Is It Really AI Chat? A B2B Buyer’s Guide to Evaluating Chatbot Tech

 

Speaker 1 0:04
Okay, hi everyone and welcome. I am Pallavi Shukla from the marketing team at botco AI, and we are so glad you’re here. We’re excited to share with you practical insights about evaluating chatbot technology. So let’s get right into the topic, because we have just about half an hour and we have a lot to cover. Today’s topic is, is your chat bot really intelligent? AI, or is it just a scripted interface, a scripted bot with a sleek interface and clever branding, there’s a lot of noise, and Everyone claims to have real AI, but as we’ll show you, the difference that real AI in a scripted bot is very different, is very different, and the difference is huge, especially in industries where accuracy and compliance and personalization can be optional. I’m going to introduce myself again. What I do at botco Ai, which is helping people understand what our technology does. But more than that, it’s helping them to understand that it’s not just about efficiency. AI, chat bots can change the way you do your business and over time, AI changes the expectations that people have from businesses and how customers are building relationships with each other. So I’m joined by two of my brilliant colleagues, Joe Terrell, our AVP of sales and Dawn Tong, do you want to say a quick hello, Joe.

Speaker 2 1:41
Yeah, absolutely. Hey everybody. My name is Joseph tarome and the AVP of sales been with the company since our inception, little over five years ago, with how fast things move in AI, I almost feel like a grandfather in AI, so I’m really excited for the conversation day and go over, kind of the differences between AI and what scripted AI has kind of been in the past.

Speaker 3 2:04
Don, yes. Don, I’m doing sales and business development over here at bodco. AI also close to five years came a little bit after Joseph and, yeah, long time, but it’s flown by. I am a big fan of language and languages, and so it’s just been fun to be a part of and super excited to have this conversation.

Speaker 1 2:33
Let’s look at the agenda so you know what to expect today. Here’s what we’re going to cover. How to tell the difference between real AI chat and scripted bots, a five point checklist to evaluate if your AI is really intelligent, and examples and use cases for marketing and regulated industries like pharma, insurance, verification places, innovative places where chat bots are really making A difference. Our main goal for this webinar is so that you can walk away with clear takeaways that you can apply, whether you’re evaluating new platforms or you’re simply optimizing your current chatbot and chatbot solutions. So let’s start with unpacking some common assumptions that people have when they are evaluating or buying a chatbot solution for their organization, Joe in your sales conversation, conversations, what are some of your biggest misconceptions buyers have when evaluating chatbot tech?

Speaker 2 3:38
Yeah, absolutely. There’s always a ton of different things that people kind of have this preconceived notion of what AI is. And a lot of times people think that AI is all the same, which is not always true. There’s so many different applications for AI, from note taking to recording conversations to large llms And then even, obviously, chat bots. So there’s always different things where AI can be used, and not all AI is the same. And then second, what I hear a lot is people still have this older notion that AI can’t handle complex conversations and tasks, which, if this was five years ago, that might be true, but things have come a really long way since then. Everybody’s kind of using chat, GB, team and AI and different things to handle a lot of different projects and a lot of complex situations. And AI, chat is no different. And then lastly, that it can be really time consuming to get a bot trained and deployed when it is real. AI, obviously this kind of comes from that grandfather’s chat bot where you had to pair every question and answer. Think of everything under the sun that’s no longer the truth. We can now easily train and deploy within a matter of weeks without having to bog down the team with a lot of admin time just to get something up and running that’ll do the job.

Speaker 1 4:59
Yeah, I just. Mentioned that my grandfather was part of Mensa, so

Unknown Speaker 5:04
that’s all nice. That’s why you’re so smart. Pauline,

Speaker 1 5:06
thank you. So that brings us to the core of today’s session. How do you actually evaluate an AI solution? So let’s walk through the five must haves dawn. Do you want to kick us off with the are you going to kick us off?

Speaker 3 5:24
Yes, I am talking about the knowledge base. So the knowledge base you got to have, it? Is there a source of truth? Can you form a brain of the AI? So you know different from chat GBT, because it’s not pulling from everywhere. It’s pulling from what you say that you want the AI to answer questions on. So think of what you’d want to give your new employee to be an expert on your business. You know the website, PDFs, white papers, Google Sheets, Google Documents, anything with text that will allow the AI to recognize questions and then say, Hey, this is where the answer is, and it’s not hallucinating, because, again, this is approved information that’s up to date. And as Joe was saying, it’s really easy. Once you have a source of information or repository, if you make any updates, it crawl. You can crawl that information regularly, weekly, however you want. And so when it when you have these updates in an ad, you get new questions or the same questions, it’s just pulling from that source information that’s been updated. It makes no difference to the AI. And so you’re not like Joe was saying, meticulously creating Question and Answer pairs when you update your information. And you might not even need a repository so sophisticated AIS can connect to different content management systems, you know, through APIs, for example, two on one databases, or, you know, Microsoft SharePoint. And so if you have this place where your information and your content wise, already, we can dynamically, you know, an AI chat, true. Ai chat can dynamically connect to that, and once the information is updated, nothing’s changed. This is the source information. The AI understands intent and the issues that the user is asking and pulls from the source information your content. So that’s what’s really powering that the answers that the user is getting.

Speaker 1 7:36
I’d like to add I believe intelligent chat is got the sentiments, context and intellect of a human, but it’s got the capabilities of a robot. So your best employee can’t do some very complex things. And sometimes you know what happens is like your business, maybe you have like, 10 franchises, and there’s like information, like, It’s Chinese whisper sometimes, or it’s like, okay, one your every employee can relay the information exactly the same. Sometimes information is changing, and by the time the employee learns the new information, they are giving old information. So I feel like that’s why, in some industries and some businesses, a chat bot works better, and now, as we’ll see as we move along this webinar, why we don’t have the issues like earlier about sentiment analysis and context. Because we are going to cover those in the in the five must haves. So thank you for that. Dawn.

Unknown Speaker 8:39
Yeah, definitely

Speaker 2 8:44
awesome. And then what we really want to focus on, also, with regards to knowledge base, is also prompt engineering. If you don’t have a prompt, you probably don’t have real AI in your chat bot, just because it’s everything that you need to have the bot be able to speak and talk about your business the way that you want it. It really helps you control the voice of the chat bot, because this is going to be potentially the first thing that potential customers, patients or employees say is the chat bot. And so we want it to be in the same tone, the same voice, empathetic, non judgmental, anything that you guys already do, and how you train your employees, that’s going to be super important to train your chat bot, because it needs to be talking in the same voice that your employees aren’t. Because your chat bot will be an extension of your brand and should talk the exact same way as your brand does. And so that’s what we do through the prompt engineering, is being able to understand and make sure that the bot is responding in the same way that you would want one of your employees to as well.

Speaker 1 9:49
And then you can also tell it what not to say, right? Oh, absolutely, yeah,

Speaker 2 9:54
yeah. One of the best things in the prompt engineering is being able to set guardrails so there’s a. Of things in our clientele, our regulated industries, or just any industry that you might want the bot to talk about. But then there’s also things that you don’t want the bot to talk about. So we’ve had to do things like ensure that it’s not going to act as a medical device. It’s not making recommendations, it’s not doing diagnosis. So we’re always, constantly making sure that there are the correct things that the bot can talk about, and then we put those guardrails in place to make sure it doesn’t go off the tracks or hallucinate.

Speaker 1 10:30
That’s great. Now let’s go deeper into what makes this intelligent, actually intelligence, actually usable, after you’ve done this prompt engineering, which is trainability,

Speaker 2 10:44
absolutely so training the bot is paramount. There’s always going to be a point where there’s going to be a gap in content that there’s something that we never thought of before. There’s going to be new things and new things that are coming up, like with some of our senior living clients, they never thought about pickleball. And with pickleball exploding, my goodness, we had so many different questions. Do we have courts? Where can I play? Is there anything near you guys, if you don’t have it on premises? So those are the things that will always show up in a training module, anything that we haven’t thought of before. Maybe there’s something that we want to attend to to make the bot smarter. It always just allows us to make the bot smarter and improve it, day over day, month over month, during kind of that startup period when you’re building the bot, that’s kind of the most important time for trainability. We always like to say that we won’t launch a customer before they’re 85% confident in all the conversations, and then after that, we work with them month over in that first month to get the bot up to about 90 95% confidence in all conversations. And we’re doing that easily through the training module, because we can also show where was the bot less confident on a certain answer. So we have what’s called like answer scoring so that we can see, oh, the bot was only 20% confident on this conversation. So let’s attack this one. See what was going on in the conversation, what questions were being asked. As you can see in the background, it’ll show what the question was, what the answer is, and then if we want to revise that answer, we can quickly add something in. Maybe we wanted to be a little bit funnier. Maybe we want to call out a specific topic or trend that you guys want to focus on. It allows us to have that control without always having to rely just on the AI to provide those answers,

Speaker 1 12:27
right? So it’s like it feels like, instead of, you don’t want to be stuck when you program, you just want to be evolving. So difference between a tool that keeps up with your business, rather than hold it back. Because if you didn’t have training, and then all of a sudden you have, like, a new, new product, or you have new, you know, some change in policies. You don’t want to go back to prompt engineering. You just want to kind of keep evolving and grow with the business and and makes, you know, tweaks and iterations and that. And the great thing is that it just keeps getting smarter.

Speaker 2 13:05
Yeah, absolutely. Because one of the big things is that we don’t know what we don’t know until we know it. And being able to track the conversations and have this training, like one of our clients had been seeing a lot of questions around acne. They were a massage place that was focusing on only on massages, but people were asking, Hey, I have acne. Do you have anything to treat my acne? And so they had never thought of that before, and so we didn’t have any content about it. So it continuously showed up in the training module. And after about a quarter, we were able to realize that this is really important to their clients. So from the training data and the conversational analytics, they made the business decision that, hey, maybe we should bring out a product that helps treat acne while they’re doing massage. And so it actually turned out to be one of their top three add ons to a massage, and all of that business learnings came from the training and the conversational analytics. So training is not always a bad thing. It just helps us see what we might not have thought of, so that we can attend to it and even potentially bring out a completely new service.

Speaker 1 14:11
Yeah, and as a marketing person, I’m like, You got a free survey. You can survey like all the people who come to your website just because they ask the question, and you can create new product lines because your chat bot is able to, you know, understand context and ask them those questions. And so that’s great. So since let’s, let’s move to the next just perfectly segues me into personalized experiences, and one of the quickest ways to lose a customer is to treat them like everyone else. And today, personalization isn’t just nice to have, it’s foundational. So if you see Ashley, she’s coming in from a Facebook ad and she’s pregnant, and then you know your chat bot should be able to tell. So if there’s another you know gentleman who’s coming in from maybe LinkedIn and is looking is on your page for physiotherapy, they should not land on your chatbot and have the same experience the questions that the Chatbot asked them is different. Even the referral source should guide what is the should guide the Chatbot guide the customer as to what’s the next step they should take. Right? You’re moving them down the funnel with all this context and information. So an intelligent chat bot is able to do that, and the best part is that you don’t have to do the heavy lifting to manually code every different interaction. The chat bot will adjust language, tone next step, and it’s automatically guided by the prompt engineering that you did when you first onboarded your brand. And once you do that, you know that that’s a lot of different different things that your Chatbot is doing now. So you’re not going to just launch it without checking that it’s that it’s working, okay? And that takes us to the AI playground, which Joe is very passionate about,

Speaker 2 16:12
absolutely, yeah, the AI playground is one of my favorite things, just because we don’t want AI to be this scary black box where you don’t know where this answer came from, or how a question might be responded to before you’re going live, or even after you post new content and you want to test things before pushing it to your live deployment. We can always do a lot of great testing. So what you’re seeing here on the screen is our AI playground. One up here, you’re able to put in the prompt that you’ve been testing, we can even make changes to the prompt to see how it reacts, so that we can dial in and tune that prompt to be exactly what you want, once again, without it affecting your live instance. And so here is where you can ask all of your different questions, see how the prompt pulls from the knowledge base to trigger that answer. But then, not only do we stop there, but we’ll also show you on the left hand side, what are all the sources that the AI looked at to curate this answer. So this is really important, because if, God forbid, during our testing, or at any point in time, we seeing in the chat bot that we might not be getting the exact answer that we want, or there’s a certain maybe price difference in the answer, we can come in ask that question inside the AI playground to see where that information is being pulled from, because our chat bots, and really good AI chat bots, should not be just making answers up. They should only be pulling from the approved content that obviously, you uploaded in your knowledge base. So we can identify where was that information coming from, that might be old or hasn’t been updated, and we can identify that make that change and update it. And so a real life example was from one of my clients. We were getting wrong pricing during our testing phase when we were building a chat bot. And so we came into the AI playground, we put the question in, and we actually identified eight web pages on their site where there was old information, and that’s where it was pulling from. So without the AI playground, it would have been really difficult for them to even know what web pages it was being pulled from. But since we’ll show the source document, whether it’s a document, a web page, or any source that we’ve uploaded into the knowledge base, we can easily, quickly identify exactly where it is, make those updates, either re crawl the website or web pages, or upload a new version of that document, so that we always have the most up to date information

Speaker 1 18:37
I see. And can you do you always have to upload stuff? Or can it link directly to their it can, because I know that some, some, some companies have, like, 1000s of white papers and pages like that. So are you able to directly get to all of those 1000s of, you know, manuals and instructions and or do you have to upload each topic?

Speaker 2 19:04
So for our clients that are larger and have those 1000s of documents, we can always connect to their source of truth, whether it’s a database, a domain knowledge base, or even a document repository, like Don was kind of talking about, we’ve done a lot of different things so that we connect through an API so they don’t have to actually upload it into our knowledge base. We can just look inside of theirs. And so like Don was saying, anytime that they make an update in their system, since we’re doing an API call, we’ll dynamically update that so that they don’t have to duplicate work by updating it in their system, then updating it ours, it’ll automatically take effect in the chat bot anytime that they make a change in their system.

Speaker 1 19:48
That’s great. And also the AI playground helps there, because it’s through it can scan in seconds, or however faster, all those 1000s of. Documents and manuals and tutorials. And can you imagine trying to do that

Speaker 2 20:03
manually? Yeah, that’s why the painstaking part of old times when you had think of what’s every question and answer pair really was not allowed to do anything complex, because one you had to think of what’s every single question anybody could ask. Now when you build a knowledge base, as long as you have the documents and enough information from your website, PDFs, Word documents, pretty much anything, it’s able to pull that automatically. And then also, once again, if it’s a little bit out of context, or you want to make the answer shorter or longer, you can always do that in the prompt to ensure that you’re getting the exact responses that you want, which always can be tested in the AI playground.

Speaker 1 20:43
All kind of fits together. So once you’ve tested your AI, so you tested it, it’s live. The main question is, how is it performed? Dawn, can you help us? Take us through what AI insights does? And you know, if there’s more, what’s going on here?

Speaker 3 21:00
Yes, indeed. And so the data insights super important if you’re trying to have evidence behind you know you’re improving the user experience or improving your business. And so this is where the AI chat conversations are a gold mine of information. So you’re getting, in real time, what the customer cares about, what they’re asking about, and then you can actually track and record and save these conversations if you have a sophisticated AI chat that has a HIPAA environment and is allowed to record these conversations. And so, you know, there’s some term called attributes, and in the conversation, there’s many different, infinite amount of attributes, keywords, topics, sub topics for health care in our you know, an enterprise level, you know, you might be wondering, if they’re asking about inpatient or outpatient services for behavioral health, are They asking about spravato or TMS, all of these things. You can tell that, you know, chat bot to start tracking, and you can have these custom metrics that are relevant for your business. So obviously, these engagement metrics are awesome, and you know the kind of cookie cutter metrics, but you really want to know what’s important for your business. And you know your business, and you can actually prompt and tell the AI to record. You know, you don’t even need Predefined Fields. It’s just in the conversation, and you just are able to pick out the things that you want, that you want to track. So think of it as like a Google Analytics for your chat bot, just very able to give you a clear view of what’s going on, what people are talking about, and what they’re not talking about. So maybe you’re running a campaign or a service and you know people aren’t asking about it. That’s an issue, that’s that’s something that you didn’t know, and now you know that people aren’t asking about it. So it really allows you to pivot and make really well informed decisions when you’re trying to change something about your business or the user experience. And so kind of you know, to sum it up in here, this is a very top level, like heat map, of the things that could come up in a conversation, but you can actually dig a lot more granularly. So what are the sub topics? What are the keywords? What’s the actual conversation? So you can actually see the conversation, see how it went. You can track sentiment. So the possibilities are endless and long story short, you just get a clearer picture and a clear pulse of your audience, which is valuable for any business.

Speaker 1 23:45
That’s awesome. And like we say in marketing, you should be doing your social listening. I would call this. You should be doing organic listening. People are coming to your website and they’re talking, and you should be listening to what they’re saying.

Speaker 2 23:59
Yeah, 100% because, yeah, even Paula V that’s a great point. Because for marketers and all the people like when you’re able to hear the voice and the actual words that your customers patients are using, not only can you then put that back into SEO, but also we can track what topics are important to your constituents in the community. So whether it’s new blog posts, different kind of campaigns around certain topics that now we’ve heard are really important to the community as well as it’s almost like the AI is taking a tally every single time it hears something. So as you can see, like on product sustainability quiz, like, that’s 512 people that were interested in that you get to see, like Don said, what’s important. But then it’s also like Don said, as important to see what you’re not seeing. Because if you are running a campaign, you put out a new product, a service line, and no one’s talking about it, you know that your messaging might be off, and so it gives you that feedback instantly, without having to wait. A survey or anything like that that might not even get filled out, you’ll always have that information right here at your fingertips.

Speaker 1 25:09
Yes, and that when you said blog, that reminds me with our own chat bot, right? We have a chat bot who’s called what Queen app. And I asked our engineer to you know, if someone says certain words say, Senior Living. We have, we have a bunch of clients there. Then can the chat bot give, I gave him some blogs that I would like to be them to be offered based on, like, where they were in their journey. Because sometimes, before asking people to consult with you, or, you know, asking them to meet with you, to see how you can help their business. Maybe they want to learn a little bit more. So it’s nice to have a chat bot to be able to offer these resources when it’s appropriate, when they’ve asked you a question about it and they’re interested to look for the information, then they’re not going to jump off and go to my blog and then start trying to hunt for that particular word, because they don’t know it exists on the blog page, but the chat bot knows that it exists, that pickleball blog is there, and the question that they’re answering maybe they want to know more about it than the I’m giving them a small answer, but maybe they want to learn more. So that’s helpful, too, for for marketing, 100% we are here at the exciting part to just see a couple of use cases as to how industries, and also some regulated industries, are using chat bots in innovative ways where they’re making a huge impact.

Speaker 2 26:36
Joey, absolutely. Yeah, one of those is definitely in the farm space that we’ve been working with, which is a highly regulated industry, anything around these adverse events, when somebody says that they have a rash or they had a reaction, these are very complex conversations that need to one be handled with care and be able to be routed correctly so one we can answer them, get them the information that they Need and alert their their healthcare provider of what’s going on. But also the AI is always looking for this, because in this situation, adverse event detection always needs to be reported. And so we’ll be having the AI always look at these conversations and be listening to anytime that we have this, because once this happens, we actually need to report it, and we’re able to automatically trigger the back end automation with emails as well as scoring. So we can also then understand how advert how I’d say, what’s the level of acuity of this adverse event, and depending on the level of score, we can treat them differently whether we need to automatically report them to the FDA based upon if it’s over an eight and above, or if it’s below that, we can always then share it with the team. And if it’s an adverse event, maybe it’s something’s wrong with the product and it’s just effective, or maybe the box or the product had a misprint or was just damaged in shipping. Any of those things can be handled automatically without having to always rely on an agent, because some things can happen at two in the morning and you might not have somebody there. So having this AI be able to trigger this automatically for the team has saved them a bunch of time and also helped them be staying within their compliance rules.

Speaker 1 28:18
And I think that’s a great nod to generative AI and to intelligent chat bots, because pharma used to be sort of the lead adopters, and they only used to use very question answer type chat bots. But now they’re starting to use chat bots for they’re gaining confidence because of the guardrails that are in place. So that’s great, Dawn. Let’s look at the next case.

Speaker 3 28:44
Yep. So verification of insurance or benefits, so obviously, at an enterprise level in healthcare, you know, Pallavi, like you were saying, you’re able to ask and get questions. You’re not handed over to someone, and eventually you get that value and you’re willing to give value. And so, you know, obviously, how am I going to pay for this is a big question, and you know, reimbursements, obviously for the provider are super important these days as well. And so this is a great opportunity to be able to move along with this chore that you have to do in healthcare, it can be a little overwhelming. And just being able to do something, at least anything, in this moment, you know, where the user is engaged, and then be able to move it forward is huge, especially this tedious task. So for on the user side, you know, being able to, because you have a HIPAA environment, you can collect personal health information, we can safely capture it and handle it, and then, you know, make an API call to like a clearing house or database, and just quickly say. Hey, you’re a network. Here’s maybe your copay or deductible. And just make them feel at ease. Peace of mind that, hey, this is, this is a place that can help me out. I’m I’m probably not going to look anywhere else this. This is, I’m good to go. And so that’s, that’s really great for the user. And then for the admissions team, you have a lot more information to work off before you talk to the person. So again, handling the personal identifying information in the chat, which is safely encrypted at rest and then encrypted also in transit to like your EHR or your CRM. And then now the admissions team has the information. They’re able to do some, you know, legwork behind the scenes before they actually hop onto a call with a referral. And so now it’s a lot smoother conversation, and when you talk to that person, you’re not getting interrupted doing like, email and phone tag, like I have to get back to you and figure this out, or whatever. That’s just the momentum killer. And so, you know, just being able again to have that smooth conversation and give you know, time back to the admissions team to be able to delve into the ins and outs of the treatment and what the user the referral is going through. That’s what you need the time for, and time is of the essence. And so just being able to take off these tedious tasks that can eat into that time is huge, and it’s what the admissions reps and the user would rather be doing is like talking about, you know, what the treatment looks like and and how that’s going to go.

Speaker 1 31:46
Yeah, as a consumer right now, if I was on that chat bot, it says, Okay, your network. And then next box that I see is book an appointment. And I’m able to book an appointment just like 4pm because it’s connected. So it knows, like, if I’m trying to look for an ENT appointment for my nose, it knows that this doctor is free at 430 and it even knows, like our location analysis can let them know, like, you know which location I’m in. So if you have like, 10 different specialty clinics, like, look at how I’m trying to navigate their website, trying to be like, Okay, where do I go and which doctor has this appointment? And here I’m done in two clicks. So thank you, generative. AI

Unknown Speaker 32:33
big time scheduling appointments is a pain. Sorry, yes,

Speaker 1 32:41
yeah, yeah, I had resemblance over the whole I had to do journaling about about the whole system. So I’m very glad about this. Let’s you know, we covered a lot today, so let’s quickly recap what we’ve spoken about. First, AI, chat isn’t just about automating responses when done right. It’s about delivering at least a 98% accurate answers within weeks of deployment. Second you wanted to integrate with the tools that you already use. It could be HubSpot, Zendesk, care, logic, whichever ones you use, because that means faster adoption and less friction and better outcomes. Third, it should be built on a no code platform, so that your team can deploy, they can train and they can analyze, and you don’t need a developer on standby to make these changes. And finally, your chat bot should be HIPAA, validated if you’re in a healthcare environment industry SOC, two compliant and basically built with enterprise security from day one. Yep, that’s our those are key takeaways for you. If anything today sparked your interest, or if you’re wondering or and I forgot to say, if anyone had any questions, they could have put in the tab and we would answer that at the beginning. You’re free to do that now too. We have a couple of minutes, but if not, please just use your cell phone. Schedule a free consultation, and we’re happy to just chat about if you want to learn more about this technology and understand, like any additional questions that we didn’t cover today, I’d like to give the mic one more time to Joe and Dawn if you have any any last words,

Speaker 2 34:33
yeah, no. I mean, I really appreciate everyone’s time if you have any questions, even if it’s just about AI in general, it doesn’t even have to be chat bots. I’m here to be a wealth of knowledge, so doesn’t have to be about chat bots. But obviously, you know, I’m very passionate about those

Speaker 3 34:49
Indeed, indeed, don’t settle. Don’t settle for, you know, not a good chat bot, not a great chat bot. Your business deserves more. Make your life easier.

Speaker 1 35:02
Absolutely, I love that don’t settle. Thank you, everyone so much for being here today. We hope this session helped to clarify what real AI look like. And we’ll be sending out a recording for everyone who’s registered so you can share it with your team members and on behalf of all of us at bot co AI, stay connected, and we hope to talk to you soon. Bye, bye. I am so okay. I said, bye, bye. And some people may have left, but I just saw that we do have three questions. So anyone who’s asked those questions, if you can stay on a minute, and we’ll try to get to them. There is someone who asks, Are there use cases when you can use a scripted bot?

Speaker 2 35:54
I mean, absolutely, we do a hybrid approach, so we don’t always only rely on AI. There’s certain things where, if you want certain, what we call blocks, to be triggered and have the wording the exact same way every single time with no variation. A lot of times we do that for pricing, insurance information or just overall kind of legal language that has to be set a certain way. We definitely will use that as well in a combination with our uh, AI answers as well.

Speaker 1 36:26
That’s yeah, that that’s great. So you can have some peace of mind to be like, I want the chat bot to say exactly this about about certain topics. Um, that’s great. We have another question. Are you partnered with any PSO patient safety organizations.

Speaker 2 36:43
Oh, that’s a good one. I mean, there’s always a part of patient safety that we take into account when we’re working with behavioral health, mental health and addiction, even home health organizations as well, with regards to just a specific PSO only, we haven’t run across one as of yet, it definitely could be up our alley, especially if you need to kind of focus on follow ups and understanding, because we’re even working on a project right now where we’re sending daily messages out to patients in a clinical study so that we’re always checking in with them, seeing what their side effects are. Are they doing okay? What help will they might need, and if we need to escalate that to a certain HCP that they’re working with? So with regards to a specific PSO, we have not but it’s definitely kind of in our wheelhouse of things that we work with on a daily basis.

Speaker 1 37:38
Thank you, Joe. And the last question is, what is the cost?

Speaker 2 37:42
Oh, good question. So we’re built on a SaaS model. So our pricing is built on a monthly subscription and then a one time integration fee depending on what systems we need to integrate to. And this monthly SaaS being, the pricing is always kind of bespoke, because we will give a flat rate depending on a few different outliers. One is, how many different channels are we going to be on? Because we can be on web chat, SMS, social, WeChat, WhatsApp, Slack. I mean, we can be anywhere where people like to have conversations, and we always like to meet the community, patients or customers where they are. So that will help determine that as well as then, how many unique visitors on the website, SMS or social will we be interacting with on a monthly basis? Because we give you guys a flat monthly fee so that you get unlimited messages on that tier, so that you have really good control over what the budget is, and you won’t be worried about fluctuating pricing or being nickeled and dimed for every conversation or message that goes out.

Speaker 1 38:51
I think that was an in depth answer. Usually people just kind of dismiss that question. Sounds like you said that before, Joe,

Unknown Speaker 38:59
uh, only once or twice every, every meeting,

Speaker 1 39:01
yeah, well, I’m glad I was able to catch those questions before we go. And thank you so much for everyone who stuck around. And finally, have a great day again. Thank you. Thanks everybody. Hello and thank you, Dawn,

Unknown Speaker 39:16
yup, bye. Have a good everyone. Bye.