The Real Role of AI in Healthcare Performance Creative

Announcer: Welcome to The Ignite podcast. The only healthcare marketing podcast that digs into the digital strategies and tactics that help you accelerate growth. Each week, Cardinal’s experts explore innovative ways to build your digital presence and attract more patients. Buckle up for another episode of Ignite.

Lauren: Hello, everybody. Welcome back to Ignite Healthcare Marketing Podcast. We are here with our creative director, Jean, today to discuss, in what will probably be a multi-part series, the topic of AI in creative strategy and creative production. Jean, what’s going on in the landscape? How’s it changing? How are agencies utilizing it? Why are we still paying agencies to do this work? All of the questions that have been coming our way. Today we really want to focus in on the setup, the background, why it’s so important. Then, in future episodes, we plan to share some tests and results with you guys. Jean, thanks for joining.

Jean: Thanks for having me. I’m excited for this one.

Lauren: Jean, I think you’re the best one to frame up, what’s going on in the landscape, rather than me doing it. For the audience listening, give us a little bit of the change that you’ve seen. What are the threats? What are the concerns? Where is everybody at?

Jean: AI has really changed the landscape of what we can do from a creative perspective, but I think one thing we’re really starting to learn as we play with the tools is that it is very much a supplemental tool for creative teams to be able to create in a way that perhaps feels more tailored to the brand. It allows us to do some of the iterative work that we would do almost as homework or brainstorming as part of the creative process ahead of time in a way that gives us more to work from.

I think the thing that I’m most excited about from a creative perspective is, it really arms smaller brands that perhaps don’t have massive budgets to spend on video the ability to create something that feels custom, feels a little bit more authentic, borrows from things that exist within their creative portfolio already, and bring it to life in a way that can be more engaging, more geared towards what platforms are needing right now for performance, and really gives them more firepower, broadly speaking.

Lauren: Jean, you mentioned the AI platforms. What specific ones are you guys thinking about using, testing?

Jean: From a testing perspective, there’s a lot in regards to creative reporting now that enables us to learn faster. From my previous comments, and really where I’m seeing a lot of movement from a production standpoint is within the video and image generation spaces. I don’t believe that can ever replace truly the entirety of the creative process. You need a strong concept, you need it to speak to a human truth. You still need all of those pieces. You need a video editor to take the different pieces that the video generators can make and put them into a compelling narrative that then tells a story that resonates with users.

On that side of things, with the production piece, what we’re seeing and what we’re starting to dabble in use a lot of are some of these AI tools that allows you access to multiple video generators, for instance. It gives you access to more options within AI, and it almost self-selects which tool is going to work best for whatever your immediate needs are. We find as a creative team, that allows us the ability to tap into more. That’s what we’ve been leaning into recently, especially from a video perspective.

Lauren: Awesome. Definitely name the tools as we’re going through this, because that’s what people really want to know.

Jean: We’ve been using a lot of Higgsfield. We were using Freepik as well, which has now rebranded, and I don’t remember the name right off the top of my head. Those are some of the ones that we’ve been using more frequently, especially within the image generation and video generation space. One of the things we’ve seen a lot of success in that I think is a really interesting area for us to play in is all things voiceover related. We can now use ElevenLabs, you can clone voices.

Lauren, if I wanted to clone your voice and have it as a VO overlay over an asset that has you narrating, I could record you now, give it to ElevenLabs, mess with the intonation, and then use that in any assets that we would want to put in to market.

Lauren: It’s a good plan for my maternity leave coverage.

Jean: Exactly.

Lauren: Keep the content coming. I don’t have to be here. The underlying context here that you did a good job of laying out is what’s changing in the landscape. I think the safe, templated approach, specifically in healthcare, is just not working. I think that, plus the intersection of AI, is where it can get a little scary. If you’ve got this safe, templated approach, and then you’re using AI as a multiplier, you’re making the problem worse. What would you say are the foundational elements that are really necessary to make AI work harder for you?

Jean: Acknowledging that AI cannot do it all is going to be critical. I alluded to this before. If you don’t have a strong narrative, if it doesn’t feel authentic to actual people or authentic to the brand, you’re going to miss the mark. You asking AI to generate a clinic space, generate some copy, and put it together without giving it the cues that are specific, of course, to your brand and shaping it accordingly, it’s going to fall flat.

One thing we’ve really learned with AI is it works so well. It is such a helpful supplement if you give it the right prompts, if you shape it the right way, and provide enough to get you to 80%, and then have you as an actual person, then afterwards go in, iterate, update, and make those changes to really bring it to life in a way that will resonate and will feel true to brand.

Lauren: A lot of people think things like maybe the creative briefing process or the underlying brand work can also be done by AI. What’s your point of view on that?

Jean: AI can certainly help you compile a lot of it, but I think the reality is good ideas come from unique spaces. While AI can get that foundational work of here’s the data, here are the human insights, what it can’t necessarily do really well is be creative outside of them.

It can brainstorm within the capabilities of the LLM, but what it can’t do is put together six creatives that have been creating videos and statics within performance creative space for years, and get that brain trust together of here’s an interesting idea leveraging something that I saw on TikTok last week that I think could resonate really well. That’s where AI falls short a little bit, and it does a really good job of giving the starter ideas, but it doesn’t always take it across the finish line. What you also risk there is it may be overly generic. You just have to be careful with how you approach it.

Lauren: Yes, I’ve heard in a lot of discussions lately with creative teams, ours included, that a really good use of the savings of time that you might get on the back end with the production and the iteration, what you normally had to do manually before, now getting to generate sizes, versions for the different platforms, that can be a real time saver. Reinvesting that time back in the front-end work, interviewing providers, spending the real human time to get to know the brand so that those inputs going in, the stronger they are, the stronger the whole workstream is after that point in time.

Jean: One thing we’re seeing, not within our creative strategy and production area, outside of agency altogether, is a shift specifically within Meta towards more creative being critical for broader performance. What I’m speaking to there is really that Andromeda update, that algorithmic shift that was able to be created because their AI has also gotten smart enough to read creative itself.

Along those lines, meeting in the middle, being able to create more concepts by leveraging, for instance, maybe some AI-generated video B-roll that helps you tell a different narrative of a story from a patient standpoint, and then a patient’s caregiver standpoint, and then another concept C. AI allows us to create more in regards to that. You don’t have to do a video shoot for some of those pieces if you didn’t have the budget for it.

Lauren: Hundreds of thousands of dollars-

Jean: Hundreds of thousands of dollars. What I will say is, we’ve done some preliminary testing for AI assets, specifically within video versus video production assets. The reason we wanted to do that was to figure out if the video production costs was critical to drive performance. It’s preliminary, but what we actually saw was that the AI assets performed us as well, if not better, in comparison. I think a lot of that is also how far AI has come in terms of being able to create video that feels real.

Lauren: I did an episode with Brandon recently, and in that episode, I alluded to this one coming up, so I’ll do it in reverse, too. We were talking about specifically Meta and Programmatic, and Andromeda, and how it’s almost reverse now. Andromeda and Meta are looking at creative itself as a signal to who they should be targeting. That never used to be the case. It used to be these two silos of this is who I’m going to target, and then this is the messaging I’m mapping into each of those targets. Now it’s much more fluid and open.

The creative assets are no longer the step two, it is part of setting up who you’re going to algorithmically be able to get in front of. That variation, if you serve a super broad audience, and you want to ensure that multiple segments of those individuals are represented in the creative, you need to have a lot of version so that people see themselves represented. If you are super niche, you may be in a little bit of a better shape because you have one specific audience, but especially for those groups with multiple audiences, the creative iteration is almost like table stakes now. We shouldn’t be launching without it.

Jean: It is. Meta has some of this data already to help validate that point, which is that I believe, across the board, when there were more variants and more things within market, typically speaking, those types of campaigns perform better in comparison. Then, from a creative diversity standpoint, having things that truly felt different, there were a few, I guess, user research pieces or case studies that I’ve seen where it showed the more diverse types of creative you have, meaning you looked at the portfolio, they didn’t all look like the same thing. The more often that happened, the lower CPA was and the better the engagement rate was. They’re starting to, of course, also show data to validate the broader promises of Andromeda, and we’re also running some of those tests ourselves. TBD on the results, but I’m excited about it.

Lauren: That’ll be up for it, too. Jean, I think it’s important to always discuss the element of trust, brand, HIPAA. Have you heard what concerns are circulating, and how are we addressing those?

Jean: Yes. Trust is an interesting one, because I think while we’re at this interesting juncture of AI taking off and being able to do all these crazy things, we’re also at an interesting place where some consumers are repelled by it. We have to balance that. We have to make sure that what we’re telling from a story standpoint doesn’t feel like we’re lying. AI can’t be used to say, here is a testimonial that I made up that isn’t real, and it’s going to stand for this brand. Rather, we have to find the space between the true authentic brand and telling the message to users that will resonate with them, to engage with the brand or product, and essentially tell the story within that space.

Lauren: Making up a case study, and then individuals go to Google, and your reviews are 2.5 stars, and that’s not reflected in the experience, is never going to land. Got to be consistent.

Jean: Yes. There are spaces, though, that we’ve seen where it’s going to play a role that is actually quite helpful. We’ve had some funny conversation. There are times within our space, within healthcare, where people aren’t going to want to stand up and have their story. They don’t want to be the face of a certain story. There are certain things, and of course, with healthcare, that is part of the game. AI can also play a role there. We’re not going to make a fake person to tell this testimonial, but we can assign AI as a face, in a sense, to help take the place of someone who might want to talk about their story but might not necessarily want to be the face of it.

Lauren: I think that leads into the HIPAA component that you were just about to get into. Concerns, thoughts there, how are the brands you’re working with thinking about it?

Jean: I think even HIPAA is interesting because we’re not exposed to PHI. We don’t have that data. It’s easier for us to stay contained within our storytelling space and remove ourselves from it. There are a few guardrails that we just have to make sure we remember. There are certain things that we could and could not say, especially in regards to patient testimonials that we might be trying to illustrate or bring to life, but generally speaking, as far as performance creative itself is concerned, as long as we stay within the guardrails of what we’ve always done, we are typically speaking more than safe.

Lauren: Of course, if you’re representing real patient stories, the brand typically has full approval to represent those, or should if they do not. Jean, I think it’s an overwhelming space right now, and a lot of it is where to begin. If you have tips or tricks for any of our in-house marketers who we know have very lean and oftentimes zero creative support, where should they start? What’s something they could do tomorrow to start this journey of really leaning into creative and the AI landscape?

Jean: Some of the easy wins, I guess I would say, that feel fairly doable, assuming you have someone who is willing to play with the technology, is to, one, you could probably do something interesting with broader concepting. I think LLMs do a good job of giving you a starting point and a few brainstorming ideas. If you’re trying to come up with new taglines or broader throughlines for campaigns. What I typically will find is the LLM can give you 15 good ideas, and two of those are sound for you to then internally riff off of. As I mentioned, it is a good starting point for brainstorming. I think that’s absolutely a space in which you can leverage it.

From a copy perspective as well, I think riffing on copy, especially for iterative channels, for instance, how do I transform something that I made for Meta into PMax? You can have those conversations with LLMs to help you shape that as well. Then the other thing that I think has been a really fun win for us with some of our brands that don’t necessarily have a wealth of, for instance, a video portfolio, a video library for us to tap into, what we’ve found with them is that we can take, for instance, a still photo of a space, a clinic, something related to their actual in-office brand, give that to AI, and have AI bring it to life.

It essentially feels like you’re bringing video within the space without physically having to send a team there. We’ve been able to use a lot of that to stitch into our video to, again, make it feel authentic to brand. It’s not a critical piece of storytelling, but it does feel more polished when you put it together, assuming you have a video editor on the other end that could, of course, put together some video. I know that’s a big if.

Lauren: Yes. A few tips there for everybody listening. We are going to be, as Jean mentioned, not only is Meta in a big push right now to put out case studies and metrics to really validate some of their promises around Andromeda and what the outcomes are going to be, but we’re in some active testing right now with a few clients. We plan to bring you guys some of our own first-party data on how this is all playing out here in the next couple of months. Stay tuned for that.

Jean, I’ll look forward to regrouping with you on that data when we get it.

Jean: Awesome. I’m looking forward to it.

Lauren: Thanks for joining everybody. Wherever you listen to your podcast, please like, share, subscribe so that we can continue to connect with you. We’ll see you all next time.

Announcer: Thanks for listening to this episode of Ignite. Interested in keeping up with the latest trends in healthcare marketing? Subscribe to our podcast and leave a rating and review. For more healthcare marketing tips, visit our blog at cardinaldigitalmarketing.com.