What’s Working Now
A Health System’s Winning GPT Playbook
How one major health system is using GPT to improve access, outcomes, and patient experience—today.
Care Sherpa CEO Jessica Walker and CHRISTUS Health’s Digital UX Manager Jimmy Epperson break down how real healthcare teams are using GPT today to improve patient access, strengthen SEO, and streamline digital workflows.
This 60-minute replay reveals how to:
Make content and UX more readable for both humans and LLMs
Refresh existing pages for stronger AI visibility and patient pathways
Shift from content “volume” to content “value” in an AI-driven landscape
Use GPT to accelerate audits, metadata fixes, and optimization
Avoid the early AI pitfalls costing systems traffic and conversions
Meet the Speakers
🎤 Jimmy Epperson - Manager of Digital User Experience, CHRISTUS Health
Digital experience and SEO leader dedicated to improving access to care through strategy, content, and design that serve the user first.
🎤 Jessica Walker - CEO, Care Sherpa
National speaker and patient conversion strategist helping healthcare organizations increase access, improve retention, and turn patient interest into scheduled care.
Transcript:
JESSICA WALKER: All right, good afternoon, and welcome, everybody. Thank you again to another one of our series related to all things GPT. I'm super excited, this month, to have one of our guests with us. This is Jimmy Epperson from CHRISTUS Health System. Jimmy has been so gracious to come on and share a little bit about his experience in the trenches as someone specifically working within GPT optimization. So Jimmy, if you want to come on screen here. So, Jimmy, your title is Manager of All Things Digital, or Digital Experiences, at CHRISTUS Health. Before we jump in, I'd love to have our attendees hear a little bit more about your role at CHRISTUS and the work that you've been doing, and why this topic was near and dear to your heart.
JIMMY EPPERSON: Yeah, well, first of all, thanks for having me, I really appreciate it. This is a passionate topic of mine. It's… so, I manage a team of content writers and a team that manages the website. We do a lot of digital content from email to blogs, service line pages, location pages, just, you name it, we're writing it. And one of the big things that I came to CHRISTUS to do was to increase non-branded traffic to the site, and I did a great job. Working on top of funnel stuff, and we really knocked it out of the park, and then, AI overviews came along, and so did the threat of these LLMs and chatbots. And so, naturally, I'm very interested in this, what do we need to do? What's the next phase to keep making CHRISTUS Health relevant and in the search results?
JESSICA WALKER: I love it, and when we connected about you being a part of our series, the part that I'm most excited that you're going to share with our attendees today is you're in the trenches, you're doing this every day, and you've learned a lot along the way, so I'll let you get started, and then I'll stop sharing so you can share, and ultimately I think our attendees today are going to love hearing from you around the things that have worked and what you've learned along the way, so I appreciate it.
JIMMY EPPERSON: Great, let me share my screen, and we'll get started here. And I’ve gotta say, programs like this, and the overall SEO community, they really do like to share and help each other out. I think that’s great.
Okay, so let’s jump in.
This session is called the Health System’s Winning GPT Playbook, but what we're really talking about is how AI search is reshaping SEO. Briefly, we're going to touch on the modern marketing team. There’s a big debate happening right now in the SEO community: Is AI search just SEO with a new interface? Can we follow the same SEO principles and expect to show up in LLMs? Or is this an entirely new game?
I would argue the chessboard has completely changed. We’re playing a different game. AI search doesn’t always follow Google’s rules, and because of that, we have to think differently not just about our content, but about everything in our business if we want to show up in LLMs.
I’m going to go through three or four slides and talk about what we’re actually finding, and then I’ll show you some examples of what we’re doing at CHRISTUS Health to make our content more friendly for AI Overviews, LLMs, and chatbots as they continue to evolve.
To start, and this is just my opinion, this is a new channel. This isn’t SEO. It’s going to require different approaches and different ways of thinking. It’s not as simple as “follow SEO best practices and you'll show up.”
There’s a lot of debate out there. iPullRank is a big SEO firm and they constantly post about this on LinkedIn, and the comments get very passionate. So… what do we know so far?
From GoFish Digital, we know that LLMs don’t read full pages. They only read about the first 90 lines. Right off the bat, this tells us we need to ensure the right information appears above the fold.
We’re also seeing that LLMs seem to strip out everything except text. They ignore layout, design, icons; all the UX elements we obsess over. What survives is plain text and structure. This means clarity sometimes beats design.
Another big shift: in many cases LLMs rewrite the user’s query, trying to guess what people are really asking. So this is already totally different.
What we're also seeing is huge: only about 20% of Google results appear in LLM answers. Four out of five high-ranking pages never make it into an LLM result. Traditional ranking factors are no longer the deciding factor.
LLMs pick content based on clarity, structure, consistency, and preserved authority. That means brand is becoming even more important. And ranking well on Google does not guarantee visibility in AI search.
From SEER Interactive, we’re also seeing that content recency matters. Refreshing content, updating key sections, tightening structure, all of that can make a difference. In healthcare, this exposes a big challenge: many older pages that haven’t been updated in years are unlikely to be surfaced by an LLM.
Another interesting shift: the domains AI cites are changing. Models are drifting away from the trust signals they leaned on six months ago; Reddit, Wikipedia, etc. What worked six months ago likely won’t work six months from now. We have to stay adaptable and plugged in.
This is why I'm glued to LinkedIn and following the SEO experts who study this stuff full-time. I’m in the trenches launching pages, I rely heavily on the researchers.
So what does all of this mean?
It means LLM visibility does not equal keyword ranking. Those are not synonymous anymore. Your content probably needs to be recent, not always, but recency does help. Strong signals must appear at the top of your pages. I’ll show real examples of what that looks like in a minute.
Brand consistency across all platforms is going to be more important than ever.
And we must optimize how AI interprets content, not just how humans read it. A beautiful UX design might still hide content behind scripts, images, or structures that AI ignores entirely. We have to rethink how we build.
Now, I’m going to go over three tests we ran and share the results. These are simple changes, nothing overly scientific, but they produced meaningful outcomes. Interestingly, everything we did made the content more user-friendly for humans, and we found the LLMs responded to that too.
With SEER Interactive, we identified four articles and several location pages for a content recency study. We rewrote the content, updated the data, and changed the structure so the most important information appeared at the top.
We added an “At a Glance” box. We added actionable buttons. We summarized the full article so AI could quickly scan and determine its helpfulness.
The results were solid. One blog saw a 300% increase in AI traffic , though to be transparent, that was a jump from 1 click to 4. But where it got interesting was our ER pages. After applying the same approach, those pages saw 63% growth in AI traffic, while the control group declined 31%.
This made us rethink our strategy. Instead of always creating new content , “Let’s write another cardiology story!” , we’re now asking, “What do we already have that we can update and make AI-friendly?”
Another shift: we’re creating friendlier, more engaging content. I was working on a spine/back pain article for paid advertising, and it felt like the shackles of traditional SEO were removed. Instead of writing a classic blog, I created something interactive , knowledge checks, FAQs, clickable sections, mini-quizzes.
We don’t yet know how it performs in LLMs, but the AI Overview already cites it. It’s also ranking #6 on Google. That tells me we may need to stop writing “for Google” and start writing “for humans,” trusting that LLMs will catch up.
This approach made content creation fun again, and our internal stakeholders loved it.
Finally, one big change we made was rethinking top-of-funnel content. At CHRISTUS, we had a massive amount of top-of-funnel content for breast cancer. It performed well in email and search campaigns. But with AI Overviews handling all the surface-level questions, top of funnel is shrinking fast.
So we moved that content under the main breast cancer service line page, making it a supporting, centralized resource that reinforces why CHRISTUS is the right choice. It strengthens authority and answers questions in one place.
We haven’t launched it yet, but honestly, we should’ve been doing this all along. We just didn’t need to , because we were playing the Google game.
So those are three strong examples. I’m going to pause there in case there are any questions about these tests or anything else I shared.
JESSICA WALKER: Jimmy, thank you so much. Just to remind our participants that we do have the Q&A that you're welcome to send in your questions live, for Jimmy. And again, he's… he's been doing this for some time with some testing. We're going to talk a little bit about how he was able to get internal buy-in and executive approval to do some of this, and what are the things he's testing all on his own, but by all means, he's ready to share any of his insights he's learned along the way. While we're waiting, Jimmy, I do have one question you just brought up. It was a really good point, which is, you mentioned about going beyond Google, right? So not just focusing on the Google trap of that being the only channel, the only engine we've all been paying attention to, right? And so… as you've been doing these strategies, thinking beyond Google, what have been some of the other indicators or signals that you've seen that are making a difference across other LLMs or Bing or other search tools?
JIMMY EPPERSON: Other signals that, we're doing this right, or…
JESSICA WALKER: Yeah, that's making a difference
JIMMY EPPERSON: Yeah, I think it's so early, it's hard to tell. a big thing that I thought about was, what is our playbook for LLM? Well, we're building that, as we go, because… and so is everybody else. Everybody else is trying to figure it out. In… in figuring it out, we're… we're doing a lot of tests and… and, So, I… I don't know if we really have anything else. We have some other things we're testing. Just, we're putting a reason to choose CHRISTUS, and really getting detailed on each location page, and trying to figure out, okay, what… What is the true, unique selling proposition here? And how can we clearly say this at the top of the page? It's obviously there, but how can we maybe package this differently? There's also the schema markup that we're trying. I've seen some benefit to that. I was testing a page in Longview, Texas for one of our hospitals that had a stroke care program, and I put schema markup on there. And it immediately showed up in the AI overview. It was very interesting. And so we probably need to be doing more of that, too.
JESSICA WALKER: Just a quick plug, our presenter next month is all about schema markup, and there's an opportunity, so preview, if you're on those, come back next month, it's gonna be a great topic, because you hit it nails on the chalk there, Jimmy, that we've seen that as the most, successful, immediate return. If you spend that… it's painful, right? But there's partners out there that can help you with it, but it truly shows a return on investment. Sorry to interrupt you on that.
JIMMY EPPERSON: No, that's great, and I will probably watch that one. I'm not an expert there, but with the chat box, you can basically make anything. So we're able to put some on some specific pages and test them, and so we have seen pretty… but that's just an AI overview, We don't have any data if that's showing up in LLMs.
JESSICA WALKER: And we did get a question while we've been chatting, so as you incorporate your blog content into your service line pillars, will you keep the blog content, or are you going to sunset your own blog internally? What are your thoughts there?
JIMMY EPPERSON: Yeah, we're gonna forward the former URL to the new service line page, because otherwise we'd have duplicate content, and there's really no point in doing that, so… Yeah, our intention is to not sunset the whole entire blog, because we do have some great, stories that are in pages that are ranking well and get a lot of traffic. But whenever you look at, the real data, analytics, and how much traffic those blogs were bringing in, it wasn't that much, and so I think it's okay in that case. But we have a… we have a page that's, pain after knee surgery, and it's still doing very well. We've updated it quite a bit. That's another thing we did. We interviewed doctors and started putting doctor quotes on there, and really started to answer a lot more questions on that one. And we're never gonna get rid of that one, as long as it keeps generating traffic.
JESSICA WALKER: Yeah, using the data to guide you, right? So, outside of the LLMs, but seeing that the traffic, the analytics show you that's working, and then now thinking about enhancements, that's fantastic. Well, we'll save the next couple questions for the next segment, but I'll go ahead and go off and keep going.
JIMMY EPPERSON: Okay, and I only have two more slides. The other thing that is happening, and the way we are working differently, is something I want to explain clearly. At CHRISTUS Health, we do not use ChatGPT; it is too risky. There have been leaks in the past, including searches appearing in the search console, and we cannot afford to be in that situation again. Because of that, we internally use Copilot, which works very well for us.
Now, for my own workflow, I use my personal computer on a separate network. I do not access anything tied to the CHRISTUS Health network. That allows me to build automations, and these automations are incredibly helpful; they speed up work significantly. The first thing I will say is this, do not go against any internal policy you have. I am only scraping publicly available information, and since all of it is external, I believe it is safe; they may tell me it is not, but for now that is how I work.
One example is an automation built in N8N, which is similar to Zapier or Make. I can drop a URL into Airtable, which functions like an Excel or Google Sheet, and the automation will crawl the page. It returns the H1, the H2, the meta title, the meta description, and the OG title and description. Then it sends all of that to ChatGPT and asks whether the metadata follows the parameters I have defined. I paste in the rules, and it evaluates everything for me.
This has accomplished two major things. We have more than 500 location pages; when we launched our site four years ago, our brand was in an earlier stage. Over time, our voice evolved, which means some of our H1s and H2s no longer match, and much of our metadata is inconsistent. Running this automation allowed me to gather all 500 pages at once. You can do something similar in Screaming Frog, but the difference is that I can then send the results to the LLM, and it will generate suggestions based on our brand rules. It flags what is not aligned and offers fixes. It has saved hours and hours of work; honestly, it would have taken us weeks to audit 540 pages manually.
Because of this, we are getting everything aligned with our brand faster, and our audits take a fraction of the time. Marketers often think only about how the landscape is changing and how to get our brands into these new channels; but there is another side of the coin, which is how we can work faster and more efficiently. These tools make that possible.
Another thing I have been able to do is use Google Gemini’s deep research. When we have a high-profile provider who is starting or updating their profile, their resume or CV alone is not enough to tell the full story of their work. With deep research, we uncover things like contributions to rural health initiatives, involvement in clinical trials, community work, publications, and other accomplishments. We add this to their profile, which improves SEO and eventually helps them show up in LLMs, because it strengthens the authority and credibility that LLMs look for.
For anyone who has not tried this yet, I highly recommend it. That brand audit alone allowed us to work five times faster, and it improved the quality of our content at the same time. It makes the team quicker, more efficient, and ultimately helps us deliver stronger content. And that is all I have in that section.
JESSICA WALKER: Jimmy, I love that, and a question, and we talk about toolkit, but I might hit you up afterwards to see if you want to give me your prompt that you're using here that we can share with our listeners. It's wonderful what a great use, like you said, going beyond… not only is it serving you to get your brand out there, but your efficiency and your time, and it's all public available information, right? So it's nothing that you are doing behind the curtain that you're, organizations' policies don't allow, and that's a tremendous example. I know a couple webinars ago, we had our friend Andy from RatingsMD on and talked about… and so what I love with what you just shared with the deep research on your provider, linking that strategy with increasing reviews and getting their their pages claimed and linked, and then also giving them credibility through those backlinks. So is that a big part of when you updated their profile, of adding in where else they've shown, or how did you use additional information that you found during that deep research?
JIMMY EPPERSON: Yeah, we didn't do backlinks or anything like that, it's more just, I can probably send you all an example if I search for it, but it's more… Showcasing the provider's authority and credibility, and why they’re good at what they do. That's more. A lot of times, we don't have that information, and we just have, where did they work? We don't have all the extra education and the big things they really accomplished.
JESSICA WALKER: And I'll say from a consumer mindset, so as many of CareSherpa, we think about that journey from someone who is interested as a consumer before they become a patient. And don't sleep on that strategy of bringing in their care philosophy. If they've done any YouTube videos or things where they're expressed, or patients talking about them, those really do make a difference, and linking that back, so how is that showing up when you write their profile, right? So, about their personality, their care philosophy, their approach, patients as consumers look to that information and help it to decide, right? provider. Another question, we alluded to earlier I'd love to touch on right now with you is, this has been a journey. You've started this your own, independently, out of curiosity, seeing the trends and whatnot. So, could you share a little bit about how it's been to get internal executive support for the work that you're doing? Anything that you had to do in particular to bring facts or data or insights to the table that gave them a green light to keep going on the work?
JIMMY EPPERSON: Not yet. I have the trust of my leadership. I think there's still more to come.
JESSICA WALKER: Well, so they were supportive, though, of you starting that, too, so that was… you already… they already had support for working on optimization with AI.
JIMMY EPPERSON: Oh, yeah, absolutely. We have a team here that is working to figure out how we can automate a lot of stuff through AI, on the digital team. We have a new team member who does that. And then also, I think the thing that I would have to get approval on is, am I allowed to use ChatGPT in this instance? And it may not be, I may need to change it to Copilot. And they are starting to really look at it because they know that there is a threat in doing this, so… Yeah, and so I, I definitely have the support of my boss and my boss's boss, and, so… I think it's just, what tool, what LLM tool should I be using, and can I use it in this way? And if I can, that's great. If not, then… We'll have to go back to the drawing board.
JESSICA WALKER: Well, and so, for some of our attendees, I want to double-click on the story you shared a little bit ago with this example on the screen. you demonstrated that that saved you hours and hours and weeks and weeks. And so when we think about building internal support, even in that, you use publicly available information, right? You use the tool in a non-threatening way, but you could show, here's how quickly we were able to get this accomplished using these tools, and quantifying that with labor savings, and you were able to accomplish that quickly and move on to more, additional projects. And those are the small things when you're thinking about building credibility with your executive team. And help them see that there's… within the guardrails, there's things that we can do that are not scary, right, with using GPT optimization.
JIMMY EPPERSON: Yeah, and I think some of this probably… we have a group, a board that really looks at each instance, and some of this has to be taken up there to see if we're allowed to keep working this way. Yeah. And so… and that's actually a new thing for CHRISTUS, and… I think the story is that, CHRISTUS is really engaged in this. They want to do this the right way. They… they don't want there to be any threats to any… any type of patient data, or any type of anything like that. And so they're very careful, but they also… do see the true value in it, and they want to do the right thing. So we're taking small steps to get there.
JESSICA WALKER: Yeah.
JIMMY EPPERSON: And, this is honestly just an example of what you can do. I couldn't imagine what we could do if we had everything locked down in a personal GPT and just had all the data at our fingertips. that would be amazing.
JESSICA WALKER: Well, and I know that the work that you've, like I said, you started this on your own independently, and you're just in that learner's mindset of it. So, what are things as you look ahead to the end of the year and in 2026, what are some of those areas that you plan to continue to expand or look into? You talked about content optimization. What are some of the other ways that you're planning to deploy GPT optimization in your day-to-day?
JIMMY EPPERSON: In my day-to-day it's everything, so some of it's super simple, I have a paid account, and I usually talk to it on the way to work and say, hey, these are my to-dos, and my drive's about 30 to 40 minutes. And by the time I'm at the office, I know exactly what I need to do. Sometimes I'm… I already have work finished, because I've talked it out. And it's just a matter of just getting in a Word document and publishing. And then… some other things I'm doing, too, is, I am looking more at this automation to try to figure out how can we shorten this workload? we're… we're a scrappy team, and we… and that's how marketing teams are, and so what… what is it that… we really need to, figure what is… what is blocking us from getting the best work done we can. And so, when I started using ChatGPT and the LLMs, I thought it was really just for writing, and now I'm seeing it's for so much more, and you can just get so much work done, sometimes while you're sleeping. I created another automation that will look at AI overviews, and it looks at it every week for a certain set of keywords, just so we can figure out, is anything changing? Are we still being cited? Are we not being cited? I feel like with this stuff, if you do it right, the sky is really the limit, and there is just so much you can do.
JESSICA WALKER: Yeah, one thing I find really interesting is that, I had another partner that was going through and actually using it to listen to what's happening, like social media listening and global listening, and then giving their content team suggestions on content, right, in real time. So every morning, they'd come in, be hey, here's a starter content to consider, or blog post, or social media post based on it. And that was, like you said, working overnight, and just publicly available information, and serving it up in a way to say, here's a way to keep your content fresh and out there and relevant to whatever events may be happening. We… Oh, go ahead.
JIMMY EPPERSON: Oh, another one I made was, so… If we're gonna write about a keyword, you can put it in a form and put the city in, and all it does is, search, what comes up, and it finds the related searches, it finds, the competitors, and it comes back with a pretty big analysis within, 15 minutes on, hey, this is where your competitors' content is really good, this is where… you're really bad, this is where we think you should probably go. And then we also have another one with locations, too, where you can plug it in, it'll go on the… internet, and find your competitors and figure out where your strengths are.
JESSICA WALKER: Yes, that is amazing. And I will say, for anyone listening, where if you've not played with any of the LLMs in this category. As I coach my team, as we talk about it, which is, you don't need to know how to be an expert to use them. You literally can go and say, I would like to build a process where I'm checking my competitor's content, and walk me through, step by step, how do I do this? And literally, the GPTs, the LLMs will walk you through all the way, and I've written code. I've never written code before, but now I know that. I tell a lot of my friends, just get in and play. So even looking at what Jimmy has here on the screen, just a simple ask it to walk you through, how do I accomplish this goal using a tool such as this?
JIMMY EPPERSON: Yeah, it'll help you, and I think the first one I made was so complicated. I should show you, it has… I think there's 30 nodes here, and it cost $2 per API call to ChatGPT, it was ridiculous, but… All in the learning process, so…
JESSICA WALKER: Exactly. Jimmy, we had a question that came in to say, could you talk to me more about why you chose to do prototyping with ChatGPT instead of Copilot, knowing that it may be a hard sell internally?
JIMMY EPPERSON: Well, so, I think, we're transitioning. We transitioned recently to Copilot, and… I think that's really why. I'm still, I'm getting associated with the co-pilot now. All my examples are in ChatGPT. That's the only reason. For this automation, I don't think, you can connect, Copilot to it, so… That's really the only reason. We were transitioning from having no LLM for our services, for our marketers, to adopting Copilot, and that's the only reason.
JESSICA WALKER: We had another question that came in. Earlier in your presentation, you shared the discovery you had about, the landing page and the first 90 lines and information. Could you share a little bit more about, once you had that discovery, what were some of the now, how have you changed your practice or protocol to, use and leverage that?
JIMMY EPPERSON: Yeah, so think of a, like a newspaper, and think about above the fold. At a traditional newspaper.
If you're old enough to remember printed papers, it was folded. And the big headlines were up top. I, I think, I think, I think that way now. I try to think, okay, this is above the fold. Do we really need this gigantic hero for this article? That is… only has two buttons and two headlines on it? Or can we do something a little bit different, where maybe there's a smaller picture, and we can put all the necessary information, not just for the LLM, but for the patient and consumer making a choice? I really think, that… that is my leading, that's how I … determine that, I don't know that there's a fast rule on, hey, you should have all of this in the beginning. I wish there was. I think my team wants that from me, but, because they keep asking for clarification, and I'm sometimes you gotta live in the gray area and see what works.
JESSICA WALKER: Yeah.
JIMMY EPPERSON: But, for example, that cardiology one, we had, hey, here's what this story is about at a glance, and just had a very nice summary. We had two buttons that said, this is where you can find a cardiologist, this is where you can find a location, and then we had, basically, a robust summary under the picture. And, we didn't have that before. We had a headline, and we jumped into the article.
JESSICA WALKER: And it's like you're building for two audiences, right? You're trying to optimize for human consumption, and then optimize for machine consumption. And I know some of my other client partners working with their marketing content partners. It's experimentation. Like you said, you saw with your one ministry page that it made a significant difference pretty quickly. I have a client partner right now that's testing two formats, where at the beginning of every blog article is a TLDR, for those of you who don't know, too long, didn't read, right? And it gives a very snippet summary, because again and then right below that, they're creating the question-answer format, right? So what's a common question that someone… to your example, how long is recovery after knee surgery? So they make sure that all their content leads the page with those two things. And so, I don't have anything to report back yet in terms of it. We saw early indication, though, that it was making a difference with how they were showing up. And… we talked a minute ago, Jimmy, about your competitive analysis, looking at your competitors. If you want to be the one ranking the highest for knee replacement in Irvine, Texas, well. if you're not in those top 3 results in the LLM, why are your competitors there? What are they doing differently, and what do you learn, and then how do you outsmart them? Any other things that you've discovered, you just talked about with your team, that you guys are doing currently with experiments, or anything else you're trying to see what works?
JIMMY EPPERSON: Let me think we have a lot… of things we are trying. I think I probably hit the… top ones? I really think the lesson of, unshackling yourself from the Google rules is probably the biggest aha moment I had. I… because I've been doing SEO work for 10 years, and… I started by working with this guy who had a checklist, and I was just a writer, and he said, follow these rules, and pretty soon I was… I was ranking, and… and so I have… I have been that way for so long, just, following that checklist. you… there… I think there is no checklist, there's no playbook, and…
JESSICA WALKER: Yeah.
JIMMY EPPERSON: Everything has to be tested, and… everything has to be… we have to test everything in a way where we can see the results, and Sear Interactive, our partner, is really good at that, testing and making sure that we're doing the right things and keeping track of the results.
JESSICA WALKER: Yeah, I love that. The reason CareSherpa is hosting this series is that we have seen consumers specifically referencing GPT of how they found their providers, or how they led to the questions they asked us on the phone when we were interacting. And so, my plug that I've been doing with my client partners is pay attention to what's happening in that last mile, right? When that patient goes from a prospective consumer to a patient, what were those questions they asked? What were the deciding criteria? What were their objections? That is your leading indicator of our content, is our information that's out there guiding that more quickly? Because patients are asking the human because they couldn't find it digitally, right? And so, for us, that's been a big piece that we're experimenting with a lot of our clients is, especially in those high-value, high-margin service lines, let's dig in and really understand the consumer mindset, and then therefore, how are we then feeding the machine to then find us more easily?
JIMMY EPPERSON: What kinds of things have you seen, for that last mile, where people have seen the ChatGPT?
JESSICA WALKER: Yeah. Well, and I share with you the story, and I talked about it on a prior webinar, you use the example of a woman who had a very specific medical condition, and because our provider had content in multiple places on their page about this condition she's in the Bronx, New York, and he's in St. Louis, Missouri, she found him, and got on a plane, is now seeking care with him. And so, it was that that came up, where it's specific about conditions, or specific you're a great example. I want to know the common patient. I want to know what's the procedure, how is it going to free, how do I have to prepare? What's it like after, what's my recovery time? The other interesting thing is that we see of… it depends on how people decide to buy, right? But those who are in emotional buy, the questions they ask us about the provider has less to do with their credentials and more about, hey, what's their expertise? How many of these cases have they done? how long have they been practicing in this specific treatment procedure? And so those little nuances where we go back and adjust the content to say. Dr. Jones has done 300 of these XYZ procedures in the last 12 months. Credibility gets popped, right? Another example is we hear questions around… it's silly sometimes, but it's, everything environmental as well. So, meaning, how easy is it to get to the location, or the parking? These are the hesitancy moments of the… can my… can my child come with me to the appointment? Can my spouse come with me? Or what do I need to know? And those, again, very service line specific, very treatment-specific. So for our clients, we track what are those frequently asked questions, but I would… I would welcome all of you who are listening, if you're using, a CallRail, Invoca, whatever, recording calls, there's wonderful AI tools, even within those partners, that they can give you themes, right? And they can give you… and that's the goldmine of trying to say, here's an opportunity for us to continue to optimize, because if it comes up with a frequency with every call. There's an opportunity to leverage it.
JIMMY EPPERSON: That's… that's really good. Yeah, we, listened to some calls on… for… through Invoca to figure out why people were calling hospitals, and it definitely changed how we… presented content. And… one thing that I… this reminds me of is, I used to think that 2 months ago, just from listening to everybody, that, our content had to change drastically, and that we needed to have different content from everybody else. And I think that's another thing that I'm probably working on, too, is that we don't. Actually, there's research that says that LLMs want to see content consistent with other things they're sourcing for. So, WebMD, they want to… they trust that, so they expect content to be that way. And so, really, the trick is not to make something completely different, but to make it basically have the same information, but make sure that you're really showing something else that has your value proposition, or why you do things differently on top of just having the traditional content. And so your questions, like if we had that data, we would definitely… Just put it in what we already have.
JESSICA WALKER: Exactly.
JIMMY EPPERSON: I changed up a lot.
JESSICA WALKER: And I share with my clients what we look at when we study this consumer last mile mindset is, let's, for sure, look at the people who actually converted, right? What did we hear during their calls? What were their questions? What did we share during that call that made a difference that they felt comfortable to schedule? But then we also study those who called in and were clearly an opportunity, but did not move forward. So what was that determining factor? Or. Was there something that the agent didn't share with them, whatever the case may be, but that's that whole, focus group, and especially those of you that are recording your calls today. It's a great… even if you're just doing a random sampling, you learn so much. And another thought, too, with what you said, and we think about, this, where how do you make sure that not only are we responding to the consumer mindset and the questions, but then the credibility, right? You mentioned of… LLMs want to be able to verify it's changing every day. We said on our webinar a couple months ago, all about Reddit and other external credibility. Well, now that's not the case anymore, right? And so, as frustrating as you said, we all want the playbook, we all want the checklist, we just don't know what it is today, because it is changing.
JIMMY EPPERSON: Yeah.
JESSICA WALKER: And so, I say, if nothing else, all the marketers on this call, as you all are, you're data-driven and experimental, and that's what you have to do, right? Just try and see what works, but you also can't sit on your laurels, because it's going to change, possibly, too.
JIMMY EPPERSON: Yeah. Yeah, and my… our leader of strategic marketing, Jeff Stewart, whenever I first started, yeah, I was… this was 4 years ago, and I was always let's just get on page 1, and he always would stop and say, hey. you need to make sure that you're answering the questions for the consumers. You need to make sure you're answering questions for patients, and you need to make sure that, it's written for them. Everything else will follow if you do that, and I I really … uphold that, and I have. And I think that's true. if you see the, that cardiology story I wrote, and that we changed. We really didn't do anything drastic except make it easier for the person reading the story to find information faster.
JESSICA WALKER: And I love that from a good summary. I think, like you said, at the end of the day, we want to learn the magic checklist, but we're not. It's a black box in a lot of ways, so what are the right practices? And making it easy for patients to do business with you, making it easy for them to understand you, and not make it a complex concept, and like you said, writing for humans. We have seen that the LLMs respond really well to that. So, the basics of blocking and tackling are still there, and now just do it at scale, so…
JIMMY EPPERSON: Yeah, and I'm curious on what your opinion is. Do you think AI search is… which camp do you go into? Do you think it's SEO, or do you think it's a completely new channel and different?
JESSICA WALKER: I'm maybe going to add this as my ongoing question as every session we end with, too, Jimmy, but, and just, I wanted to put this up here for our attendees. So, Jimmy and I are both on LinkedIn. We'd love to connect with you and continue the conversation. I did just get a request, Jimmy, around your cardiology, if you'd be willing to share that information with our folks around that tool. So we'll follow up with you and send that out to everyone in our follow-up with the webinar today.
But to answer your question, I believe that the pie we talked about who moved your cheese. This is the same cheese, the plate has just gotten bigger. This is another channel. We still have to think about SEO, we still have to think about all the great tools that we know, but now, and in some ways, I appreciate it, because it's no longer just a Google game that it has been for the last 15 years, right? And it's going to continue to evolve, and consumers are evolving with it. So, for me, it is now another slice of the pie that we just have to be mindful of.
JIMMY EPPERSON: Yeah, absolutely. I totally agree. I do think it's its own channel, and I think more… Obviously more is going to be revealed, and as ChatGPT and Perplexity have their own browsers, and they become safer, and more people use them, I think we're going to see a big difference in just how people are finding information.
JESSICA WALKER: Jimmy, I really thank you for taking the time to share with everyone today about your expertise and the way that you've experimented and learned, and we just got a great comment in the chat around, this is really it, right? So try things and go to leadership. Just learn, be curious, we're not going to have all the answers, we're in a brave new world, and so coming on sessions like this, like you said, seeing what everyone else is talking about in LinkedIn, and that's why we've hosted this webinar, just to get more content out there and have all of us share learning and see what we get from there. But, this is a QR code to our toolkit, so if you've got tools from Jimmy and others from our previous webinars, we encourage you to go and download this. But, Jimmy, again, thank you so much for spending your afternoon with us and our attendees today.
JIMMY EPPERSON: Yeah, and I would just say, be curious and be careful. That's what I would say, and then go to leadership. Because they want us to get work done faster. They see the value in it. It's just got to be safe and right, and it's a balance between those two things.
JESSICA WALKER: Absolutely. Well, thank you, everyone, and we look forward to seeing you again next month. We're going to be talking all things Schema Markup and what that looks for you to practically do it, bringing some resources to the table that can help you, because you all dip your toe in the water if you've not been there, done that yet. And then, as well let us know, reply in the email, let us know if there are other topics or presenters you'd love to hear from. But Jimmy, you've been a pleasure. Thank you again so much.
JIMMY EPPERSON: Thanks so much, I appreciate it.
JESSICA WALKER: Alright. Bye, everyone.