Episode 05 | Conversational AI & its Impact on Patient Care の全文書き起こし

virtual agent or a chatbot for example. It encompasses a variety of technologies that work together to enable efficiency, automate communication via text or speech by understanding customer intentions, right? So it's this five, it's this. It's really understanding the language, the contexts, and being able to respond in a humanlike manner. In a nutshell, that's what it is. Chris, I think there's a lot of folks as you said, you know, kind of, uh, confuse that. But in a nutshell, um, you know, I think those are the key elements of conversational ai. Chris: 4:02 So Charlie, I mean, number one, conversational AI is such an interesting topic. Um, and you know, it's AI in general or artificial intelligence is used in a bunch of different industries across a bunch of different mediums. I'm curious to know, like from a life sciences standpoint, where is it used within life sciences that makes the biggest impact? Charlie: 4:26 Yeah. Over the years, it's really picked up in the life science industry, right? I mean, if you, if you go back, you know, the pharmaceutical organizations have really been kind of slow, in really embedding, you know, artificial intelligence in general. Right. But in the last couple of years, it's really picked up. You know, if you look at some of the top 50, you know, life science organizations, Fortune 500, life science organizations, they really started looking at where is there an opportunity to improve, you know, the self-service from our, for their consumers or for their healthcare professionals. One of the, the key areas that we've seen a dramatic push right now is, you know, I'll mention a few, you know, financial assistance, for example, you know, financial assistance for consumers, you know, is really, you know, providing them help instantaneously getting access to that information without really, you know, speaking to a healthcare professional. So it's having that self-service drive. You know, for those consumers, that has been probably one of the, the easiest areas where, you know, conversational AI has been implemented. Um, things like product availability. I want to know about, you know, what kind of products do you offer? Same thing, you know, these are areas that are really easy to implement in the life science space where it's not, you know, it doesn't require, you know, additional scientific ways that a healthcare professional needs to help that consumer. It's self-service, right? These are all self-service tools and implementations that are really easy. To integrate, you know, into, for example, a medical information contact center. There is others, you know, sample requests, sales rep information, like, you know, having virtual assistance for the sales rep to understand, you know, what's available. Um, devices, right? Medical devices, another area of opportunity. How do I, you know, operate this medical device. These are all how to questions. You know, a virtual assistant or a chat bot could answer.

Uh, one area that's really starting to kick off is really program enrollment, right? Enrolling, you know, consumers, um, into specific programs. And this is an area that we believe is going to be, you know, very popular in the next, you know, uh, six to 12 months. Another area that, you know, we have started to see. It's a little bit more complex, but medical information, uh, groups such as, you know, areas where you want to search for medcom information, large, you know, read large documentation and be able to make a decision, you know, that you require for your patients, right? So implementing what we call optimize, you know, search index, um, N L P capabilities. Is something that we are, uh, heavily involved in with, uh, some major life science organizations. So these are just a few areas of opportunities, but there's many more Chris. But I think these are the key areas that we're seeing right now. Chris: 7:45 Talking about conversational ai, um, could you really hone in on, and obviously, you know, AI in general has really started to mature, especially over the last 3, 4, 5 years or so. Can you kind of hone in on those qualities of AI that you kind of see standing out that ultimately are gonna lead to the success of artificial intelligence being crucial in things in life sciences, like patient engagement and, and different types of, uh, engagement factors. Charlie: 8:20 Chris, this is probably one of the most important things that I feel, uh, you know, folks have not been concentrating on at all. For me, you know, implementing artificial intelligence in the life science space is going to revolutionize the industry. Life science organizations that are not doing some form of AI today are going to be left behind. And let me explain why. First of all, AI for me is providing you data, right? It's all about the data. And data is enables life science, you know, leaders to make strategic decisions. So having that AI tool and having those analytics is going to be able. You know, to engage not only patients, but your healthcare professionals to make better decisions. Now, how do you do that? Right? Um, having the right tools, having the self-service tools, having the omnichannels where you're engaging, you know, with your consumers, your patients, and your healthcare professionals, all on one model. Right. And how do you do that? Well, you know, if AI is not knowledgable, if it's not engaging, if it's not empathetic, you're not gonna wanna do business with a virtual assistant, as an example. So this has evolved, you know, um, I'll give you an example of our own conversational ai. You know, we've been, you know, learning the, the bot, if you may, or the AI power behind it for over 15 years. It understands the lexicon language of life science organizations. And this is a critical feature in the success rate when we're implementing AI solutions. Leaders expect instantaneous results after day one. Well, guess what? It's not gonna happen, right? Uh, I, we believe that artificial intelligence is an enhancement to your current process to make strategic decisions, but you always have to have that human interaction. So we believe in human-centric technology. So let me explain that. When you implement a chat bot as an example, you always have to have a human that is reviewing and in real time those transactions why? That's what makes the conversation, you know, a lot easier for the next patient if that individual did not get the answer the first time. You want to make sure that you know those self-service tools you're giving those patients, you know, they, they want to come back and use those self-service tools. So it's really critical that your chat bot, for example, is empathetic. It understands natural language processing, It understands your intention. And if it, it's not satisfactory to the patient, give them that option to speak to a live healthcare profess. The full circle customer experience is extremely critical, you know, in, in measuring the quality of your success. The last thing I'm gonna touch on is quality and compliance. This is probably the area of biggest concern for a lot of organizations, right? How do you detect, as an example, artificial, you know, AEs or PQCs? This is, this is something that we believe is gonna take some time. You know, there's not one chat bot out there, or technology that I believe is a hundred percent detecting artificial AEs. Um, you know, from a chat bot perspective, you know, we, we have linguistics. They're reviewing this data on a daily basis and being able to engage and mature that chat bot along the way. So it's really having that human-centric, you know, philosophy, but taking advantage of the technology. Chris: 12:21 So Charlie, what are some of the challenges that AI or conversational AI in general faces today? Or challenges and or myths? Because I feel like sometimes, you know, ai, there's some misunderstandings about AI in general, and then why ultimately, and you know, this is kind of a two-prong question, why is differentiation so important? Charlie: 12:46 Yeah, Chris, I think, you know, there's people that accept artificial intelligence technology, they love that opportunity to self-serve. And then there's people that don't like it, right? It, it's really those, those are the two variations here. I think there's more and more acceptance, um, of artificial intelligence because it's everywhere, regardless of what you do in every industry, there's some form of artificial intelligence that's being played out, you know, from a consumer or patient perspective. So I think the first thing is that people think that, you know, AI is gonna take over, you know, human jobs. And I think that's the first myth. I think it's, it's just gonna enhance and speed up the process to get you the information quicker. That's what AI does. A human will always be involved to make sure that you know that AI technology is improving on a daily

basis. Let's face it. How many times have you transacted with an AI technology where you were frustrated cuz you didn't get the right answer? Right? These are not myths. I mean, it still happens today in the life science space. It is super critical. You canot afford doing those kind of mistakes. So before implementing any solution, you need to stress test it. You need to make sure you have the right people working on it. Um, this is why having a healthcare professional linguistic on your team that understands, you know, the patient journey, understands that what the doctors are looking for, you know, quickly. For example, you know, if I'm going through Medcoms and getting information through Viva Vault and I'm being frustrated cause I have to read 300, you know, pages to get one document. Well that's where ai, you know, is very powerful. That's where, you know, AI can help. They cypher through the information, streamline it, and be able to get you that information instantaneously. And I'll talk a little bit later about a use case, but the, the, the fundamental you know, thing about AI is that we're still, you know, at the preliminary stages, you know, I feel in the life science space, I think there's still a lot of work to do. It's not at a hundred percent mature le maturity level at this point, but I think in the next couple of years, we will definitely be at the, you know, at the same level as some other industries in this, in this space. Chris: 15:20 So for me, honestly, Charlie, one of the things that I find the most interesting about artificial intelligence is that in order to be successful with ai, The success is judged by whether you know it's there or not. So, like, if it's not there, if you feel like it's not there, then you're successful. You know, that, you know, you're not chatting with a bot, with a bot, but you feel like you're chatting with a human. Uh, you know, that that's, I think that's what's interesting to me a little bit about AI too as well. Charlie: 15:52 Yeah. I mean, you know, you, you wanna make it, um, that it's, you know, conversational, right? That's why that word is always being used, but you're absolutely right, Chris. I mean, I get frustrated when I'm testing technology and I feel that I could tell that I'm talking to a robot, right? Um, in the life science space, this is super critical. If I'm talking to a patient that, you know, requires advice, you want to feel that, you know, that chat bot really understands those intentions. Right. Uh, our technology has the conversational intention, you know, lexicon, and we've been very successful in implementing this in many organizations today. Um, so I think this is the future, um, is providing that, you know, empathetic way of speaking to somebody. You're absolutely right Chris. Chris: 16:45 Um, Charlie, so when you're working, you know, on the variety of different artificial conversational, artificial intelligence, um, projects that you do work on,

um, what for you is the biggest challenge in, in development overall? Charlie: 17:02 Number one, education. Right? It's really educating, you know, first. Um, you know, the life science organization that wants to implement the solution, um, you know, a lot of, um, a lot of leaders, you know, have a perception of AI is gonna solve all their problems, right? And the reality is, you know, AI is an enhancement to their existing process to improve efficiencies and let their healthcare professionals take care of more you know, difficult transactions, I always use that analogy. So education is number one. We spend a lot of time educating, you know, leaders, patients as well, obviously, and caregivers because there's an expectation level when, as you said, speaking, you know, to a chatbot for example. So setting up that education is really critical upfront, so people don't, uh, necessarily expect to have, immediate you know, results, uh, you know, in within their business. Just to give you an example of that, Chris, you know, when you implement a conversational AI chat bot as an example, typically you start seeing some really good results after four to five months. Um, and the reason for that is the bot really starts understanding, you know, the lexicon language of your organization. And it has enough data to really, you know, focus on the key, you know, the key lexicon that you're looking for to implement, you know, within your omnichannel. So that is the, the number one reason, the number one challenge that we've always faced. The second one is training. Not only training the bot, but treating your people right? Training your people on detection, and what are the things to look for, right? As you, uh, look at your AI brain power. You know, are you analyzing the data, what you think you will, you were trying to implement? Is that what your customers are telling you? So make sure you survey, you know, your chat bot. Every single transaction that you're going through ensure you, you provide, you know, the patient or the healthcare professional a way to provide you feedback. Its that's gonna go a long way, and that's what's gonna make your, your next, you know, transaction a lot more fruitful. So I think that's the, the key. The, you know, another area is really, you know, reliability, you know, or having access to resources within your organization. A lot of times we start, you know, with a discovery session, right? And we start laying out to the life science organization, Hey, these are the steps that it's gonna take for you to implement this solution. And you know, they, they always think, Hey, you, you know, here's ai. You have the software, switch it on, and the way you go, Well, it doesn't work that way. I mean, there's implementation, testing, you know, making sure that we have the right flow. So it does take some time to get it to the level where you want to be, but typically, you know, an implementation, you know, in a life science organization could take anywhere between 16 to 18 weeks. And those are the things that we like to ensure people are, are, are aware of, and we have the right resources, you know, allocated to both, uh, to both areas, to both parties. Chris: 20:30 So Charlie, obviously, you know, uh, you know, conversational AI continues to mature and grow and expand. Um, and obviously it touches such a wide variety of different industries and within our own industry, life sciences, it continues to expand, um, every single day. Um, where do you see conversational AI going into the future? Charlie: 20:56 Yeah, I mean, you know, Chris, um, I look at it this way. Imagine smart digital assistance that can provide unlimited amount of conversational information, power to help patients navigate the administrative complexities of healthcare, right? It doesn't replace the human teams, but rather it augments them, right? And most importantly, it provides the engagement scale that healthcare desperately needs today in this industry. Covid 19, forced healthcare to virtualize and conversational AI is becoming part of that experience, right? Most patients don't even realize the change because the user experience is so simple. It's simply works. You know, that I could reassure you this is gonna be the future. For me, conversation AI or digital assistance. Needs to be part of a business's DNA process. Automation initiatives, handling a range of customer support needs... all of these things have to be part of your DNA to be successful. Organizations that are process heavy and resource constraints stand to benefit most, and life science industry may be the biggest beneficiary of conversational ai. I really believe that. Pharmaceutical companies need to have, you know, need to have this part of their SOPs or standard operating procedures to continue to innovate and truly understand the benefits of what patients require. You know, you look at some of the statistics of Gartner or any other publication, um, you know, only about 35 to 38% of life science organizations really have budgeted for AI initiatives. Right? And I really believe this is the future. They really need to get on the train because it is probably the only industry that is still not heavily, heavily involved in this area. But I really believe in the next five years it's gonna kick off. Chris: 22:59 So, Charlie, I know you've been working within, um, and implementing AI for, for quite a while now. Um, so you have, I'm sure a ton of great case studies and, and, uh, good examples of how AI has been implemented successfully, um, and how it also has created efficiencies, um, and, you know, really help streamline processes and interactions. I'd love for you to share, you know, one or two different cases, um, that you know, that you really feel kind of stand out and show off the capabilities of our conversational AI platforms. Charlie: 23:38 Yeah, for sure. Chris. I mean, you're absolutely right.

There's quite a few, but let me, you know, let me talk about one in particular. You know, this is a major, you know, Fortune 50 life science organization. You know, they came to us to Alphanumeric and said, "Look, we have a global contact center and we really wanna reduce the number of interactions that our healthcare professionals are handling on a daily basis". So they were, you know, we, we looked at their portfolio. Basically, they were handling approximately a little over, um, you know, 50,000 transactions or interactions, I'll call 'em. And basically, you know, year over year they were seeing anywhere between 28 to 30% increase in interactions. So more and more transactions were actually being handled by the contact center. More call volume, more omnichannel transactions, social media was up. All of these different omnichannels, whether they were, you know, email, you know, conversations, whether they were social media conversations, they kept on seeing a steady increase in terms of contact center. And obviously it's costly, right? So how, you know, they, their biggest challenge is how do we help them really reduce the volume going into those contact centers and enabling AI to take care of those transactions as much as they can. So what we did basically is we implemented, you know, a virtual assistant in the front end of a lot of their websites or their areas where they, their patients were actually interacting. Whether they were sending a web form through the website, whether they called a one 800 number. You know, what we started doing is, you know, deflecting is the word that I like to use all the time, Deflecting those transactions from going to the contact center. But enabling the patient or the healthcare professional to self-serve themselves. Right. So what happened by implementing our virtual assistant, the number of live contacts that were transferred to a healthcare professional decreased by over 34%. That is a substantial amount of percentage when you look at it, you know, from a contact center perspective. The accuracy rate of this virtual, uh, virtual assistant is actually still running. It's, it's been running for the last couple of years. You know, we just keep on adding to it, but the accuracy rate is 91%. Now, you might say, Well, you know, why is it not a hundred percent? You remember me telling you that I don't believe that AI will ever be at a hundred percent, or at least not in my lifetime. Um, That is a phenomenal number. So what that means is the, the virtual assistant was handling a tremendous amount of transactions, you know, and not, not deflecting that to the contact center. It was 91% accurate. Those, the defects that were not accurate, you know, the remaining, you know, the remaining, uh, 9% was really the linguistic coming back and starting to work on, you know, those areas where to improve. And what we ended up finding out, it wasn't because the virtual assistant couldn't answer, it's because the product knowledge that it was type, you know, going and get the information from, from that life science organization was not correct. So they had to rewrite some of their product knowledge. Which enables them to say, Hey, let us look at all of our information now. Right. So, uh, uh, definitely something that is critical in the life science space. They actually, overall, have been able to deflect 77% of their transactions. This results in a $1.4 million saving. If you want to put it into FTE's in the space, it's approximately, you know, seven fte. Now you might say, you know, what does that mean? What this organization in particular did didn't necessarily reduce seven resources. They, you know, they felt that those resources could be used in other areas where they could be concentrate on more difficult, you know, transaction. And that's what they did. They just basically reallocated those resources in areas where they could be more useful in their scientific ways. So, uh, a true story of success. We have many others like this, Chris, so I could probably talk about it for a very long time. But this is truly a, an example of the power of conversational ai understanding intentions. Chris: 28:30 So Charlie, as, as, as you're aware, obviously at Alphanumeric here we have our Make Your Mark campaign, um, and all of our podcast, um, you know, subjects, our interviewees, um, we've been asking this the same question. Um, you know, how, how do you go about making your mark? Um, we'd love to hear from you what you do in your day to day, um, and how you make your mark. Uh, Your industry, the light, your life in general. Um, we'd love to, to learn a little bit more about Charlie. Charlie: 29:02 I'd love to, to speak to folks, learn new, um, you know, new concepts. Uh, I love to innovate. The way I make my mark is really finding ways that I can help, you know, our patients, our, our, our staff, our employees to do their job, you know, quicker, more efficiently, and provide that ultimate customer experience. I think. You know, a lot of organizations, you know, have great processes in place, but that experience for me is making that mark what's gonna make that employee continue to work for alphanumeric and continue to serve our customers. Right? It's really going above and beyond and being innovative. I challenge my team every single day on finding innovative solutions, and for me, that's making, you know, a mark. Chris: 29:54 Well, number one, wanna thank you. Um, you know, I know, um, you know, your time is extremely valuable, but we love hearing from you, Charlie, um, the insight that you provide. Um, you know, number one, uh, to us and, you know, understanding artificial intelligence, um, understanding, you know, the, where it started and kind of where it's evolved and where it could potentially go. You know, it's exciting. Um, and, um, you know, we definitely feed off of your excitement and your passion towards artificial intelligence and just in general. Enhancing patient engagement. Um, and that's, you know, that's very exciting for us here, all at alphanumeric. So want to thank you and, um, you know, wish you nothing but the best and, uh, we'll talk to you soon. Thank you for listening to this episode of Make Your Mark an alphanumeric podcast. For more information on Alphanumeric go to www.alphanumeric.com and follow us on LinkedIn, Twitter, Facebook, or Instagram. Thank you, and remember to always strive to make your mark.