Episode 30: Privacy, Accuracy, and AI – The Future of Healthcare Content Marketing

In this episode of the Digital Clinic, we explore the strategic implementation of AI in healthcare marketing with Diane Hammons, Director of Digital Engagement at WG Content, uncovering how forward-thinking organizations are navigating regulatory requirements while unlocking AI’s potential.  

The conversation reveals practical insights into how LLMs are fundamentally changing healthcare marketing, from requiring new approaches to content structure and user experience design to the need for AI literacy training even in highly regulated environments. Whether you’re navigating compliance requirements or seeking to maximize AI’s impact while maintaining the trust and accuracy healthcare audiences demand, this episode offers actionable strategies for balancing innovation with regulatory reality. 

Listen & Subscribe:

Introduction to Diane Hammons

Aaron Burnett: Your title is Director of Digital Engagement. That is an intriguing title. Tell me what that entails. 

Diane Hammons: It is a title that came about with a lot of intention, and I appreciate that. Because at WG Content, one of our pillars, one of the things we’re known most for is building relationships. That’s our mission. We build relationships one word at a time. So when I think of Director of Digital Engagement, I then take that as building positive, healthy relationships with technology, our clients, internally certainly. So that whole engagement – it’s experiential, it’s relationship building. It’s more human. So yeah, I appreciate that title was supposed to, operator, director, blah. Sure. 

Aaron: Yeah. Tell me a little bit about your background. I was looking over your LinkedIn profile, and it appears to me that this position is the culmination of several important things that you’ve done throughout your career. 

Diane: It’s funny you make decisions, and then it does seem – I’m intrigued that you saw that linear. 

Aaron: It looks like a plan in the rearview mirror. 

Diane: When you’re just moving through and making decisions, moving where maybe your strengths and interests are, and all of a sudden you end up here. And if you tell one person, “I started as a graphic designer” and they don’t know the in-between, it might be like, “How did you get here?” 

But you’re right. A part of, I always say being a designer – at least back when I started, nobody was allowed to use Mac, not in a corporate environment. So really every designer also right away became a troubleshooter. You had to know a little bit about tech. You had to be not afraid to get in and poke around and try to fix things because IT wasn’t going to support you. 

And all the places I worked and corporate requirements I worked, it was, “Okay, you could use them, but you’re on your own.” It’s okay, most we will be. I think that naturally then designers in general have less aversion to technology change, and so you adapt it as you go. So there’s a little bit of that mindset carried through. And then, as the growth of my career, it becomes not just design, but focused on marketing, focused on brand, focused on strategy. And then you get to the place of not just healthcare, where I spent most of my career, but now working for a JD Edwards consultancy – completely different environment, different audience. 

We’re talking manufacturing for the most part, manufacturing industries. But now I am a single marketing person surrounded by technology-focused people – software developers, cloud, network specialists because they’re also a cloud provider. So learning all the lingo, learning what makes things successful. Learning about a company like that is certainly – they were talking AI eight years ago. So then coming to WG Content, it was great. I couldn’t not apply what I learned. So you start to look around and you’re like, “Holy, all these things should be talking to one another. What’s our source of truth? What are we doing with this data? And oh, here’s this thing called AI.” 

I did just naturally then fall into “We need to focus on this.” And I really give a lot of credit to our leadership – forcing not waiting, saying “We need to start exploring and learning and not be,” as I say, “dig your head in the sand about how might this impact what we do as creatives.” Sure. Of what? Let’s approach it positively and see what it can do for us to benefit us and help us evolve. That’s when my role got carved down to focus more specifically on technology. 

Healthcare Content Strategy Fundamentals

Aaron: WG Content focuses on content strategy, content production for healthcare in particular. How does content strategy and how does content itself differ for healthcare versus other industries? 

Diane: I think we’ve seen for years now, there’s been that trend to be more casual, more conversational in our business speak. You don’t have to be quite so formal, all of that. But with healthcare, it’s always been story matters more than anything because what’s really going to get you to connect with the system, connect with a provider, but those stories – understanding that how much that person, that provider deeply cares, understanding how much they made a difference in a person’s life. 

And so more than maybe a lot of other industries, story matters. As content generators, as writers, I see even in the world of AI that we still have a place. Never say never maybe, but I can’t imagine a world where that human-to-human connection isn’t needed for some of those stories to be created and resonated. 

Regulatory Considerations in Healthcare Marketing

Aaron: How is the approach different for healthcare than other industries? Principally because of the sensitivity of the content, the regulatory environment. How do you need to approach that work differently than you might for other industries in general? And then we can shift to AI as well. I’m very curious about the AI world that is powered by AI and how that applies there. 

Diane: Yes. Obviously it’s highly regulated. Some other industries are too. You do have to – for those who maybe aren’t in healthcare, the marketing is – they have to be very careful about the when and the where and how they’ve received information. What databases it’s coming from, what’s being collected, and when they, when and how they choose to use that information to market to potential patients, existing patients. So that means being sensitive in what’s shared with us and us respecting that sensitivity and what we do with that information that might be shared as. 

Certainly safeguarding patient names. It may – it will, if it’s shared with us, it’s going to eventually be public, right? Most of the time we are creating white papers, blogs, web content, all of that. So it will be public. But in the meantime, you still want to safeguard and hold true and not make assumptions about what can or can’t be made public. Our policy is the client’s policy. We keep track of each individual client and respect their preferences. 

Measuring Content Performance and ROI

Aaron: One of the key things that our internal stakeholders need to do is to clearly demonstrate the value of a content strategy, a content initiative to those folks who are providing budget, who are managing the business overall. What approaches do you take with clients to either ensure that the right approach to analytics are in place or implement some other measurement protocol that does tie back to business value? 

Diane: That definitely comes through most when we can have the biggest impact when we’re looking at those large overhauls, website overhauls, when our content strategist can get in and look at Google Analytics. Look how things are performing now, and make those sweeping changes that aren’t just one-off.  

But you can say, “Okay, now that we’ve done all of this, now let’s go back six months, a year and see how things have changed.” That’s the type of – when we’re looking at analytics, that’s what we’re looking at. How’s it trending now? Is it improving? Is it going up? And can we be sure that is related to what we did on the page, as opposed to – so we’re very aware of not claiming that this spike is because we made a change in content on the page. If you happen to be pushing a marketing campaign at the same time, sure. It’s likely you were also driving more people there. So we’re very aware. We ask a lot of questions to make sure we’re evaluating fairly and we can make sure that what we put in place is really having an impact. Or do we need to continue to go back and evaluate and tweak it some more. 

Aaron: Has the approach to measurement changed in a world of LLMs? Are you now measuring on-site and presence and sentiment in LLMs and other things? 

Diane: I would say that’s where we are at right now as a company where we are aware in looking at AI visibility tools so that we can start to – as we said, increase the literacy and help our clients understand that this is a thing that. It’s not just anymore where it was hard enough to think about Google and guess what was impacting, what could have the positive impact on the algorithm. And that was hard enough when you only had one thing to focus on really. And so now it’s getting everybody to understand that it’s not just Google anymore. It will be with AI mode, but ChatGPT continues to get more and more visits, more and more month over month. It’s important to understand support, to continue to look at your visibility across all, which is going to be harder. That’s going to be a lot harder if our marketing managers or our partners were overwhelmed with just observing Google. 

Imagine now they have to observe across all the LLMs. Yeah. Which will have their slightly different personalities and maybe trying to understand more and more, get more data around where do your particular patients live? What channels are they using? We were thinking before when we said that meant social. Now it means LLMs as well. Yeah, so it’s a lot broader, and in that way I think our clients are still going to need our help because it’s so much to monitor. So we’re looking at AI visibility tools. We’re looking at just even trying to help clients understand where they can, how and what kind of things they should go out and look just manually. What types of things should you go out for? Where we used to say, “Put a keyword in and see if you show up.” Now it’s “Consider these phrases, have a conversation over here. Then try to duplicate the conversation over here and over here.” 

Adapting Content Strategy for AI

Aaron: Have you started to modify content strategies to account for LLMs? Are you now thinking of producing content that not only ranks on a website, but also is prominently figured in AI environments? 

Diane: If we can call them best practices when this is such an infancy, this AI, but – 

Aaron: New practices? 

Diane: Yeah, new practices, yes. We are in agreement with a lot of what we read out there, that it’s a lot of what we were already doing for our clients – following SEO best practices, but also making sure when you’re looking at the structure on the page, schema markup, all of that. So it’s a lot of reinforcing business as usual, or at least reinforcing all the things that we’ve been asking you or suggesting you do all along. But on top of that, now we’re starting to find – here’s a great example. We just uncovered last week that it might be even more – we might start to give more recommendations about how the page is structured, the actual user experience of the page. AI mode rolled out, and we could more apples-to-apples compare a conversational experience in ChatGPT with that and Google and looking at finding a position. If you say more nuanced because that’s what you can do with conversational – “Help me find a female OBGYN that’s accepting new patients and is in this area.” 

That’s all in one. You couldn’t do that before. What’s coming up? And we found in this one area where we knew there were four major health systems, only three kept surfacing in either two. So we asked, “Why didn’t you surface this one? Do they not have high-ranking OBGYNs?” They said, “Oh, certainly, but it’s so much more focused on intent now. It said, ‘But you asked me who’s accepting new patients? I can tell that with these three because it’s listed on their website. I can’t tell that with this fourth one because it’s not.'” 

Aaron: Oh, interesting. 

Diane: That’s a little more information than we would normally give when we’re talking content strategy. We’re seeing you might miss the button, you might be missing somebody of your user interface that you wouldn’t go back to the table with. So suddenly, it’s just – there’s a lot more to it than there used to be. 

Aaron: That’s interesting though, that you can query and ask for the intent signals that are and are not being used. And you effectively get an ad hoc guide to optimization. And that’s all we can do right now. It’s being curious and continuing to ask why. From some of the background research that I was doing, implementation or at least exploration of AI at WG Content is a big part of your role. I think maybe for the last year or so, you’ve had some working groups within WG Content. Can you tell me a little bit about what you’ve done, and then I’m very interested in what has worked and what has not worked? 

Diane: Well, when the leadership realized we’re not going to wait, we need to be aware, we formed – we call it the AI Pathfinders team. And then we asked for volunteers within the company of people who would want to be, were open to exploring. And there was a great sign that some of the people that volunteered, or people you would expect to, the people were quick to raise their hand because they’re always eager and enjoy learning and whatnot. But there were also people who said, “I’m intentionally – I’m pushing myself. This scares the heck out of me, but I’m intentionally going to join because I need to force myself to face it.” 

But it was good because then that gives you a range of experiences and opinions. So from there, the important thing was finding those business cases, not just, “Okay, go try ChatGPT, go try Grammarly. Just try it.” It was, “What are our pain points? And how might – small, they could be very small because I’m not trying to intimidate anybody in the beginning. We just want to get comfortable with the concept of AI. So starting small – what are some of our pain points? Let’s – we have a theory that AI could help with this. Now let’s explore it.” 

So things like note takers, email integration, and that type of thing. And those small things – and they sound like, “Okay, big deal.” But I think that’s what made us successful in a team like Pathfinders is it gets people over that hump. Yeah, they just needed to have some little wins or little easy and non-threatening ways to start working with AI. And then they could get a little bigger. So from there we could go into things like custom GPTs, or someone raised their hand the other day and said, “I want to start looking into agents.” Awesome. That’s great. And again, not because “I just want to look at agents,” but because we’re realizing, “Okay, we have all these custom GPTs that are working now, wouldn’t it be great if they could be strung together?” 

Aaron: Yeah, you mentioned custom GPT. Can you give me a couple of examples of custom GPTs that you’ve created that do work well for you? 

Diane: Yeah. I’m trained designer, not writer, so I’m one of the few people on the team that take a lot longer to write things because it’s not natural for me. Here comes Brand Guru – it’s part knowledge base, so you can double-check things. All those things like, “Do we use a serial comma or not?” But then it’s also a writing assistant, especially for people like me, for internal communications. For, “Ah, I shouldn’t be sweating and spending an hour on an email invitation to a lunch and learn.” Yeah, just get it out there. Stop overthinking you, Diane. So Brand Guru does that for me, and we call her Jeannie. So Jeannie does that for me, but she can help review even things we write for our own blog, just to see, “Hey, did I get the tone quite right?” It’s kind of like an editor before the editor to hopefully lighten our editor’s load. 

So we’re applying that same approach now to a lot of our client’s work. So we have same like a guru for a lot of our main clients. So same way we can run it through just to check, “Did I miss anything? Did I get all of the – did I capture all the writing styles correctly?” Because they vary. They vary so much overall, our client base. So did I get them right? Double-check me. Did anything in the tone sound off that I should go in and redo? What we’re not talking about using this for all-on content generation. It’s just that check. As acceptance is growing and the ability of these LLMs is growing too, we’ll simultaneously – some of our clients have given us that permission in certain situations. Interesting. Sure. So we’ve built another one recently that helps write web pages that are just really treatments and conditions. It doesn’t have a lot of marketing fluff. It’s pretty nuts and bolts, and the clients and it’s, “Okay. This is a perfect place for that.” 

So in that case, we built a custom GPT, not just that knows the branded voice – it does – but also spits out the information in exactly the format that they need. They can analyze the source content and structure it exactly as they need it, and it’s taken a lot less time. 

Balancing AI and Human Expertise in Healthcare

Aaron: One of the things that we think about and a reason that we’ve not used AI to actually produce content is, in healthcare, accuracy is sacrosanct. And in terms of the performance of the content, authorship matters a great deal as well. You were referencing it, the intersection of authorship and also the right schema markup to ensure you’re making the most of that authorship. Do you use AI to ghost for authors, for physicians, scientists, researchers, that sort of thing? 

Diane: It’s thinking about the best place to use it. Yeah. So again, in this case we were talking about treatments and conditions are pretty universal. Only so many ways you can say it. That’s not what’s – yeah. So yeah, that’s not what’s really – it’s important to have, but it’s not a differentiator so much. If you’re familiar with Marketing AI Institute, I’m guessing everybody is at this point. Yeah, so Paul Roetzer – they shared the Human-AI Scale, and we apply that to our healthcare writing. 

So you’re looking at zero, I think might be all AI to all human anyways. Yeah. A scale. So now we can talk – we’ve adopted that as a tool, as a method of talking to our clients to say, “Okay, where does this kind of content live? Does it – is it important that it’s all human? Is it important that it’s – could be – is the middle like we’re going to write it with a human, but then we can put it in and we can help fill the gaps maybe with AI? Is it AI can do a first draft and human just can review it?” Right now, there’s nothing in healthcare that we would say talk fully autonomously, give it to AI. No, because of that accuracy you were talking about. Everything is human in the loop. We can’t stress that enough for accuracy, for brand integrity. But when it comes to authorship and expert expertise and how that shows up in rank continues to rank. And that’s totally where that should really still be all or mostly human. 

Aaron: Do you also, as part of your content delivery services, provide design support to your clients? And if so, have you been able to successfully use AI to support design services? 

Diane: The short answer, no. Yes. Yes. Okay. Then I feel in good company because that’s good. That is the trickiest area for us. It is, yes. So yes, we do support. Which scientists need it. But wow, that’s a whole other thing with prompting. Oh my gosh. And another area where I’m talking to the Pathfinders about exploring a little bit more because I theorize. It – some of the best prompters are going to be writers because they can visualize, they can put into words that’s in your head. It’s like blank page for me, staring at that prompt window because I want to generate some kind of image like, “Oh, how do I explain that? How do I say that?” 

Writers are great at that. I think there’s even more opportunity for them to continue to teach me and others how to get those words out of your head. But for also of them to continue to evolve as more visual thinkers, just image generation within some of the basic frontier LLMs. Yes. Wow. Even though now they’re better at editing and maintaining some of the core focus. Sure. Your visual – the image, it’s still, “No, that’s not what I said.” 

Aaron: Yeah. But I think that your admonition to your internal teams, which is not yet, but eventually, is definitely our perspective. So we’ll continue to perfect and refine the skills that we need to have and the systems that we need to have internally in the fairly strongly held belief that the technology will advance in lockstep. And it will be there as we become more proficient as well. 

Diane: As long as we continue to explore, which is what we’re doing in – yeah. That unfortunately, I don’t know that everybody’s doing, and part of it I think is that their hands are tied. If they want to do it, they’re going to have to be doing it on their own time. Is that what you’re finding as well? 

Aaron: Like you, we’ve made AI a big focus internally. Again, we don’t use it to deliver client assets yet. There are some specific areas where we will focus much more on AI-enabled insights and data analysis and strategy development and that sort of thing. And sort of democratized everyone looking at AI, using it as much as possible to support internal processes. I think what’s been curious and surprising is the types of people who’ve embraced AI and are excited by it, and the types of people who are at the opposite end of the spectrum – dismissive, don’t like it, skeptical, far longer than is justified by anything they could see. And it’s often surprising who’s in which camp. 

Diane: And do you find that internally as well as externally? 

Aaron: I see it more clearly internally because we can have those kinds of conversations. Although I think that it is true to say that at least within Wheelhouse there’s pretty much universal enthusiasm for AI because of the ways that we’ve been able to implement it, and it so clearly delivers. One of the things that we’re about to roll out to our clients is AI-enabled business intelligence. So you can use natural language queries, and you can ask effectively the data warehouse that contains all of the performance data and first-party data for our clients. You can ask questions, and you don’t need to know SQL. You don’t need to know anything other than how to frame a question. You can ask it to model scenarios. “What if I did this? What if I did that?” And we’ve been working toward that for a long time. 

Like you, we evaluated lots and lots of platforms and providers, and I have never seen so much vaporware in this space, just really good presentations and really good descriptions of what it will do until you get to the point where it actually has to do it. Well, it’s on the roadmap, right? But we’ve finally – we’ve been able to put together the right set of platforms and services to deliver on that. 

Diane: That’s good. Yeah, we’re working on a pilot with our third-party software developer and Salesforce right now to start to do a little bit of that with all of our client data in Salesforce. So how do we get to that? We can get – we have pull reports, we can do all that, but yes, the analysis, the easy analysis. 

Aaron: Yeah. And creating a context in which a non-technical user can actually get the answers to the questions they need or can create a dashboard that they need that would make them much more effective without having to go through a BI specialist. 

Diane: Yeah. Yeah, you’re exactly right. 

AI Literacy and Industry Adoption

Aaron: You mentioned earlier on that you’ve really shifted to a much more consultative approach with your clients. Can you tell me a little bit more about that? 

Diane: We’re excited and we’re passionate like you, we – because we’ve been doing some exploration. And even though you’ve got different sides of the camp, you all – some people may be a little bit reticent, but they’re still doing it because they realize they need to, at least to a degree. Sure. Like the passion might be there, but they don’t have a place, a playground for it. And again, I get that. A CIO, CTO of a health insurance, health system – they can’t afford for a slip-up risk. The risk is high. But at the same time, I’m a little bit – I’m concerned that in the interest of making sure you get it all right and it’s all buttoned up, which is really hard to do when it continues to change all the time – that in the meantime, people, the employees are a little bit floating in the wind, right? You can lock down your system. It doesn’t mean they wouldn’t do it over here. 

So to me, the miss is you have to run AI literacy. At least in parallel with – you can’t wait. An element of AI literacy has to go along. So for example, our AI Pathfinders, they were inherently learning a lot more because they were going through some of the use cases. The rest of the team do – we just let them – you didn’t join, so you don’t get to learn about this? No, absolutely not. So certain things we continue to change to the full team because they need to at least understand the basics of security and why it matters and what it means to train the model or not, and how to evaluate what should or shouldn’t be uploaded. You got to at least share that much with them while you’re figuring everything else out while locking down on trying to be very conscious for the whole marketing gets handcuffed. 

Reminds thing up right away is those early days of social media when people are like, “Ugh, people are going to waste time. So we are shutting down social media. Nobody can access it. We’re going to block it for the whole company,” and you’ve got marketing going, “But that’s where our consumers are. And I’m supposed to build for it, and I’m supposed to monitor, and I’m supposed to analyze it, but I can’t access it. How am I do that?” And that’s where I feel we’re at with AI. But I feel like that it’s such a miss because I hear too many of our clients that don’t have access. They can’t experiment like we are. So as a consultancy, that becomes important for us then to be their eyes and ears and hands to do that when they can’t. But it’s unfortunate. 

The Future of AI in Marketing and Final Thoughts

Aaron: I’m glad you’re doing that. Anything else we should talk about that we haven’t touched on? 

Diane: Who are your go-tos? Where do you get your information? How do you continue to stay up on things? 

Aaron: I read fairly broadly. I don’t have particular sources that I go back to again and again because everybody in this space has their own particular agenda. The landscape, the technology, the use cases are changing so quickly. They’re changing more quickly than people are able to adapt. And so I have to confess that I tend to hop from person to person for different kinds of things. There are certain folks that I follow principally on YouTube and then their own sites for agentic AI. That’s a very interesting space where again, a lot of smoke and mirrors, some real – getting to the people who actually are very practically minded and talk candidly about what’s possible now and what’s worth doing and what’s just a science experiment. 

In the early days of AI, I was very interested in the people who were pontificating. “This is what I think about AI. This is what I think it means. This is what I think will be able to do that.” I don’t find that very useful anymore. Everybody has thoughts about what they think it will mean. I really gravitate to people who say, “This is what you can do, and I’m actually going to show you how you would do that,” or, “I’m actually going to – I’m not just going to describe the potential for AI and conversion rate optimization, for example. I’m going to give you five different approaches in very detailed, well-structured prompts, and I’m going to tell you what the important elements are within those prompts and the underlying data or knowledge that you need to include in a custom GPT or a Claude project.” 

So I think that probably is the most succinct shift in the information that I now pay attention to. I pay attention to people who know how to do it and show you how to do it and give you practical, real information. And I’m not very interested in people who have a lot of thoughts and pontificate. I belong to a network of CEOs called the Entrepreneurs Organization. 

Diane: Cool. Yeah. Christy Pretzinger does as well, our CEO.  

Aaron: And EO has a really robust AI-related community. And a couple of WhatsApp channels where information flows really quickly. 

Diane: I think she may have just joined that. She was talking about it. Yeah. 

Aaron: It’s worth joining. It’s a lot to keep up with. When there are new announcements, you can very quickly find that you have a hundred WhatsApp messages in a day related to new AI technology, but you don’t have to respond immediately. You can consume those in sort of bursty fashion. And again, it’s a group of people who are very practically minded, super pragmatic, and they’re talking about what you can actually do and then sharing what you can actually do as well. 

Diane: That’s what I love about all this. And we found the same. The people are very okay with sharing. It’s “Let’s put all this together.” It doesn’t feel quite as possessive. We have a good community, AI community here in Cincinnati. That’s that way, and that’s probably one of my favorite things to do. There’s monthly meetup. But it’s just hearing what’s possible, and you’ve got developers who will say, “I hear something I’ve been tinkering with.” 

Aaron: It’s a really interesting space. It’s wildly different from any other major technology. I’m just going to use the word evolution. Evolution doesn’t really cover it. Seismic event, whatever. In the past, let’s say the rise of the internet, the technology was the thing. The IP was the thing, and that’s what people developed, protected, defended. That’s how vast sums of money were raised. The IP is not the thing in AI. It’s actually the utility that you derive from it. So our thesis around AI is because we’re in healthcare where our clients can’t simply layer AI over the top of their data. They have to be in a HIPAA-compliant environment. They have to have much more rigorous data protections, and we have a HIPAA-compliant data warehouse that gives us a really interesting opportunity to deliver value. 

But the AI – our philosophy is it’s moving so quickly and we don’t have tens of millions of dollars. So we’re not going to attempt to build much of our own proprietary technology. We’re going to assemble best of breed, and our point of differentiation will be the way that we can bring together client data in this HIPAA-compliant environment and deliver insights on top of it in a way that maintains privacy. It’s not the service on top of it, it’s not the AI. I think the AI becomes the commodity through which service and value is delivered. 

Diane: I think you’re totally right. And we in our small way are starting to realize that too. It’s the – a little bit shift in thinking of “Can’t we build a GPT for you?” or “How do we just – the one and two will help the flow of your work.” 

Aaron: Very interesting. Yeah. I’ve really enjoyed the conversation. Yeah, thank you.