Loading summary
A
This is adtech God and I command you to listen to this house ad. So if you're listening to this show, just know that you've really stumbled upon a giant network of content across advertising, marketing, media, publishing, and of course the people that work in this great advertising industry. So go to market, check out all of our brands. We have multitude of shows from the Brand Forum, the Advertising Forum, the Monopoly Report, the Ad Tech God Pod, the Market and more. We are bringing more podcasts to our network. We are consistently and constantly bringing on new shows. So check it out. Market or search for any of those brands in the app that you use to listen to this podcast. Enjoy the show and see you all soon. Welcome to the AdTech Godpod, your window into the world of advertising technology and the people behind it. I'm your host, Ad Tech God. Hey, this is AdTech God. Welcome to another episode of the AdTech God Pod. Today we're rolling out a session from Architecture Live called Can Neural Networks Save Healthcare Advertising? It's with Brad Fox, the SVP of Health Media at Dentsu, Josh Walsh, the co founder and CEO of Branch Lab, and of course Zach Rogers, founder of Sensical. This talks about how pharma and healthcare ads have always had to walk a very fine line when it comes to data driven personalization. The three of them really dig in and share a lot of insights that I was unaware of personally having never worked in pharma. I hope you enjoy this episode and if you're interested in any of our brand conversations from Architecture Live, you can listen to the Brand Forum. And if you're interested in anything related to ad tech or technical deep dives from Architecture Live, you can tune in to the Market podcast. Hope you enjoy the show.
B
So AI and healthcare. I think this is maybe the original consumer use case for AI is we all self diagnosed because we can only get 10, 15 minutes of our doctor's time. I feel like everyone I know has some story about learning something through using AI that they hadn't realized whether that's a specific condition, disease or health related question. So maybe we could start with a general question about patient adoption of AI agents for health information. What are the. This is moving fast. What are the implications for advertisers? Josh, do you want to start?
C
Yeah. So I mean you kind of touched on it right in the opening here. The stat that everyone's kind of focused on recently came out of OpenAI, which is 40 million people per day. Go to ChatGPT and ask a health related question. That's Chatgpt that's not including Gemini, which I'm sure that number is much higher. And so if we just think about that from a consumer's perspective, you know, instead of going to search, instead of going self diagnosing on WebMD, as Zach mentioned, or even like retargeting opportunities from a brand website, the experience is changing, you know, rapidly across the whole industry. In healthcare, that's probably even more acute.
B
Brad, do you want to pick that up? When you look at this landscape, do you see all of these different Internet destinations that used to support health journeys, search WebMD, all of the health related websites collapsing into the agent chatbox?
D
I think there's still a place for both. I think the interesting thing that we've seen is that patients have started using ChatGPT like Josh said, and as they've announced their own health specific platform, they're using it outside of clinical hours. So the times where they used to maybe look to WebMD as their kind of source of research, I think that's where it's really changing. Like I think of Josh, the 40 million people per day, like 70% of that is happening outside of clinical hours, which is kind of fascinating. I think that's the real behavior change. And you have people that will always be the loyalists to the WebMDs of the world and things like that. But I think as consumers become younger and more comfortable with AI as a whole, it'll continue to shift in that direction. But we're also seeing it from doctors too. I think Josh and I were talking backstage about open evidence. And open evidence at this point is seeing over a million searches a day from physicians and open evidence is chatgpt, but specifically for doctors from clinical sources. So I think that's also telling you it's a trusted source for both patients and also the physicians too.
B
Do you think that the information people are getting from AI as well as the information doctors are getting is leading to substantially better outcomes already or more accurate diagnoses or improved journeys? Or is it just a replacement for what the Internet has been providing up until now?
D
I think if you have asked a doctor, they'd probably say it's driving them crazy because you have people coming in that think that they're doctors or that they know everything because all of the information is there. I think it's creating more informed consumers and there's certainly a lower barrier to be able to do the research and ask the questions that you might not feel comfortable asking a doctor before your appointment. But I think it's also creating more informed and focused physicians. I think the ability for them to not have to comb through a medical journal or a textbook on a regular basis to be able to find a very specific answer is certainly making their lives easier. And that's what we're hearing too.
C
Yeah, I would say for consumers, like probably too early to tell I think. Does it replace, you know, what used to happen on search and self diagnosing on publishers? You know that that is largely already happening. Although there, there will be some loyalists as well. I would say like if we, if we look at, you know, if we were to turn to academia, we would, we would know that when people self advocate and when they seek care sooner, health outcomes do improve. It's probably too early to try that correlation now, but generally I'm hopeful that that will be the, the outcome here. Also for advertisers, as we think about, you know, the opportunities in search or publishers or retargeting kind of shifting, you know, perhaps becoming less and less, audience targeting becomes sort of a key focus here. And the way that we've been performing audience targeting for the past five, ten years in healthcare, it's just, it's critical that we re architect that using these newer types of technologies.
B
Yeah, right. So if we lose overtime slowly you have an erosion of search advertising of as we've known it in the health category. You have an erosion of contextual which has been huge in health and you have a erosion of retargeting to the extent that's happening in health. And it's, you know. So my thoughts go to a couple of different places. One is what's going to replace that for pharma marketers. And the other one is what are the frontier model companies, the ChatGPTs Geminis going to build for this specific purpose. So we're not there yet. Obviously we don't have a hyper targeted healthcare opportunity inside of AI environments. Presumably over time we will have one and it'll be pretty powerful. But for now, where do your dollars go as the addressable audience shrinks?
D
Yeah, I think for, for us on the agency side and working with clients, it's been a delicate balance between the super targeted side of digital and trying to find ways to introduce that to things like streaming TV and otherwise and balancing that out with the traditional forms of search and mass reach and things like that. But as, as you said Zach, the ChatGPT health and, and even sensitive searches on ChatGPT and other platforms are going to be excluded from advertising at least to start. But I think we wouldn't be Surprised. And I wouldn't be surprised if Google, as they start monetizing AI overviews or something like that, that's kind of the first dip, first way of dipping their toe into monetizing AI content in health specifically.
B
Okay, great. So I'd like to shift gears a little bit and talk about some work that you guys are pioneering around neural networks and how they can help challenge, overcome this challenge of reaching patient populations in a way that's sort of privacy protecting. Could you speak to that to start, Josh?
C
Of course, yeah. So, you know, I mentioned this earlier, like as you know, and this is too, I don't know, this is too extreme, but if there were to be like, you know, or will be a collapse in search contextual retargeting options, we become limited only to audience targeting. Audience targeting in healthcare is typically a derivative of medical claims data. So that data source is protected in this, in this evolution, but also carries a lot of regulatory constraints. Our approach here, the branch lab approach, is to model that data, that clinical data with best in class demographic data. And in effect, what we're attempting to approximate are digital signals that in combination can predict a health care population. Think patient population. So not deterministic. We're sort of one step removed. And so that's been our approach. Neural networks are fantastic at this. We all know, you know, things like ChatGPT are built out for language. So what is the next sentence going to be? Or sorry, what is the next word going to be in the sentence? Similar concept with healthcare, like what is the next event in this journey? And then looking at, for example, you know, dentsu data or other things and modeling to understand like how those can be predictive of a population. So our thought here is targeting options are becoming more limited. The way that we've been performing, audience targeting needs to be rearchitected. This is our vision for how we re architect that once we do have search like or search quality advertising opportunities in a chatgpt or in a Gemini, that largely replaces what search used to do, which is people saying, hey, I just received this diagnosis or hey, there's this weird thing on my skin. What is it? But audience targeting is after that. So maybe I'm diagnosed with something, I might live with that condition for the next, you know, 10 years, 20 years, 40 years of my life. At which point, you know, throughout that lifespan, new therapies come out, new treatments become available. How do you talk to those people? Audience targeting would be the mean. So that, that's kind of the way we view the world. That's the way we see this evolving. And yeah, deep neural network architecture we find works really well for this.
B
Could you talk a little bit more about the data that you're feeding it? Like, this is real medical case. These are real medical cases, right?
C
Yeah. So the data is, you know, think of it as historically a data scientist, a human or set of humans, typically at a consulting firm, would mine this data, read this data and build reports, and marketers could read those reports and maybe talk to those people. The concept here is we can train a series of models, put an agent on top of those models, and. And now you can talk to the agent about what the model learned. You can never see the data. It's never exposed to you. It's a combination of both medical as well as demographic and other signals. So it effectively gives you this hyper intelligent data scientist in the room that can help you build out, you know, whether it be patient population models or HCP or caregiver, etc.
B
Okay. And it's a combination of, like, you're getting insights that, that map conditions to demographics. Is that fair to say, or am I oversimplifying that?
C
No, that's a, that's a fine way to think about it from a technical perspective. We're not mapping them. That's an important distinction per regulations. Rather, you can, you can uncover demographic characteristics that in combination can predict patient populations. More specifically, they can predict populations who are likely interested in a specific treatment.
B
Okay, thank you. So, Brad, speak to the patient journey. Like, what's the opportunity from a care. A diagnosis and care standpoint?
D
Yeah, I think the important distinction and what we've started learning and I think our mutual shared vision and the way our partnership started was the belief in the importance of patient journeys and the way that healthcare audiences have been built in the past, is we on the agency side or our clients reaches out to a partner and says, hey, I want to reach an audience that is people diagnosed. And the example we like to use is psoriasis. And that was before SNL spoofed Tesla over the weekend. But it was really an oversimplified way of thinking about it. It's just, hey, we're going to lump a bunch of people together that match the profile of a diagnosis without any regard or totally disregarding the journeys and the different ways that people get to a point of diagnosis. And the reason I think psoriasis is an easy example is for every person that lives in New York City and sees a specialist on their first sign of symptoms or is referred to a Specialist immediately and has strong commercial insurance and ultimately ends up on an advanced therapy. There's someone that doesn't meet any of those criteria and maybe is older and has pushed off their symptoms or doesn't trust the health care system, doesn't have the right quality insurance, and they get to a point of diagnosis when they end up in an emergency room because they've scratched their psoriasis or their rash to the point of bleeding. And it's okay. Now I have to handle this. And the way that patient audiences have been built in the past treats those two people the same, and as a result, brands speak to them the same. And I think that's why you see so many, aside from a lot of other reasons, so many pharma ads on TV. It's okay. Our audience is 20 plus million people. How do we reach such a large population with one message? But the partnership that we have with branch lab and thinking about patient journeys and building audiences to understand kind of the different demographics in the segments allows us to speak to each of those populations differently.
B
Yeah, really interesting. So where does something like this work that you're doing together fit into, from your Dentsu point of view and client point of view into your overall media mix? Is this still in the category of like an experiment or how does it scale?
D
No, it's worked fully into our addressable strategy from essentially anywhere that we can use an audience. And we're using our branch lab and Dentsu audiences side by side with a lot of the traditional players in the space, and using it in everything from our programmatic campaigns to our social campaigns and streaming campaigns, as well as addressable continues to become more and more prominent.
B
Great. So once you've identified audiences that fit into, like, narrower groups, whether you're not mapping but assigning, working with demographic groups and regional groups, how do you then locate those audiences? Is that something that you all do at branch Lab or is it more a data exercise?
C
Yeah. So the historical way to do this was you would have an audience brief. You would pick up the phone or send an email to the legacy audience vendors. They do their thing and then they ship that wherever you want to ship that to. The integrated approach is these models sitting on top of multiple sources of data with an agent that now Brad and his team can talk to the agent, explore that patient population and build an audience. Takes a couple minutes. And that could be one audience, it could be 100 based on whatever you and the agent want to explore. From a distribution standpoint, you distribute it straight from the integrated platform and it goes anywhere that you want to go. Social, ctv, you know, programmatic to a dsp, wherever. So the front to back process takes I don't know, five or 10 minutes versus say, you know, four, five, six months.
B
Right. And there's no need obviously to buy health media here. It's an audience buy.
C
This is an audience buy. So we're, you know, we're just shipping it to where your media buy is taking place and informing that end destination, which house, which device should receive an ad and which.
B
Yeah, really interesting. So love to sort of shift again and talk a little bit about the future and what we might see in the AI, health and pharma vertical. There's so much promise around personalized healthcare, like the cold exploding world of peptides. GLP1s being the first huge one to find a mass audience. But there's other peptides too that do different things in the body. And you can imagine a future that's AI driven where we have more sort of a personalized cocktail or, you know, a shot, who knows what it looks like. But in this world, how, how, how does the AI powered personalized health care journey impact how pharma marketing happens? I mean, is there still a need for ads or will AI just send us our drugs and invoice us?
D
That would be nice, wouldn't it? No, I think there will always be a need. I mean, look, the two of us saying there's a need for healthcare advertising, but no, I think the level of personalization will always be limited to an extent. I think Josh and I were talking backstage about some of the point of care companies that we work with have started tapping into opt in patient data for the ability to reach people as they're logging into their patient portals and things like that. So that's been a step towards really not necessarily one to one, but one to smaller groups of patient audiences. I think as long as HIPAA is in play, we're going to stay away from true personalization. But between opted in patient data and the ability to really further segment the audiences that we're targeting, you can get to a more detailed level of customer that we're targeting.
B
Josh, would you anything you'd add there?
C
I just, I like everything, agree with everything here. I think also just like the concept of, of patients becoming like more like better educated and advocating for care more, I just think it's something that will be a net positive. So I think the more, you know, the more we move in that direction, I suspect the better health outcomes that we see broadly and I like to think that marketing efforts, advertising efforts actually play a role there. You know, when we're not just, you know, running the same linear television spot to the whole universe, but you know, trying to better speak to folks specifically while considering privacy and regulations. I don't know. I mean, I'm hopeful that we can see an improvement in health outcomes as a downstream effect of that.
B
Got you. Great. And then a question on the big AI platforms and how quickly they are going to be able to move here to capitalize on what's obviously a huge opportunity around paid media and healthcare. If you had to guess, what do you expect? How long before we may see a, a product or an ad offering that clearly serves this vertical?
C
Yeah, I mean, I think from a, from an H, from a HCP perspective, targeting physicians perspective, we all, we already have open evidence. OpenAI will fast follow, I believe from a DTC perspective. The, I mean, I don't know, maybe reading the tea leaves too much here, but the integration between OpenAI and Critio is suggestive that OpenAI will allow third party data, audience data on that platform long term. And who knows, maybe long term means 18 months, I don't know. But 18 months is a very long time in the long term. Yeah, I think. Will the agents replace the search portion of the industry? I think probably, but again that's in the moment, like I'm diagnosed with something, I have something, I'm going to ask an agent about it. There is also the majority of pharmaceutical advertising which is focused on chronic care for conditions that you live with the next 20, 40, 50 years. Yeah, um, and we're not asking agents about those things every day. Um, so audience, you know, audience target would be a means to get in front of those folks.
B
Yeah. Brad, anything you would add on the timeline before we may see some mature AI ad products?
D
No, I think the ChatGPT health thing is going to be something to watch. I, I know they're keeping that separate from ads for now. I'm, I'm curious to see what adoption looks like for that maybe six months from now and how much people are really going to trust uploading their patient chart, their Apple Watch data, whatever it may be. Good point to an AI plat and seeing what comes out of that. So advertising aside, I think the adoption of that will be really telling because I'm sure if they see that people are using it, they're going to want to monetize it.
B
Yeah, gotcha. Just. I'd like a 20 second answer from you, Josh, on the question of Make America healthy Again, the Maha movement. Because if we don't mention it, I feel like it looks like we avoided it, but we're not avoiding it. And regulation in general is a factor in this space.
C
Yeah. So plenty of rhetoric from RFK and others about barring pharmaceutical advertising on television with the thesis that the reason why that occurs is to influence news. I think a more important voice here is probably Marty Macary, who's the head of the fda, who properly recognizes that that's not why pharmaceutical advertisers put TV advertising out in the world. It's because it works. That's also been a First Amendment issue. Like, it's been challenged and it's been held up. So that's kind of that. I do think moving forward, there could be further regulation on how data is used in advertising. We, we actually want to like. It's my opinion that there probably should be more regulation in the way DA health data is used in healthcare advertising. So we, we're trying to build the answer for that while also encouraging regulations to catch up with what's now possible with new technology.
B
Right on. Thank you, guys. We're going to leave it there. All right.
A
Thanks for tuning in to another episode of the AdTech Godpod, a podcast for the people about the people. Stay connected with me for more insights, trends and interviews in the realm of ad tech. Don't miss out on the latest updates. So follow me on X Instagram and connect with me on LinkedIn. Don't forget ATG Slack community has insights, networking opportunities and jobs. Keep the conversation going and stay at the forefront of ad tech innovation.
Podcast: AdTechGod Pod
Host: AdTechGod
Guests:
This episode explores how AI, and specifically neural networks, are fundamentally changing the approach to healthcare and pharmaceutical advertising. The guests break down shifting patient behavior, the challenges of data privacy and compliance, and how neural networks are being leveraged for smarter, privacy-preserving audience targeting. Throughout, the panel discusses the evolving healthcare advertising ecosystem, speculates about the future of personalization and AI involvement, and addresses the impact of regulation.
On AI's reach:
"The stat that everyone's kind of focused on recently...40 million people per day go to ChatGPT and ask a health related question." — Josh Walsh [02:45]
On patient AI behavior:
"70% of that is happening outside of clinical hours, which is kind of fascinating. I think that's the real behavior change." — Brad Fox [03:58]
On privacy and modeling:
"You can never see the data. It's never exposed to you." — Josh Walsh [11:23]
On old vs. new audience targeting:
"The way that patient audiences have been built in the past treats those two people the same, and as a result, brands speak to them the same." — Brad Fox [13:56]
On regulatory direction:
"There probably should be more regulation in the way health data is used in healthcare advertising. So we, we're trying to build the answer for that while also encouraging regulations to catch up with what's now possible with new technology." — Josh Walsh [22:38]
The episode delivers an expert roundtable on how AI—accelerated by patient and physician adoption—is reshaping healthcare advertising. As search and retargeting wane, neural networks are paving new, regulation-compliant paths for reaching and engaging patient audiences. The next wave of innovation will see further personalization, greater collaboration with AI platforms, and an ongoing conversation between technology advancements and evolving regulation. As the guests agree, all eyes are on AI platform monetization and the real-world impact on public health outcomes and advertising strategies.
Recommended for:
Adtech professionals, digital marketers in healthcare, compliance officers, and anyone interested in the intersection of AI, health, and advertising.