they're spending on bandwidth? I mean, every time you query Gemini, it uses a lot more bandwidth than querying and Google. Let's be realistic. So there's a cost of that. And those companies that are serving us are going to have to figure out how to absorb that cost.

Jeff Mosier:

What have you seen with AI that has not lived up to the hype so far? What is disappointed on your end?

Morgan Chemij:

There hasn't necessarily been the transfer of information or the synergies between some of these platforms that you would hope. Amazon has Rufus, the Amazon desktop experience or mobile experience of best in class if you're trying to find a product. Gemini experience that Google offers, I feel like, is best in class if you're trying to learn more about something or do recon, and yet somehow they haven't transferred it over to some of their other technology yet. So I think some of the personalization elements, be it in marketing or otherwise, is still not quite where it was promised to be. So I think you've got a variety of different things that could be worked on.

But this comes down to prioritization for these companies. What's the priority? Is it building tools to monetize these new platforms? Is it improving the home experience on some of the hardware that these companies have manufactured that are installed in a pretty significant amount of homes? I think they're figuring it out just like marketers and consumers are figuring it out.

Jeff Mosier:

For you as a marketing leader, what is the ultimate thing you would love AI to be able to do for you at work that it can't do yet?

Morgan Chemij:

I think one of the other big opportunities, frankly, is just around project management. At a company our size, we deal with a lot all at once. And I realize how AI can simplify or eliminate the bandwidth required to execute a project, but there's still a lot of minutiae before that project gets off the ground. And so, being able to really go ahead and not just explain to ChatGPT or Teams or Copilot what I need to do, but to actually take all the different software pieces that exist on my mobile device and on my desktop and actually have them all talk to each other to successfully execute something.

The big challenge I feel like is the devices aren't necessarily, or the software applications aren't necessarily talking to each other. And that's one of the biggest challenges. I'll ask something, I'll ask Gemini to do something and it'll say, "Oh, I don't have access to that." And unfortunately at this point, there's not something I can go into settings and grant access, it just doesn't have access and that solution is not going to happen in the immediate future.

And so I feel like we still have this Nintendo versus Sega or Apple versus Android more recently. The ecosystems are still not talking to each other for the betterment of the consumer experience. And I think at some point when all the different ecosystems are able to talk to each other effectively, we'll be in a different place, and we'll potentially be in a better place than we are right now. And we need for all those pieces to talk to each other and to synergize, to improve my experience and my productivity.

Jeff Mosier:

Do you think agentic AI will help with some of that? Obviously some things where platforms refuse to talk to each other, maybe not, but what do you see?

Morgan Chemij:

I think agentic AI in general has the capability of making life a lot easier for myself and a lot of my peers. And I feel like agentic AI over time is going to be the scale that it enables is pretty significant in terms of there's scraping of prices and product details and things like that. And we pay a myriad of vendors to do that. Agentic AI can do all that for us.

I think the other piece is not just internally within HP, but even when dealing with partners, how are we setting up our agentic AI future where let's say Best Buy has a question about a product. The Best Buy team, their retail team says, "Hey, what's the battery life of this laptop because we're updating our product detail page and we need those details and we can't find it."

For them to be able to go and talk to the HP agent and just ask that question. Or, "Hey, please give us the most up-to-date product detail page content." And then for our systems to be able to talk to the 1WorldSync or whatever platform Best Buy is using, and they can actually do all that themselves without my participation or anybody else's participation. So again, how do we get all the platforms to talk to each other and kind of take us out of it with the exception of QA, which I think is still going to be an important component? So I think there's a lot of promise there.

Jeff Mosier:

What is one piece of advice you would give to others who are trying to figure out how to use, how to incorporate AI into their marketing?

Morgan Chemij:

I think this is definitely a trial and error thing, to be honest with you. I think, don't be afraid to test and fall on your face, because I think nobody has mastered this yet. And frankly, I think the other piece is, don't be afraid to ask for proofs of concept from a lot of these companies. Whether you're a big brand or a small brand, a lot of these companies are looking to build their book of business, and sometimes just getting you on board for 30, 60, 90 days, they'll be willing to go ahead and bend over backwards. Because they realize this isn't a proven space. This isn't direct mail where I'm switching from vendor A to vendor B and everything's a known. This is literally the great unknown and brands need to, I think, flex a little bit when we're talking to new vendors or existing vendors in this space if they've got unproven technology, which I think a lot of them do right now.

Jeff Mosier:

Fantastic. These have been great insights. Thank you so much, Morgan, for taking time out of your day and giving us a little bit of your knowledge here.

Morgan Chemij:

Absolutely. Thanks for your time.

Jeff Mosier:

Thank you. Appreciate it.

Morgan Chemij:

Appreciate it.