Videos
AI Accelerates Audience Intelligence and Measurement at Bayer
Insights
- AI is enabling more sophisticated audience creation by combining behavioral signals, contextual factors, and predictive models to improve targeting precision.
- AI-powered measurement tools are helping marketers uncover insights faster and democratize access to analytics across the organization.
- Successful AI adoption requires a pragmatic, test-and-learn approach that balances speed, experimentation, and responsible governance.
How can marketers use AI to improve both the speed and quality of decision-making?
Jeremy Rose, Head of Unified Marketing Measurement at Bayer, discusses how AI is transforming audience creation, marketing measurement, and organizational decision-making. Working at the intersection of analytics, media, and marketing, Jeremy shares how Bayer is using AI to move beyond traditional targeting approaches and unlock more precise, responsive customer engagement strategies.
Jeremy highlights three critical shifts:
- How AI is helping marketers build more intelligent audiences by combining purchase behavior, contextual signals, and predictive modeling to improve targeting effectiveness
- How AI-powered analytics and conversational interfaces are making measurement insights more accessible across marketing organizations
- Why successful AI adoption depends on collaboration between analytics and marketing teams, supported by a disciplined test-and-learn mindset
Drawing from Bayer’s experience in consumer health marketing, Jeremy describes how AI is evolving from an efficiency tool into a strategic capability that improves both audience precision and business outcomes. He explores the growing role of machine learning in audience segmentation, real-time optimization, and insight generation, while emphasizing the importance of keeping humans involved in critical decision-making. Jeremy also discusses the challenges of training AI systems, the need for cross-functional alignment, and why marketers should resist the temptation to move too quickly without validating results. Grounded in a practical approach to experimentation and measurement, this conversation offers marketing leaders a clear perspective on how AI can drive smarter marketing while maintaining trust, accountability, and long-term value.
This interview was recorded at the 2025 ANA Measurement & Analytics Conference in Chicago as part of a partnership between Infosys Aster and the Association of National Advertisers. Click to learn more about the ANA and the
Dylan Cosper:
I'm Dylan Cosper with the Infosys Knowledge Institute. Today I'm joined by Jeremy Rose, Head of Unified Marketing Measurement at Bayer. Today we're going be talking about AI and marketing. Jeremy, thank you for joining us today.
Jeremy Rose:
Of course.
Dylan Cosper:
So I want to start off maybe a bit like high level here. How do you see AI changing your marketing organization?
Jeremy Rose:
Yeah, in a few different ways. At fundamental spot right now, I think in audience creation is probably the first and foremost. The way we do it right now is a little probably unique. We're a CPG, so we don't have access to first party sales data or anything like that. So we're using and syndicating audiences based on NCS data for past purchase. We're then pushing that into our various different publishers and platforms.
Jeremy Rose:
But with AI, we're starting to try and get a little bit more creative in how those audiences are being grouped and targeted trying to get a little bit more on who the consumer is as opposed to just purely based on purchase history. It's extremely important, but now we can start to get a little bit more refined in how we're creating those audiences.
Jeremy Rose:
And then also in terms of real-time adjustments or targeting, we have a few different indicators for us during, let's say, peak seasons, whether it's allergy, flu, et cetera. So if flu count is high, where it's high, we can push in. Same thing with pollen council. We can really start to be a little bit more creative in terms of how we're creating audiences and how we're targeting and pushing down into market.
Dylan Cosper:
Given what you're doing around audiences, have there been any surprises that you've encountered with adopting AI to improve how you're creating audiences?
Jeremy Rose:
We haven't had any bumps really, if that's what we're going after from an audience creation standpoint. Creative, sometimes we have little bumps. We're still pharma. So we have to have the right approvals for any creative that's being pushed out and all. So it may not be as fast in that regard for us right now. But I think that's stuff that we're working through to try and get better and faster at.
Dylan Cosper:
Have there been any kind of positive surprises, you would say, in the AI adoption around audience creation?
Jeremy Rose:
Speed is the biggest one. I'm sure that's a common answer, right? Definitely speed, getting faster. And I think, you know, one of the watch outs, not exactly surprises, is that there's so much talk in the industry about it. And that type of knowledge may not always be, not necessarily for us, but more widespread. It may not be living on the brand side, I think it's really on the vendor side where a ton of that knowledge and creation is, the ability to move very fast.
Jeremy Rose:
But on the brand side, the knowledge may not always be there. Thankfully, we have a good foundation to be able to adopt and put things into market quickly. But from talking to some colleagues across the board, that may not always be the case.
Dylan Cosper:
Going around kind of like decisions that you've made or plan to make, what would you say are some of the most important decisions marketing leaders need to be making right now to get the most out of their AI use?
Jeremy Rose:
I think just speed, right? And if we think about from a leadership standpoint, it goes beyond what we were just talking about in terms of audience creation, but how do you make snap decisions on the fly? Media planning is ongoing. It never really stops. Like, yeah, there's a media plan that's issued. You have beginning of year planning, but there's always adjustments that are being made throughout the year. Whether you're talking about in-channel adjustments, whether we're talking about cross-channel, budget cuts, incremental budgets, you have to be able to just move at lightning speed to be able to keep up with what's going on in the marketplace.
Dylan Cosper:
It sounds like it's related to AI, but it's almost probably organizational changes and processes to support that speed.
Jeremy Rose:
Yeah, AI is really an enabling tool, right, to help marketers make faster, better decisions. So we'll move as fast as we can.
Dylan Cosper:
Now you mentioned speed as kind of the value that you've seen so far. Is there any other areas where you're coming across value that maybe don't relate to efficiency or just productivity that you've seen so far?
Jeremy Rose:
So there are, right, because we're able to, now I spoke about this upfront, right, in terms of how we're creating the audiences and the makeup of those audiences, we're able to go a little bit, not just faster, but AI, the tools, the machine learning can mine those insights a lot better than any one person doing hunt and peck, for example, right? So we're able to get a little bit more precision. So while it may be speed or efficiency, but now we're able to really start to move the needle a bit more on effectiveness in terms of how we're targeting and getting those audiences to the right people at the right time.
Dylan Cosper:
How does that compare to what you expected with AI initially?
Jeremy Rose:
It's what we, I think, expected, but we're also leaning in and going stepwise. We're not just running like crazy, right? You don't want to just all of sudden hand over all of your decisioning to some crazy AI tool. We have to do a prep map. We have to do it of like stepwise. Lean in, test, see, and react to what the results are telling us. And as we start to see more and more, we're able to keep pushing the needle a little bit. So it is a constant test and learn, constant experimentation to see what's the right fit and how do we push in, right?
Jeremy Rose:
But we're starting to use it a little bit more. We're looking at how do we bring in propensity modelling to create audiences and all that. It may not be some AI tool, I'll lump it all together. How do we mine the data better and faster to really push things out into market?
Dylan Cosper:
Taking a step out a little bit on measurement, how do you see AI kind of changing the way you measure marketing effectiveness?
Jeremy Rose:
I think from a speed standpoint, number one, but number two, not in terms of how we measure, but the insights that we're able to get. If you think about a mix or an MTA, there were experimentation or whatever, there's so much data in there. You can have well-trained people to be able to know how to systematically mine through the data, but it's a slog, right? And it's really hard to get to all the insights.
Jeremy Rose:
So what we're starting to do is build better AI tools and build LLMs to sit on top of the same data set to try and get those insights to us much faster. What we're also starting to see in the marketplace and also we're trying to develop on it is how do we now put a ChatGPT feature on top of our own data stack. This way now you can put that into the hands of your average marketer who doesn't know data science. They don't know how to mine the data necessarily. They just want to ask the question. And now how do we arm them to be able to get answers faster?
Jeremy Rose:
There's only so many of us on the measurement or on the analytics side. We can't answer the entire organization's questions at that type of speed that necessarily everyone wants to go to. But if we have that type of AI learning, in-house we can get to some of that again faster. I know I've used that word a lot. But it's enabling our teams to react and put things into market and just move much, much quicker and smarter.
Dylan Cosper:
Yeah, I mean it sounds like it's not replacing the role in any way shape or form, but it is augmenting that role and they to be a little bit more savvy around maybe some prompt engineering or things like that, but it's also taking a bit of that work off of the measurement team and analytics to answer probably some of these maybe more entry level kind of queries and things like that.
Jeremy Rose:
And that's where my head is going is like, how do we automate and get rid of the things that are just piling up? And how do we better use our time to go after the big things? Think of the rocks, right? Get rid of the pebbles from the jar and let's just focus on the big ones. So sometimes a big rock in that sense will be how do we automate and build a system to deal with all the little stuff, right? And shove that out. So that's what we're trying to do again to what you said, not replace, but change how we're working to be able to go after the big things. And that more strategic value that AI is supposed to deliver.
Dylan Cosper:
Now, I'd like to get into maybe some disappointment here. I think we hear with like, there's AI tools that may not do exactly what you want, but what about use cases? Have there been any AI use cases that you all have pursued, either in a pilot or something, that have just kind of proved to be a bit disappointing?
Jeremy Rose:
Yeah. I think it stems from what I, it sounds all nice. It's tough to get there. An AI tool, especially if you're developing one, is only as good as its learned prompts. So you have to train it and teach it. So for example, if you're having a tool sit on top of a structured data set, you can't just ask it, hey, what's an ROI? What's my, like it has to learn. What does that mean? What, how do you do that?
Jeremy Rose:
Easiest way of explaining is like you have an entry-level analyst really really smart, but you got to train it. So it just takes time to get it up to speed to where you are light years ahead of it, but you know eventually it will get there so I think sometimes it's just that that learning curve is probably a little bit longer than a lot of people would have liked and hoped. At least as it relates to like mining some stuff that we have from a measurement standpoint.
Dylan Cosper:
What have been some of the largest challenges that you all faced in your AI adoption so far?
Jeremy Rose:
So I'll spin it a little differently, not just in terms of AI adoption, just straight up adoption. I think one of the keys is to make sure analytics and the media slash marketing team are really working together. And it's not just analytics or an AI tool talking down from a pipe like here's an output run with it. Don't question it. That's not the right attitude.
Jeremy Rose:
So like whether it's the human element or whether it's an AI element just trying to make sure that they're working together seamlessly that there is collaboration there. I think that's the biggest thing to help push that forward.
Dylan Cosper:
So do you think these challenges are getting better or worse as you've kind of continued on your AI journey?
Jeremy Rose:
It's for the better. Sometimes maybe harder to solve, but for the better, right? Especially if you're coming out with a collaborative approach. We all have the same end game in mind. It's just challenge that we got to work through and we know it's going to get us better in the long run, so it's totally fun.
Dylan Cosper:
I'd really like you to kind of maybe pull out your crystal ball here and looking ahead, what do you kind of see as the long-term implications of AI in marketing?
Jeremy Rose:
I think creative is probably one of biggest things. Not necessarily harder in the pharma space, but to be able to create content that much faster and cleaner and better than how you may think or be able to put it on paper is going to be really key. And then how do you realistically tie that content to a personalization, to an audience targeting, is I think really going to catapult us moving forward.
Dylan Cosper:
How are you and Bayer currently kind of approaching responsible AI? I mean I know being in pharma and it's more important than ever but is there a particular kind of approach you all have taken to responsible AI that you've seen to be beneficial and supportive of your efforts to experiment and adopt?
Jeremy Rose:
So I think we started to touch on this a little bit before, but it's really about just going with a pragmatic, stepwise approach of leaning in before just full pedal to the metal. And I would say that whether it's AI, whether it's any type of learning, before you just go all in on something, just test, ramp up, see how the market responds before you put too much investment in a thing.
Jeremy Rose:
Same thing if we're, forget about like audience creation. I know we talked about that a lot but like if we're testing new channels or new publishers that we haven't even been on before, ramp up. You don't have to go all in because if you're seeing something along the way that tells you, slow down maybe this isn't the right thing to do for your brand, you can peel back a little bit a lot easier than if you were going full pedal to the metal.
Dylan Cosper:
Before we conclude here, I'd love to hear from you and just all marketers are kind of going down this path now. That's one thing we kind of hear at all of the different conferences and regardless of what the focus of the conference is, everyone's kind of at the same point. Some are a little bit further ahead, some are a little bit further behind. What advice would you have for the one thing that you hope from this conversation marketers take that'll help them along their AI journey?
Jeremy Rose:
Take it a little bit slow. I know everyone is tempted to just start to go running with it, but these are major investment decisions. Another way I would put it is, are you willing to bet your job on it? Go a little bit slow, test, see how the market responds before you go run and do anything.
Dylan Cosper:
Excellent, I think everyone, not just marketers, could use that advice when it comes to AI. Well, Jeremy, thank you so much for the time today. I really appreciate you sharing your insights. Again, just thanks, really.
Jeremy Rose:
Yeah, glad to be here. Happy to help.
