AI is Fire (Part 1): Getting Started With AI

Andrea (00:04):
Well, hi there. I am Andrea Watkins, and if you're listening to this while juggling three patient calls, checking in a couple patients, taking a payment, selling skincare, and trying to catch your doctor in between procedures, you might be working in an aesthetic practice.

Blake (00:18):
And I'm Blake Lucas, and this is PracticeLand. This is not your doctor's podcast.

Andrea (00:26):
All right. Well, welcome back to PracticeLand. We are here to share a lot with you guys today about something that is very, very important in the world and in aesthetics. We're talking about AI, the staff, how it affects our staff, our systems, our teams, and how it affects scaling for any business. So as you guys know, we're always talking about what's going on in aesthetics. And really from the perspective of those of us that work in the practice, we're the doers, we're the people that get shit done and go through our days sometime chaotically, sometimes a little bit calmly. But a lot of things we've been hearing lately are about AI. What is this AI giant? Should we be scared? Is this going to help our practices? And so we have an expert in the arena. Two of them, as a matter of fact. First of all, I'd like to introduce Ken Bosan.

(01:20):
He is the chief strategy officer at Studio three Marketing. He's going to come at this from a data visibility and technical perspective. And then we have Eva, who's the founder of the Axis Podcast production company. She has been in aesthetics for over what? Two and a half decades.

Eva (01:38):
Two and a half years.

Andrea (01:40):
Which is crazy because she's 24. So a lot of experience and a little amount of time, but she really brings an expertise in translating brand knowledge and really will be able to provide from an authority driven content perspective what AI is really doing and how it's showing up in the aesthetics business. Our goal here is to provide you guys with some clear and grounded guidance for what happens daily in the practice. So with that said, with that introduction, Ken, let's start with you. You work with hundreds of practices across the country and AI is a big thing.

(02:18):
I know you work directly with Google, you train teams, you speak at conferences in the aesthetic space and other spaces about AI specifically. So what would you say is the overall consensus or the feeling that practices have right now about AI?

Ken (02:34):
I would say the overall consensus or feeling is a feeling of unknown. And there's anxiety. They want to know what can they do. It's that feeling of they don't want to be left behind. They've traditionally been told, "Focus on Google, focus on Google, focus on Google." And now the world is shifting where they're getting bombarded by emails and influencers on social media every day saying, "Focus on AI, focus on AI, focus on AI." And it's formulated a bit of a stir craziness around it, I would say.

Andrea (03:14):
For sure.

Ken (03:15):
But I think the simplicity is there are some stable data to adopt and ensure that your brand has a position. And we have some strategies in place we could talk about and you will not be left behind. And if you take advantage of the right things to do, then I think you can only capitalize on it and it'll help you and it'll help your brand help your position.

Andrea (03:42):
Eva, anything to add there to what Ken said about how you think people are just feeling about the thought of AI in aesthetic practices?

Eva (03:52):
Unknown was a very good way to describe it. I agree with Ken. I, myself, I think have been on the more skeptical end and the less of the early adopter end, which is not normal for me. Usually I'm the first person to jump in and figure something out. But because my early experiences using it were so poor, I just sort of put it on the back burner and said, "I'll do it later. I'll do it later." And then all of a sudden I was like, "Oh, I'm a little behind." And so fear is part of it for sure. And doctors are driving a lot of the anxiety in a good way. When things change this much, this fast, we get a lot of worry coming from them. And then for those of us who are relied on for good advice, I think we have to kind of jump and get in front of it so that we're not sending them in the wrong direction or even just being inaccurate, I think is important if we get those questions. So I learned a lot just talking to Ken last week about this episode and getting ready for it.

(05:04):
And I'm really excited that he's here to share some of that with us because I think I represent the practice well in the sense that I'm in a vacuum. I don't have data across dozens or hundreds of accounts to look at. I have two. So I only know what's happening in two. And until I call one of you guys and say, "I see this, do you see this? " Then I don't really know what's going on. So yeah, I'm grateful that I have you guys to talk to about it.

Andrea (05:35):
Yeah. I was very grateful when Ken finally conceded and said that yes, he could find some time. Not to, as my son would put it, glaze you by any means, but to have someone in the industry that understands not just the technical side, but how it's really impacting aesthetic practices is so important for our specific listeners. So to that point, Ken, before we get into very specific questions, how would you describe what AI really is and how aesthetic practices should be feeling about it at this time?

Ken (06:11):
Well, that's a great question. I would say if you look at, if you just take the simple analogy of the harnessing of fire, it was such an important tool in lifting humanity above every other animal kingdom race that existed at the time and really making us the dominant species on the planet and all the inventions that have come from that. It's not just for cooking and heating, but everything from boiling water and being able to remove bacteria to preserve longgevity to powering engines to creating the industrial revolution to thermal nuclear physics. I mean, it goes on and on and on. It has billions upon billions of uses and that's kind of how AI is, is you have to look at it. It's like it's a tool that if harnessed correctly, there's billions upon billions of uses in anything from search and user behavior there, but all the way down to systematizing analytics, systematizing accounting, billing, driverless cars. I mean, just the uses goes on and on and

Andrea (07:31):
So I think to get this kind of grounded in something practical, you had kind of said using how are we using the tool? So let's just kind of do a little round robin here. Eva, I'll ask you a couple questions, kind of ask you a couple questions as far as AI specifically with the tools that we have available to ourselves now, let's start with kind of a do list as far as with the tools available. Eva, we'll start with you for content and branding. So we've got AI tools available as apparently we're not supposed to use them just to make memes and be funny, but to get something useful out of the utilization of this cool tool, what are a couple things that you would tell people, keep in mind, definitely do use the tool for this aspect of the content and branding for your practice.

Eva (08:30):
I love asking questions like this because everyone I talk to uses it a different way and for a different purpose. And so when you're starting out, you kind of have to get over yourself and just start using it and see what it does. And you have to also remember garbage in, garbage out. So if you know that, give it a head start and help it help you. The first place that I started using it, of course, was, actually it was an emergency that I was in and I had to go to a conference and I had to write something like 14 interviews for the next day or two days from now. And normally it takes me one to two hours to write interviews of people that I don't know because I have to look them up and read about them and think about who they are and what do they know and what would I want to ask them? And there's a lot of empathy in there. And in this particular case, I didn't get the list of the people until two days before the conference.

Andrea (09:33):
Rude. Who did that to you? Don't tell.

Eva (09:36):
Not going to tell. However, be out of necessity, I was forced to try using AI to do this. And so instead of just saying, "Here's the name of a person, help me write questions." I actually took their podcast and I put it in as the source so AI knew what the show was about and what we had done in the past. And then I said, "Help me write questions for this person to be on this show." And you guys, I am not kidding when I say this thing, this project that would've taken me two weeks to do, I had it done in two hours. And the only reason I tried it was because I had zero time. I probably wouldn't have done that if they hadn't been that delayed. And if you work with me, please don't do that to me. Not very nice.

Andrea (10:24):
Warning, warning.

Eva (10:25):
We like to be prepared as best we can. So that's sort of my anecdote for give it something. If you want it to give you something back that's useful, at least try and help it know who you are and what you're trying to do. So that's one of my dos along with being open-minded to trying when you're doing this for the first time or if you haven't messed with it much yet, it's a good idea to just give yourself some freedom. And so that is related to my second do, which is understand how the tools work. And this is kind of blowing my mind because there are tons and tons of tools. And of course they all have weird internet startup names that nobody can remember, which is really hard.That's uniquely an American problem. Yesterday, we were chatting after church about how to make slides and we have a friend who's an AI developer who's really an expert in the tool landscape

(11:31):
And he said, "Oh, Gamma, that's the one for slides. You got to use Gamma. It does the best slides. They've been out in front on slides for years." And I'm like, "I've never even heard of that, but okay, I was going to use Canva." So maybe doing a little research if you're trying to do something specific to figure out what tool is known for being great at the thing you're trying to do because ChatGPT's Kleenex, but it isn't designed for everything just like Gemini isn't for everything. And it's not a strange thing to consider that the tools we use today before AI were just built to be great at one thing. And the same thing is true with AI. I like that you're nodding Ken. Yes, there's a whole new tool set that we're all trying to get our heads around now. And so that's my second do. And then the third one, I don't know if this is real or not real, but when it gives me something I don't like, I actually tell it. And it is very inaccurate sometimes.

(12:38):
Doctors looking themselves up, that is not producing great outcomes right now. And I know we're going to talk about why and how you get it to produce great outcomes, but I routinely tell it, why would you give me that person's name as the expert in this thing in that city? They have never even done that procedure. And then it goes, "Oh, I apologize." You're right. The one that you're talking about is a better choice. And then I will say, "Where are you getting information? And it'll save vitals and health grades." What? I said, "Those are terrible sources. Please don't use those ever again." I refuse to accept results that include vitals and health grades.

Andrea (13:18):
I can't wait to hear more from this specific topic from Ken as we get into this a little bit deeper because he's got some really great insights that he trained our Studio three team about at our Client Success Summit, just of what we need to be doing to help our clients so that it is more accurate and so that they do show up in those chat searches. So I'm super excited about that part.

Eva (13:40):
I'm making popcorn right now, so let's go.

Andrea (13:43):
Okay, perfect. Which takes us to Ken. So can you provide us with what are maybe a handful of these must dos for AI when we're talking about SEO, we're talking about website design, we're talking about how doctors, those folks of our listeners and the people on the team so that they do show up in those search functions.

Ken (14:08):
Absolutely. Well, I'll start with saying, Eva, I agree with exactly how you put a lot of those dos, starting with learning how to use the tool. The one great thing about these chat products, they are one of the most unique tools in that they will teach you how to use them. So you don't have to necessarily go there and be an expert. You can be like, "Well, I don't even know. " We'll just start by saying, "What questions can I ask? How can you help me? Here's a problem I have. How can you help me? " And it will essentially prompt you along the way and it'll give you suggestions. Now, to Eva's point, it can lead you down a wrong path. So I encourage you not to turn your brain off when doing the process and just listen robotically to what it's saying. You still have to evaluate the information because we do know there's horror stories online where they will hallucinate and they will pull garbage data sometimes.

(15:09):
So you got to be aware of that. But nonetheless, overall, if you're just keeping an open mind and evaluating the data, it will generally lead you down a path where you can get insights from it and get a productive session out of it, whether you're trying to create content, whether you're trying to get ideation for something, whether you're trying to evaluate something. The one good thing AI does do pretty much in any usage scenario is it's able to ingest and synthesize vast quantities of data and boil it down into, again, whatever your prompts are asking it, boil it down to either actionable insights, give you statistical analysis, give you reference rate. It can save so much time. I'm sure our aesthetics clients that are surgeons and whatnot have also seen it in medical use and being able to evaluate x-rays, evaluate, diagnose complicated problems. And again, there's all kinds of opinions, whether that's good or not.

(16:20):
Beside the point, it is able to take a ton of data, simply analyze it in a split second and be able to return answers. So in that regard, I'll start by saying a number one do that I think is you have to focus on all engagement points for your brand online, not just your website. If you have a profile there, I guess I would reference it like this. If you have profile entities across the web, they need to be consistent and they should be all saying the same thing. If you have a bunch of reviews and they're all positive and they're all referencing the same point of a brand image, that's great. That's likely what's going to contribute to a better and better referencing of your brand the way you want it to be.

Eva (17:13):
Isn't it smart enough to know to look where you actually have reviews? So if I just think about reviews in general over 15 years, since we kind of started working on getting them, if you have 3000 reviews on real patient ratings, it's going to leverage those over a profile where you have no reviews, wouldn't it?

Ken (17:34):
If you're specifically asking it to review you, yes.

Eva (17:39):
Yeah.

Ken (17:40):
But if it's going, like this is where you have to look at it from the matter of, if you're looking at who is it referencing for Facelift and Beverly Hills, is it going to reference one provider that has 3000 real patient rating reviews versus these other three that have both high reviews in Google, high reviews maybe on RealSelf, high reviews on another-

Andrea (18:08):
Yelp. Something. Yeah. Yeah.

Ken (18:09):
Yeah, another review or two plus. Plus they have digital PR links, they have media...That's where all of those consistencies come into play. You could have a ton of reviews on health grades, but you have very little on Google and maybe you have some pretty gnarly ones on RealSelf. And then it goes, oh, well, again, this goes back to one of the other dos I was going to say in a second, but I don't know if I want to segue into that just yet. Eva, do you want me to go more into this or like-

Andrea (18:48):
I think something that-

Ken (18:49):
Do you feel like we're answering this well enough or do we want to move to another point?

Eva (18:54):
Well, what is the do that we're trying to say is like, I thought where you were going and I am enjoying this discussion. I actually think it's really useful.

Andrea (19:03):
Actually.

Eva (19:04):
Actually, AI is fire. I think this is what people would be asking themselves too is, and nothing has really changed. You still have to have reviews. And my advice about where to get them was always figure out where people are seeing you and make sure that's where your reviews are, like reverse engineer it. Is that kind of what you're saying?

Ken (19:26):
I think in the search landscape, in the discovery landscape where you're asking these chat products and whether you're talking about Google AI overviews even or AI mode, as well as ChatGPT and Gemini, you do want to focus on having more than just one review point, one major profile listing. You want to be in essentially as many as possible that all have a consistent messaging. Now, I know it's like easier said than done, but we're sending all of our reviews to Google. So now we're going to have to send everybody to Google and RealSelf and health grades and real patient ratings.

Eva (20:13):
Yeah. I don't want to go back to that world. It's been so refreshing to just say Google for however many years now. I think my ethical complaint would be that none of the sites have ever figured out how to do verified reviews. So regardless of which site you're on, my very, very informed opinion about reviews is that no one ever figured out how to make them credibly true. You still can't say that was a real patient because they can all be gamed. So none of them are good. Therefore, we should read no reviews and just move on. Don't read a review is what I want to say.

Andrea (20:55):
But they're so important And we know that

Ken (20:57):
They're so important. Yeah. I mean, you're stuck in ... And again, idealistically, I don't disagree with you, but the reality is they are used and users do look at them and hopefully users look at them with an open mind and they're able to understand that they're similar to what we're talking about with AI, garbage in, garbage out. It's a lot that are probably fake. And they should be evaluating multiple sources, but nevertheless, these are items that are used and they're leveraged. And the other thing that I think is important is understanding that the sentiment of valid reviews and user mentions online, and this is something that AI is also doing pretty well at, is being able to go into content, whether it's Instagram content or TikTok content or YouTube content, and actually read through both the content itself as well as comments. So it is expanding how you have to treat your brand strategy and that we want to make sure that we're not just getting a five-star review or a cute thumbs up on our post. We'd ideally like to elicit some engagement that has some sentiment to it where we see these users saying, "I went to X because of X, Y, Z." That's a great citation right there. We want that.

(22:38):
And then we will want to highlight that. We want to extract that, highlight that and make as much connection to that as possible.

Andrea (22:45):
When you're saying X, Y, Z, is that we want them to reference ... I want very tangible things that our listeners can take back to their patients that they're asking for reviews and things like that. Would that be something such as, would you mind sharing your experience about your breast augmentation surgery and your experience with the doctor and the team, calling them out by name? How can we prompt? So we have to ask questions to AI and to ChatGPT and be very specific with what we want them to give us back. We also need to be very specific with our patients. So what would you recommend are two things that we really want to make sure if we're requesting reviews from our patients that they include in those reviews?

Ken (23:30):
Ideally, you're prompting them exactly, as you said, with we'd like to encourage them to leave their review about Dr. X and their experience with the procedure and why they chose that doctor, why they chose that specific practice. Those elements I think are if you can essentially encourage the entity, the doctor, the brand, connected with the service and what made them different. Those three, if you put it on a triangle in your review content, you're going

Andrea (24:04):
Brand, Service and

Eva (24:08):
Differentiate.

Andrea (24:09):
Differentiation. Okay.

Eva (24:11):
Search engines can't really do ... They can see images and interpret images. Is AI doing a better job of that, or are we still mostly talking about written content that's actual text?

Ken (24:27):
AI does a actually pretty fantastic job at that. So they are multimodal now is what we call it. And an example, yesterday I was uploading images because I was just trying to position a picture on the wall. And so I was like, put it on the wall, holding it and take a picture. Does this look good? Should I move it? Does it balance these kind of questions? And my wife did it. So she was holding one, then I was holding one, and each time it would be like ... And there was no prompting from me on this, but again, just as a very simple example, it's like, no, the one the woman was holding right there, that's perfect. You should move it six inches over. Oh, perfect. You got it. The male that's holding it, and it's just answering with this naturally. I'm not prompting it like, "What do you think about the image the woman was on? " No, it was just like, "Which image looks better?" And it's able to decipher that simply just in contextual reference because it-

Eva (25:26):
How are you doing this with your phone through what?

Ken (25:29):
Yeah. Gemini.

Eva (25:32):
With the camera open in Google somehow?

Ken (25:36):
Well, I have an iPhone and I just snap pictures and then I throw them into Gemini and say, "Hey, tell me where I should..."

Eva (25:44):
You're still uploading.

Ken (25:45):
Is this image balance on the wall? Should I ... No, you want to have it six inches higher because then it's going to give space to breathe below this thing and that ... But again, like I was saying, is it recognized the male, female, just from being able to read the image.

Andrea (26:06):
Ask an engineer question.

Ken (26:07):
Yeah. So they do a pretty good job. Again, this goes back to your point earlier, Eva. The more you work with it, the more it kind of teaches you and you're like, "Oh, it understands that. I don't even have to ... " Great. So then I can start telling it like, "Okay, I want this content to be female specific and it's going to understand imagery connected to that. So it's going to choose and generate images." So we haven't even talked about the creative use of this. It can generate images. The Gemini tool now has what's called Nano Banana Pro that's its image generator. And I mean, it's fantastic. It can generate some pretty realistic images. And I think it is, from an ethical standpoint, important that we call out. Obviously on a website, we're not going to be generating before and after images. I would never-

Andrea (27:03):
So that would go on our don't list. ... never

Ken (27:04):
Promote that. And I would never agree to do that for a client. I would say, absolutely not. We're not going to do that. But you can do model representation and you can call out that. This is like a model patient or this is a model blah. You don't want to say actual patient. You do not want to cross that line. And that's kind of something that's been pushed within the content creator world, so to speak, with AI is like, use it, but just make sure you're identifying that, hey, this is AI generated or whatever so that there isn't that ... We're not trying to pull one over on anybody or try and create, but there's a massive amount of content that's being produced. I had a insider in Hollywood tell me up to like 80% of their commercials nowadays are AI generated, like all the car commercials and this and that.

Andrea (28:04):
So the Lexus is not sitting outside of the garage with the big red bow on it on New Year's Eve. Is that what you're telling me, Ken?

Ken (28:11):
The Land Rover is not driving through the desert and parking in a cave that suddenly appeared in the middle of the Sahara.

Andrea (28:18):
Damn it.

Eva (28:18):
No.

Andrea (28:19):
They're broken. That's really sad.

Ken (28:22):
Yeah.

Andrea (28:22):
I'm going to have to process this. Okay, guys, I'm back. Okay, perfect.

Ken (28:26):
Sorry, I got off on a tangent there, but if I go back to what we were talking about originally, which was-

Andrea (28:33):
The uses in SEO and website design, what are a couple of things that ... Studio three, obviously we have hundreds of aesthetic clients across the nation and internationationally, but if we're not the people, because I know that you are, as our chief strategy officer, you're the one that's really leading this charge with our SEO teams and our design teams and creative teams, they're not Studio three clients. What should they be talking to their digital marketing companies or their web companies about making sure ... What should they be doing and looking out for?

Ken (29:08):
Right. So the number one thing we started talking about was just making sure that their brand is consistent and consistently referenced, and it's referenced in as many engagement points online as possible. That's kind of number one. I think look beyond just your website. Number two is on your website itself, you want to make sure that we're reframing a lot of the content beyond just your bottom funnel strategy, and you should be doing that. Definitely don't want to take attention away from that, but that's not enough anymore. That's kind of the more the point of this is that we have to go beyond just focusing on the bottom funnel and kind of the quick wins that drove a lot of business online for so many years. Now with this, we're seeing a lot more discovery happening within these AI platforms. We're seeing a lot more discovery on social.

(30:04):
We're seeing a lot more discovery on YouTube. We're seeing a lot more discovery on these adjacent channels where it's like, okay, how do we appear there, first of all? So that's kind of like having strategies for those individual platforms,

(30:19):
But then also those then lead into discovery within chat and within tying that as you're making your brand more and more prominent and you're not so much focused on just the bottom funnel, you have to maintain that, but now you're winning earlier and earlier in the user's journey. And so you're being referenced more, they're getting impressions of your brand, they're getting what we call citations of the brand mentioned in ChatGPT and Gemini. Ultimately, we're seeing users do search discovery in these platforms, they're going to social, they're getting sort of third party verification, and then they're going to Google and when they're ready to make that commitment, they're going to the website, then they're engaging with the brand directly on the website. So you may see, and we've seen this with a lot of clients, brand search continuing to uptick, and we may see impressions continuing to increase for more and more in different items.

(31:30):
And you'd say, "Well, we're seeing a little bit less in terms of the bottom funnel." We're seeing less breast augmentation surgeon in Beverly Hills, but that's fine. We're becoming more and more aware earlier, and then they're directly searching for Dr. Joe because they've now seen his content in a number of different sources. They're ready to make a move. It may not be the same day, because we know a lot of these users' journeys take weeks, if not months. I have a client, in fact, that has done this study with the patients he's signed, and on average, it's a year and a half to two years of study that his patients go through before.

Andrea (32:16):
Is that a facial plastic?

Ken (32:18):
It is.

Andrea (32:19):
Yeah. Well, typically facial plastics, it's a longer rev cycle than it is for body. Yeah.

Ken (32:25):
Sure, exactly. And I think that's exactly right. But I mean, just as an example, so that's, again, another datum to think with. It's like, okay, so we have to make sure in those cases we have a focus to get those top of funnel because we want to be showing up during all those searches as much as possible, right?

Andrea (32:48):
And just to kind of summarize that into a sweet little package is what you're saying is in order to come up in those chat searches, we have to be diversified amongst many different platforms within aesthetics, not just our website, not just social media, not just RealSelf, not just get Google reviews, but really diversify the strategy so that when chat goes out there and searches of everything that they can, they're seeing consistent messaging for you as Dr. Smith that does facelift and rhinoplasty as opposed to maybe people that don't have consistent messaging across multiple platforms. Is that fair?

Eva (33:28):
I want to get really tactical here for a second, Adria, because I don't want anyone listening to this thinking, "This is so overwhelming. How could we possibly be everywhere?" Because that's the actual problem that I've been trying to solve for five years and I think we're doing a pretty good job. And I was able to connect these dots when I spoke to Ken last week, which was really exciting for me. It's like a great work Christmas present for me and my team. And so when you're talking about being everywhere, you can do this by creating one thing and putting it everywhere.

(34:04):
You don't have to create bespoke content for every single channel that you need to be in. And so the way that ... We have a mutual client where through the combination of the great data that Studio Three gives us and me knowing exactly what we did, and we really only did one thing. And I'll tell you what that one thing was. It was that in this practice with four surgeons and like seven mid-levels, I basically, I won't say forced, but I strongly encouraged them to sit down with me once a month and record. And so they took turns every month. Through one year, we were able to make 92 YouTube videos and those were also repurposed into blog posts, web content, and then we embedded all of those YouTube videos individually on every procedure page with the help of Studio 3's team. So if you go look at a breast augmentation page in this practice, sometimes now there's 12, 15, 18, 20 videos on the page and they're all in that Q&A format.

(35:10):
And so Ken, this is where you explained something to me that I hadn't recognized yet, which I think you called it Meta query or was it somewhere- Query

Ken (35:21):
Querey fan out.

Eva (35:21):
Query fanout. Can you tell us what that is and why that was important?

Ken (35:26):
So essentially what a lot of these AI products are doing, and this is something I believe Google patented or Google actually has a patent on it. So for the real nerdy people, they can dive into that and decipher it. But essentially what's happening is you put in a prompt of X, Y, Z and users are getting better and better at being more and more specific at their questions, by the way, which is why having these question and answer format content video that Eva's talking about is perfect for that. You asked the LM this specific question. To answer that question, it's not just going and looking for that statement and trying to find a specific one answer to return. It's going and it's now doing a fan out of queries and looking in multiple directions all around that to synthesize an answer. So it's pulling data from this site and from that site and from these four or five different sources, putting them together, stitching a cohesive response that's coherent like a human would, and then giving you an answer.

(36:44):
I give this example. It's like when you go to answer something, if I ask you a question, I say, "Well, what do you think is the best place to ski at in Colorado?" When you go to answer that question, you're probably not just visualizing one paragraph you read from one site. You're thinking of like, "Oh, well, I went to Aspen and I went to this place and oh yeah, and the slopes were nice and I didn't have to wait for the chair lifts very long and the prices was decent. Oh, and there's great cabins." And you're thinking of like, but you're pulling all these different data sources to come up with that answer. And that's literally what this LM is trying to do. It's just doing it from, again, multiple different sources. And it does that because it asks all those different questions. So again, if you think about it from a computer, how do you actually replicate that?

(37:37):
It has to ask that. It has to go, "Well, what place has the best slopes? What places offer decent price? What place is it? " So it's asking these like 10 different query fan outs right away, pulls all that information together, stitches together an answer for your original question, gives you that. And when you utilize a lot of these models now, they have a deep think or a deep research or whatever they call it, and it'll show you a lot of that as it's kind of doing it. It'll kind of say thinking for two minutes and evaluating these things and asking these questions and then it kind of spits out. And it'll, again, sometimes in its answer, if you have this kind of training in your prompting, it'll give you all the ways it kind of came to that and then it'll give you, you got to read through a lot.

(38:31):
So it's like some people are like, "No, I just want the fast stuff. I just want to know blah." But even the fast models are doing that same cycle just on a smaller scale.

Eva (38:42):
The short answer to the question is really, if you're creating content that demonstrates your expertise and you're putting it everywhere, then you're doing what you need to do to be included in the answers in the models.

Ken (38:57):
That's right. And I should say it's important for our clients to understand this because the primary purpose of these chat models is not to be a ranking system. They're not trying to give rankings like a search engine does, which is-

Eva (39:14):
If you get an email that says, "We'll make you number one in ChatGPT", you just need to delete it and report spam.

Ken (39:19):
Yeah, exactly. It is not that. It's a chat. It's trying to have a conversation and it is trying to answer your questions in the best way that you prompted it based on your history. Again, now people don't understand this. So again, depending on the model you used, which is another thing, certain models are trained on data with a specific training bias. They're not actually doing a real-time search of information. So new information may not even be collectively brought in. If you're using an older model, if you're using a free version. So there's also variables here that we have to understand. I've been on calls with clients of like, "Well, I did the same search and I got a totally different answer." I'm like, "Well, of course

(40:10):
You're interacting with a chat model. It's based on your bias, your history." It's not going to be the same answer for every user that searches this. A lot of the newer models and a lot of the pro models are tied to an active search. And if you prompt it to look for things, it will, and it'll be able to pull real-time information. But those are another call out I want to make because it's just a common misconception that people have like, "Well, why am I ranked here, but I don't show when I searched for X?"

Eva (40:42):
The number I would look at, and you tell me if you would look at the same one is when we're talking about search results, we're still talking about Google. The top line number in Google search console is search impressions, and now that includes Gemini showing you in search results. So that number, at least in the ones I have access to, has continued to climb as Gemini results have been included, right?

Ken (41:11):
Search console should be one of your primary tools to use because we do know that data is ... It is good and it's solid data. The other thing is in terms of our strategy going forward, you want to primarily talk about your brand voice, how much you are being referenced for prompts that are more useful in relation to what you're trying to do. You can essentially build a series of prompts to track and to see, "Okay, good. What is the sort of ratio that I'm showing?" Because it isn't a ranking system. It isn't like, why am I not number one? We got to get away from that whole viewpoint and we got to go, "Okay, are we being referenced? Can they find my content? Do they associate me with this particular service item and can I increase that? " And that's a more holistic approach we have to take to it.

Andrea (42:09):
I love that. And I think that's a really great summary as far as what I hear from clients on a regular basis is, "Well, because we use LeadLoop, so we're able to track where people came from and it'll show that they came from ChatGPT and they're like, we had this patient and they came from ChatGPT. How do we get more patients to find us there?" So that's really helpful, Ken, in saying this holistic approach that really ... I think something that you said in a presentation I heard you do was think of ChatGPT as you're providing it feedback and using it for search and things like that is it's really going out there into the internet and trying to provide you a response that your best friend would provide you. And so the goal from the practice perspective is how do I make sure that I have enough content out there in various places so that we are chosen as that person that's served up to the patient?

Eva (43:08):
That's all we do now. That's my 100% of my work today is how do I create content really fast and really easy for the practice?

Andrea (43:17):
Yeah.

Eva (43:18):
So I'll just tease that we have a way to do that for anybody and we'll come back to that at the end too, but I just wanted to throw it in there.

Andrea (43:26):
I love that. Coming at it from my perspective where I'm more like in the practice, working with the practice teams, working with how do we manage our leads once they do find us, whether it's through chat, through Google search, through a place like RealCelf or social media, Instagram, TikTok, wherever people are finding us. And then they call, they fill out a form, they end up actually contacting our office with a couple different clients of mine right now as far as what to do about AI inside with the practice and with the team. What are my dos? My first one would be as a team member, as one of our listeners here, don't be scared that AI is going to take your job because I'm a true and firm believer based on everything that I've seen, that there are tools out there now that I'm actually pretty impressed with.

(44:16):
But in aesthetic medicine, the thing that gets patients to take action, I'm still a firm believer in the data will show me this all day long, is the relationship and it's the trust and the safety that our patients are feeling when they have true interactions with us. And so there are AI tools that can help those really great candidates kind of get to us. And I would just say, first of all, do just be open to the same thing that both of you said.

(44:48):
I have two, like I was saying, two clients right now. One is using an AI function to take after hours phone calls. And so it's an AI respondent. And to be completely honest, I was shocked at how impressed I was with this AI agent that answers the phone. It's incredible, you guys. It's so cool. And it says specifically, you can tell, it sounds like a human. The tone is really good, but they are upfront. This is an AI agent. We've trained it really, really well. And so far I've been absolutely impressed with this. And from the team perspective, the way I would encourage everyone to look at this is not to be threatened or scared AI has taken your job, but just say, "Oh, this is great because this can help us weed the right people in a little bit faster." So if they are willing to engage and to provide their information in a time that I can call them the next day, it makes your job easier.

(45:45):
You don't have to call them and play phone tag and all these things. You have this AI agent that this one specifically is integrated with the EMR of this client and it'll automatically put the appointment for the patient care coordinator to call the person the next day on their calendar, which is so cool to me.

Eva (46:04):
That's really interesting because scheduling is super hard and that feels like a baby step.

Andrea (46:09):
Exactly. And there's no phone tag. There's no sending three emails back and forth to figure out what time can I call you tomorrow. The AI agent has this nice conversation. What are you interested in? Oh, I'm really excited for you to talk to a patient care coordinator. They'll be able to learn about your goals and then give you information about the doctor, the procedure. They'll give you ballpark pricing and then lead you through the consultation process when you're ready. What's a good time for you to have a call with Melissa tomorrow? And then they take the patient's name, phone number, email address, figure out the time, look at Melissa's like this is literally integrated with the EMR. Look at Melissa's schedule, put the person on the time when Melissa's available and she just calls them the next day. This is AI helping-

Eva (46:55):
That's truly helpful.

Andrea (46:56):
Yeah. It's not taking away anybody's job. All it's doing is streamlining our ability to serve the patient. So again, totally different than the chat that we've been talking about so far. But I think these AI tools can be so beneficial for the team within the practice, which is really kind of our target listener or the people doing the work in the practice. It can save you a ton of time to be able to just ... Our goal is to get someone on the phone so we can talk to them, so we can build trust and rapport, so we can educate them, so we can differentiate our practice, so we can start to build a real relationship. It's not taking that personal element out, it's just making getting there a lot faster and a lot easier. So be open to it. And if you are, and if you are going to implement a tool like that, just make sure it's trained really well before you launch it to your patients because when it's not trained well, which we saw with one particular client, it can be a real hassle and it's potentially going to end up losing you money because it might not respond the way that you would want it to respond.

(48:03):
So you've got to go through your due diligence. You have to train it. You've got to really test how it's responding to your patients so that we're not off-putting because there's still so many people that are like, "Oh, this sounds like AI. Screw this. " But if you don't train it with the empathy, it will.

Eva (48:20):
You and I had one of those really funny texts a couple weeks ago where you were telling me that they were trying to replace a coordinator with AI and I think I sent you like 17 barf emojis.

Andrea (48:32):
Yes, you did.

Eva (48:33):
All joking aside, the key driver of a patient choosing a surgeon has always been and will always be surgeon made me feel comfortable, which represents that relationship building aspect. So if you want to replace the consult with technology, you do so to your apparel. And I know none of us will ever say that's a great idea. It's just not what women do and will never do. We will never do that.

Ken (49:01):
I think that that's a very good point. We don't want to replace the human experience. We actually want to replace the more repeatable elements where you can bring efficiency and scale with AI. Like you're pointing out, Andrea, if you're getting 500, 600, 700 leads a month, which some of these practices very well are, and even more, you're paying likely lots of money through marketing to drive these leads in. And in the old school style, you'd either have to have a call force of 10, 15 people that you're paying for that's likely outsourced. So then again, quality becomes an issue. You have no idea what they're saying. They're working off of scripts, but who knows whatever. Or you go, you're losing those leads because you don't have people answering the phones quick enough. And then just half of them are going to voicemail and then you're playing the phone tag game, trying to get them back.

(50:01):
And we all know one of the largest buckets for lost leads is out of touch, you couldn't actually reach them. So hey, if this can increase that efficiency and now you've increased your actual conversion from the inquiry phase of like a phone ringing or web form to a booked console or even pre-consult with a coordinator.

Andrea (50:26):
Exactly.

Ken (50:26):
That's fantastic. It's more than beneficial.

Eva (50:30):
To both of your points, and I just want to compliment you and your team for this piece of it, that you also have to have data and be really tracking closely where things are working or not working. And if you're not doing that, you're just stuck with your own anecdotes and opinions about, and probably will end up making bad decisions because you only will feel the ones that didn't work and not the ones that did work. I analyzed an entire year of leads that happened to be also a Studio three client. And it's just coincidence that both of my two projects happened to be using Studio three. Is it a coincidence? Maybe I just love you. I don't know.

Ken (51:16):
What's not the love?

Eva (51:17):
I love the data.

Ken (51:19):
Not paid for mention.

Eva (51:20):
Not paid for. No, I love the data.

Andrea (51:22):
Not an ad.

Eva (51:22):
You give me access to the raw data and to the lead so I can actually do all this analysis myself, which is huge for my own trust of anyone that I work with. And I looked at all of 2025, and what I noticed was that we started getting leads attributed to ChatGPT about six months ago, and every month there was slightly more than the previous month, so I was going like this. But then I went, "But were they good leads? Because who wouldn't ask that question next?" No one. And I discovered that only two were not from inside the state that the practice is in. So that's a pretty good ratio. 90 of the 92 leads were in the state of the practice. So then my trust went from here to here, and it's not hard to find.

(52:17):
And if you are listening and you want to look at that, any one of us can help you figure that out.

Andrea (52:24):
Sweet. Well, this concludes the section of what do we do? How do we really utilize this tool and start being more open-minded and really bring it into the practice in a way that can be helpful and not scary. Don't be scared. It can actually help your job really become a lot better and more effective and more efficient.

Eva (52:47):
You guys know we're long-winded and there was so much to cover with Ken that we're making this a two-part episode. So come back next week to hear how not to use AI and the mistakes you want to avoid.

Blake (53:01):
Got a wild customer service story or a sticky patient situation? Send us a message or voicemail. If your tale makes it into our "She did what?", we'll send a thank you gift you'll actually love. Promise, no cheap swag here.

Andrea (53:13):
Are you one of us? Subscribe for new episode notifications and more at practicelandpodcast.com. New episodes drop weekly on YouTube and everywhere you can listen to podcasts.