“AI is now the first place people go to for answers,” according to Jesse Wolfersberger, global lead of I/O Health at Weber Shandwick.
“That chat box is the most important battleground in health—brand versus brand, but most importantly, truth versus superstition,” Wolfersberger said in an interview.
As AI plays an increasingly important role in health and pharma marketing, leaders at agencies and organizations who spoke with Fierce Pharma Marketing pointed to a remarkably diverse number of ways the technology is having an impact—and to the significant concerns and challenges that remain.
Image generation, for one, is a key application for many marketers. AI tools offer opportunities to create dozens of images in seconds at a relatively low cost, making them especially popular among organizations with limited resources.
“Even just creating an image of someone’s body and focusing in on a specific part to explain what to watch for was very costly to create in the past,” said Kyle Smith, Childhood Cancer Canada’s director of development. “Now, they can be generated in 30 seconds or less.”
The charity is currently constructing a website that will provide introductions to the 48 different types of pediatric cancer and what to watch for, a task that Smith said would otherwise be too time-consuming and too expensive to take on without AI.
AI’s image generation capabilities were also at the heart of a recent campaign from Incyte and Real Chemistry that was aimed at raising awareness of the often hard-to-see symptoms of myeloproliferative neoplasms (MPNs) and dubbed “Unseen Journey.”
“These patients with MPN look healthy, and they function in society, but inside they have a lot of pain,” said Mary James, president of analytics and insights at Real Chemistry. “So we asked them a series of questions, we interviewed them, and as they were answering, we created images using AI that reflect the experience. Now, you can share visually what was once unseen.”
Personalization and forecasting and focus groups, oh my!
Given James’ title, it’s not surprising that one of her main focuses has been on the analytical implications of AI, which she sees as intimately linked to another exciting potential use of the technology: increased personalization.
“We’re seeing more and more campaigns being tailored to micropopulations, which means we’re actually able to reach patients better now than ever with content that matters to them when they’re seeking the answers they need,” she said. “The combination of data processing and personalization is prediction. AI has the power to pull disparate pieces of data together and help to predict what the future could look like.”
Phreesia, which creates patient intake software, has long had a focus on personalization. That’s now being strengthened with AI tools.
“We pioneered the use of tailored point-of-care messaging to reach patients at key moments with messages that are most relevant to them,” said David Linetsky, president of Phreesia Network Solutions. “Now, we can leverage AI to further understand patient populations and help life science brands and government organizations reach the right patients with truly personalized and meaningful content.”
Linetsky added that the company is currently building tools to forecast its audiences and also analyze campaigns, the latter of which is a priority for IPG Health’s AI integration, too.
According to Graham Johnson, chief operating and product officer at IPG Health, the agency is using AI both to ensure campaigns are consistent with clients’ brand guidelines and to determine whether they are likely to face any particular medical, legal and regulatory (MLR) compliance obstacles.
“You can create a lot more content, but you still have to manage all of it and shepherd it through MLR and all of the other things that you need to do,” Johnson said.
One of those necessary steps is focus-group testing, something that IPG Health can now do with its LivingMirror application, part of EPICC, its AI-powered data and insights engine. With LivingMirror’s panel of virtual consumers, marketers can collect reactions to campaigns faster and at a lower cost.
As an operations executive, Johnson said he’s particularly interested in AI’s ability to streamline processes and create efficiencies.
“A lot of people are using prompts and generative AI to help them not just with concepting, but to write better briefs,” he said. “That’s an incredibly valuable thing.”
He continued, “Also, imagine you have a network where different groups use different brief formats, and you need them all in the same format. You could try and align all of those briefs across the network, and I think we’ve largely done that, but you can also use AI to do a lot of the busy work. I don’t know if people consider derivative content exciting, but I do.”
The trouble with LLMs
While the future looks bright for AI applications in many areas of marketing, challenges and concerns—from privacy to misinformation—remain.
For one thing, large language models (LLMs) like the increasingly ubiquitous ChatGPT often draw not only from articles from peer-reviewed journals and expert-vetted websites, but also from blog posts and other unverified online content that may lack scientific rigor.
Real Chemistry’s James referred to one study published by The BMJ in 2024 that was intended to test several popular LLMs’ safeguards against medical misinformation.
“Researchers gave LLMs information that seemed factual, and they asked questions like, ‘Does sunscreen cause cancer?’ and ‘Does 5G cause infertility?’” she said. “And all of the LLMs except for Claude were willing, with the right prompting, to produce information that was completely false and even provide figures to go with it.”
She continued, “These are high-risk, life-or-death situations. Sometimes, if you’re a patient asking for information from an LLM, you don’t know any better. Our clients are concerned about misinformation and even more about how to correct misinformation.”
Part of Real Chemistry’s answer to that lies in ensuring that the science-based information its clients create is available in machine-readable formats, to increase the odds that LLMs are receiving the most accurate, up-to-date materials.
“We’re betting that when a patient is trying to get the right information, the LLM will favor the accurate information from our clients,” she said.
Like James, Johnson at IPG Health expressed concern about AI’s inaccuracies in health contexts.
“AI is currently like a really promising intern,” he said. “It is certainly going to get better, and we’re going to put more and more trust into it, but the thing that would keep me up at night is if something happens inadvertently where you think a piece of software or a bot is going to faithfully execute X, Y or Z, but there’s a miscommunication or a misperception with an unintended consequence. In our world, that’s not OK.”
AI still needs a (human) hand
Smith, of Childhood Cancer Canada, pointed to a different concern related to a marketing buzzword of the moment: authenticity.
Much of the organization’s work, from lobbying to fundraising, is built on sharing the stories of individuals battling cancer. If its messaging is perceived as being inauthentic and created by AI, years of work building trust may be lost.
“When we look at our real-world scenarios—people with lived experiences and their families—how does that voice get lost?” Smith said. “When we look at policy changes within government, if there’s AI-generated storytelling behind that, it can cause a loss of trust, respect and that sense that this is a real problem.”
He added, “We may be able to use AI to throw some statistics in there, but we need to always lead with heart and with empathy.”
As Johnson noted, “You want to be an authentic brand; you want to be an authentic voice. And increasingly, if you’re working with a lot of assets that have been synthesized by an algorithm, it feels a little contrary to me.”
Clearly, the need for a human check on AI remains—especially as the marketing leaders interviewed for this article almost uniformly described a world in which, even if clients are reassured that problems like misinformation will be adequately addressed, they still don’t know where to begin amid the rapid pace of AI development.
“This new and relatively uncharted technology is disrupting a highly structured and regulated industry,” Phreesia’s Linetsky said.
“Each piece of content needs to be closely reviewed and approved, and this doesn’t quite fit with the AI landscape, in which everything is moving very fast and is highly personalized to scale,” he continued. “As the industry works closely with legal and regulatory teams to review and approve these new technologies, things will evolve, but uptake may take longer than desired.”
Weber Shandwick’s Wolfersberger concluded on an optimistic note, sharing his belief that while the path to the best use of AI may not yet be entirely clear, the industry will eventually find its way there.
“AI may be new, but the oldest rule in business still applies: Nimble wins. There’s no playbook yet for using AI in healthcare marketing, which means the only way forward is to experiment,” he said. “Some paths will dead end, others will get derailed by algorithm changes, but along the way we’ll discover what really works. Taking a few lumps is the price of staying ahead.”