The player is loading ...
New episodes every Monday. May we send you occasional email updates?
Explore the future of pharmacy communication with Brennen Hodge, VP of Product at Voxo. Discover AI-powered IVAs revolutionizing patient interactions, automating refills, and boosting efficiency. Unpack insights on technology, customer experience, and the evolving pharmacy landscape. Don’t miss this candid, forward-thinking episode of The Business of Pharmacy Podcast.
This transcript was generated automatically. Its accuracy may vary.
Mike: introduce yourself to our listeners.
Brennen: My name is Brennan Hodge. I'm the VP of product at Voxo, and Voxo is a telecommunications company. We [00:01:00] build business communication solutions for pharmacies and the patients, and also just across all business industries. We roll all that into a one unified platform. unified communications of the services, acronym UCAS, not too many people use that, but that is the industry we're in,
Mike: , Brendan, it seems like the IVR press one, two,
and three is almost outdated with
AI.
So I know you guys are doing some stuff with AI.
What does that look like? Run through one of
those AI conversations
For people that are maybe more used to it, push the button kind of thing.
Brennen: Sure. Yeah. So an IVR stands for interactive voice response. Been around 50, 60 years. Everybody understands you call a phone number, press one, press two, press three, you go to different options. The evolution of the IVR is going to an IVA, interactive virtual agent. And this is where you call a phone number and an AI, a voice AI answers the phone. Now I like to look at this as. A virtual [00:02:00] receptionist. Imagine you're hiring a college person and they're just sitting there answering the phone. They answer, they know everything about your pharmacy. For example, they know the hours operation, the services you offer there, um, your, holiday hours.
All your, the protocols, everything that a person would answer. And then taking a step further, if you do have a prescription you need refilled, you can just tell the AI, Hey, I need a prescription refill. Okay. And go through the same process. Just speak your prescription number, verify your date of birth. You might have to verify some other things and there You go. You have it, you have your prescription refilled, versus the old way with an IVR where you have to type in your prescription number, , you might lack that dexterity and you might fat finger a number, you start all over again, it's frustrating, with an AI, you're just talking like it's a human,
you're like, okay, my prescription number is 987, oh, I'm sorry, I didn't mean 987, I mean, 789, and you can just start over, it's just natural conversation.
Flows a whole lot [00:03:00] better. people can talk. Like a lot of people have a hard time looking at the phone, trying to find their glasses. It's, it's, it can be a difficult thing and moving forward. So with this, the IVR is very limited. Now you can request a refill, talk to a pharmacist. That's about it.
So moving IVA. You have this AI that is connected to your pharmacy management system. So patients can call in request refills. Patients or other pharmacies doctor's offices, they can call in for inventory checks. Do you have lisinopril in stock? Whatever that looks like. Query the inventory.
Check on prices. Patients and check on their, Their bill, their, their prescription, just like you have a human
sitting there, reading off the screen. Imagine now
you just have the screen talking to the
patient. So that's how I kind of like to think about the
IVA and it opens up the door for anything that a human can do sitting in front of that computer. Our [00:04:00] IVA is capable of doing so. It's really interesting stuff. And this is just inbound. We're not even talking about outbound.
Mike: I'm a big chat GPT user. Let me throw some things at you. I'm trying to think through some of the things let's say with chat
GPT, it gets a little bit verbose
sometimes, know, it's giving
me four
paragraphs and like, Hey, I want one paragraph and I
want
bullet points kind of thing.
Are there variables there where someone can say, Hey, slow that down? Or what do you mean by such and such?
is it currently, to that status of that much
of a conversation? Or is it just giving you
data in a friendly way.
Interactive form.
Brennen: You can customize this down to as much as you want. For instance, like you can have your languages, your, the tone. Like I like the, uh, some from South Mississippi. I like them, uh, so it's called Southern Magnolia. This is a voice, southern accent, female southern accent. Sounds great. [00:05:00] You can have a northern accent, California, and 32 different languages. If somebody, if a patient comes in speaking Russian, you can convert talking to Russian. Um, Spanish, same way. So imagine you hire somebody and say, all right, this is how you're going to answer the phone. you're going to be friendly, but you're going to stay right on point. If somebody wants to ask you
about your favorite pizza, say no, like I can help you out with
pharmacy stuff. So you can instruct this AI, just like, you
can
a human, and you can say like, these are your
constraints. You have freedom to go in here, maybe
tell a joke. But you're staying in these
constraints. you cannot talk about football, cannot talk about baseball, just pharmacy stuff.
Mike: Or on that point, and
you could say, it's okay to talk about the weather and it's okay to talk about local happenings, like,
there's a parade downtown or
whatever. you can set these parameters up, how much information you want to share
and all that kind of stuff.
Brennen: for sure. And you can even say like, um, try to keep your conversations to
three minutes or less. So once you get around that two minute point and be like, all right, they'll start wrapping it up. So, all right. So, really help you out with today? [00:06:00] So, Ms. Smith might want to talk, she might want to sit there and talk
for five minutes about her dog. There is an opportunity where you can build that into the patient experience. It's a new type of experience now because some people just want to be heard. They just want to have a conversation with somebody. Well now, what if you have an AI agent that's there? What if you say, okay, Ms. Smith, if you can tell in her tone, her just conversations that she really seems lonely.
She really needs to talk to somebody. Talks might extend it out to 10 minutes. Start asking deeper questions, just
like a medication therapy manager. Start asking about her medications. Go ask her, how about her,
how
she's feeling. There's a lot of insight that you can gain from that. and it doesn't have to be cut and dry where, Miss Smith calls, half a
refill prescription, 60 seconds or less.
You can take it down multiple pathways. And that's the thing with these AI agents: it's flexible. Like you can kind of learn on the fly and figure out
where to go.
Mike: Do the customers know? They're talking to an
AI
Brennen: It's a [00:07:00] really good question, and we're testing it now and it sounds really good. And the thing is, I would be willing to bet 85 to 90
percent of the people would not know it's an AI,
but also what
What difference does it make? If it is an AI, we talked
to Siri, we talked to Alexa, we talked
like we
were already growing up here with, with AI voice, they add. So is, it a bad thing to say, all right, this is, This
is, Mary. This is Mary's Pharmacy's new AI.
assistant. How can I help you
today?
Mike: I'd rather AI, I tell you, I would just rather AI. I can't think of anything. I really can't think of anything that I'd rather talk to a human instead of AI. even things like psychology and that, I mean. Psychologists are basically, I think they're trained. They're not supposed to just give you a bunch of advice. They're supposed to listen to what you're saying and, parse it out and give it back to you in a different way and restructure and stuff.
One of my daughters Is [00:08:00] in the counseling profession and I know it doesn't replace that because someone who lives can look at you and say, look, you're telling me that, but I think you're fibbing. There's stuff that humans are for, but when I'm calling a pharmacy or a store or somewhere like that.
I don't care. They're not gonna be diagnosing me, as a nutcase. They're just gonna be talking to me. across the board, I
I can't think of any place. I'd rather talk to a human
than
an ai.
Brennen: And I would say AI is gonna
be patient with you. I'm gonna have a very happy disposition.
We're out of the human, they might wake up on the wrong side of bed. They don't want to talk to
Mike: Wait a minute, Brandon, you're telling me that if Mrs. Smith calls the ai, it might even take the time
to say, well, hello Mrs. Smith. It might be more
patient than I am.
Brennen: look, I just saw a Reddit post and it was a
call
center agent. And he's like, yeah, we just installed a new, AI
assistant [00:09:00] in our call center. We're training it right now. He said, " I'm happier than everybody that works here. He's like two years old. Every call center is going to be obsolete. 300 billion industry. He's like, this is going to replace everybody cause it's always going to be happy. It's
never going to get frustrated with you. It's going to answer your questions
better than they can. Like it's, it's inevitable.
It was When chat GPT, when , the conversational, Part of came out. So I'm walking around with my AirPods in just, I was testing it out, just talking through it for about 10 minutes. And then, I thought the conversation ended and I'm just walking around for another 10 minutes. And, we got a cool lake dam over here.
I was walking up and down it
and I was frustrated about something. I was like,
oh, what am I gonna do? And then she was listening in the
whole time. She said, it sounds like you're frustrated, you wanna talk about it. And I was just like,
what? And then I thought, okay, yeah, let's talk. And then I just started dumping her and it's like, let's just work, do it, act like my psychologist.
it was very [00:10:00] weird to do that, but at the same time I was telling this AI things that I wouldn't tell anybody
otherwise, it was a different dynamic. And I know people, actually, I have a friend that he started an AI, uh, counseling.
It was a virtual reality AI counselor, which is way too early. It was with children and they were actually talking to a virtual reality rabbit. They were talking to this rabbit about everything that they would not tell their parents. They would
not tell psychologists, adults, but they would tell that rabbit everything.
And they couldn't find a business model for that, but that was a little too early. but these things, like you start to see glimpses of where the future is going to take us and, just, Going back to the pharmacy world, like I was, we have one pharmacy. We just installed it the other day and I gave him a demo of our IVA. He said, Brennan, he said, I was one of the pioneers with the, the, the robot pill counter, 40, 40, 50 years ago, he said, when that happened, he said, [00:11:00] I realized that I was never going to count another pill again. I saw that future. He said, when I listened to this and talked to this IVA, he said, I realized that we're never going to answer the phone again at a pharmacy. Because all those, like,
just these manual things that don't need to be,
Taken up by a pharmacist's
time,
like, just let technology automate that. I mean, how much more can you do at a pharmacy if you don't
have to answer the phone to
tell you, Hey, yeah, we close at 5 p. m. a day?
People, I think, are scared of AI because, for various reasons, I think it's the, you
know, the Antichrist or, it's coming to take our jobs. I mean, you can go different directions there, but I think a lot of people are scared because they feel like it's going to take their job. Like, if it can answer the
phone for me and that's all I'll do is answer the phone, then what am I good for? And I think that's the wrong mentality to take because you look at a, uh, it's going to replace some task.
But there's also some tasks that only humans can do. That
face to face talk with Ms.
Smith, like, right [00:12:00] across the counter. Like, an AI
I can't do that yet. But there's a lot of tasks that you can turn over to an AI because it is technology and it can do it
efficiently. So focus on what you can do specifically as a
human and expand
on that. Direct patient care. It's a whole
lot more important than answering the
phone and, refilling prescription or checking on inventory.
Let a computer do that. in my town, a month ago, and I've got one across the
Mike: street, like seven or
eight Burger King's closed down in our city, just like that overnight, seven or eight of them. I would imagine it's nothing more than just can't find enough help. And so, especially in the last couple of years,
it's hard to find help, competent help, and even help just to answer
the phone
or do those kinds of things.
Brennen: You do think, I mean, I will challenge listeners to say, okay, I can answer the phone all the time and I can do this myself. But in reality, across the board, every industry, it's
roughly 20 [00:13:00] percent of all phone calls are missed. Like that, that goes across from auto to banking to pharmacy,
If you average it out, roughly 20 out of 100 phone calls are going to be missed. Somebody is going to ring. It's not going to get the right person. They hang up. There's a problem. You lose that customer. they go on somewhere else or they just get frustrated. if you can answer that a hundred percent of the time, minimize their wait time by 90 percent of the time, actually get your patients or your customers the most efficient path towards whatever they're calling about. that is elevating the patient experience. That's elevating the customer experience. It's getting your customers what they want. They don't want to sit on hold. And like we talked earlier, that's being respectful of the caller.
You got to think about what the caller is trying to do, not what you think you want for the caller, what the caller is
actually wanting. And most of the time that is just to get something done as fast as possible.
Mike: It's
kind of like owning an ice cream [00:14:00] shop. And I know when I'm going up to my cottage and dairy bears there are too many people standing out there.
I ain't stopping at dairy bear. I'm going to go right by it. And so the owners can say, yeah, we did fine. There wasn't much of a wait today. People got served. We did this or that, but they forgot about Mike who went by
because he didn't want to stand too
long
out there. The same thing on the phone, we're doing okay.
We're just, and that's like, yeah, But what about the people that either got a busy
signal or they
hung up after so long, or they were in
the IVR and they got. My mother said that her children aren't Picked about something and
they just didn't call back. So there's a lot of things
that are lost like
You talk about,
friendly and things like that.My dad, back when I was a kid at the pharmacy, he had this, pharmacist. And nowadays you hear some of it, like you'll call a plumbing place.
They'll say, hello, it's a great day at sunshine plumbing. How can I brighten yours? That kind of baloney. Well, my dad used to have a guy and. He was a nice enough guy, but he would answer the phone.
I won't say a phrase. [00:15:00] He would answer the phone. Like, Bob's
drugs, like you
interrupted me. I'm trying to listen to the tigers game,
that kind
of things. And so you're just going to avoid all of
that. Or if someone's having a bad day or the customer before him pissed them off about something.
It prevents all of
Brennen: Completely. It's starting at every single phone call, starts
over. Almost like you're their best friend they
I haven't talked to in six months. You call them up, and they're there to help you. You have a very helpful
assistant that's going to be there for you. It's never going to be tiring from a long night out.
You know, Didn't
sleep
good, just broke up with their girlfriend or boyfriend. They're going to have the same disposition and it's always happy. It's always cheerful. It's always helpful. That's the thing. I keep going back to the caller's experience. Just imagine if you could design the best experience in a conversation, in a
phone call, how would that
look?
You can do that now
with AI.
Mike: had a customer just this morning. One of my main guys came to me and said, Hey, this customer already chewed out so and so and chewed out so and so, our two [00:16:00] technicians. He goes, what do I do? I said, well, just talk to her. She gets ready and you just hang up on her. and if she's on the phone
longer, tell her you
don't like the way that she talked to
so and so about that that kind of stuff,
I imagine AI can
even do better at talking some of
that
down.
I imagine they pick up on the
flow and the speed of the words and the volume and,
that kind of stuff too.
Brennen: Oh yeah, completely. Sentiment analysis. That's one thing that we're working on. We have a few large call center customers. and so the way one of these customers operates is they have a quality assurance manager there. Their job is to listen to one rep a
day and listen to one conversation. They go down an Excel spreadsheet.
It's like 15 metrics they rate them on there and that's their score for the day. One conversation, they have 100, 150 agents. We're automating the whole entire job role. We listen to every single phone [00:17:00] call, do a sentiment analysis. We actually got a graph, a trend line, if the conversation is positive, it's going up. If, the caller is Dropping F bombs, cussing, bad, angry, it's going down, gets a bad angry face, and that QA manager or any manager at a call center, they can have a dashboard like, alright, these three phone calls over here are just falling off the charts. we have to intervene. Let's say there's frustration here. How can I help? or coach. A lot of times they'll coach the call center agent. Hey, give them 50 percent off, something like that. a lot of things like that, it's already in play and it's helping behind the scenes, it's helping coaching them along. and also it's going back to that patient experience, that caller experience. if the caller is frustrated and that particular agent just isn't qualified to
help,
then step in and help. AI as a tool. It's listening to the conversation
Mike: the AI is listening to a live agent
and then offering some help
Brennen: yeah. So imagine Mike, you're a call, a manager of a call center and You have, you know, [00:18:00] 200 agents there. You have a dashboard and you see, you know, 85 percent of all the
conversations going good, but there's two conversations over here, the customers hollering and screaming, and you know that because AI is telling you, okay, they've, they've said the F word 18 times, they're hollering, their tone, everything's
escalated.
Mike: And it's a
human
talking to them, but the
AI is listening in for some of this stuff and
then waving the white
flag for this leader to, to maybe have to
step in
or add some support, things like that.
Brennen: That's exactly right. So we're doing that now. and also,
so having coaching in terms of, let's just say, uh, trying to give an example here, let's just say it's a call center
for tires and this one customer calls them like, Hey, I need some, some tires on
this front end loader. Well, that rep does not know that, but if that, the caller says it's, uh, this model number, we'll extract that model number, return all the possible tires that they might go look for. I mean, you're [00:19:00] you're skipping a step where that agent has to type it in and you're displaying all that stuff.
So we're extracting these keywords and displaying them on the screen. So you're helping agents. It's a helpful tool for the agent. But that's, that's like step one. Next step is just to
eliminate that
agent altogether,
Mike: One of my
guys, he came up with this and I use it now, but when people call us at the pharmacy, a business, an insurance company, or, PBM or something like that, they'll say, hello , this is Susie calling you on a recorded line. And my guy says, I don't talk on recorded lines and they get kind of flustered.
They say, well, I'm calling you and this is how we do it. He says, no, if you want to, you can call back on a non recorded line. And we were talking and he's like, nobody's going to call me up at the
pharmacy and
say, Hey Mike, we were listening to this conversation the other day that you had with Insurance rep.
Boy, you were polite and a nice guy. And we're thinking
of sending you a gift certificate for, Blimpies
or something, because, they're [00:20:00] not
listening for that. They're listening to spy on you and, the stuff we're talking
about. It's like, screw it.
Brennen: yeah. You, you know, one reason they do that and, I'll give you an example. You might say, you know, I
want to
do X, like I'm willing to pay or Y. Well, you know, two months later, you get a bill in, it's not
what you thought you said.
Well, they can actually go back to those recordings, those
transcripts, like, actually, Mike, you did say this. or as a company, oh wow, we uh, we
we messed that up, so
we'll eat that. That happens a lot in the logistics industry. There's uh, freight brokers, like, I know we agreed to this, but no,
that's not, not what happened.
Mike: I
suppose, it's not, did
you, did you say have a nice day correctly to somebody? It's those big things. It's like, did you order a ship full of this or two ships full of that?
So Brennan,
All right, so I'm giving all my secrets away, but worse than not wanting to talk on the phone is [00:21:00] someone asking me to do something. So then I'm like, well, here's my technician or something like that. I'm not going to do anything if I don't have to. So does
Will AI jump into that ever? Is there a time
where, AI is having this conversation
with this customer and then
tasks come out of that or, you got to
call
your doctor and we're going to call the wholesaler and this and that kind of stuff is AI getting to, to that step?
Brennen: completely that's the AI agents that I was referring to. So one thing that we'll have coming here in Q1 of
next year is appointments. So let's say flu shots, calling in. I think I need to schedule a flu shot. Like you don't really need to talk to a human to do that So with our IVAs, you sit there and be like, all right, I'm available Thursdays.
What do you have?
IVA is hooked up to your Google calendar or Outlook and says, all right, we have a 1pm, 2pm, 3pm available. What would you like? I'd like 3pm. Okay, great. Here's my email [00:22:00] address. This is like a magic moment. The first time I did this, it was with my own personal assistant and I did that.
And I told him like, Thursday at 3 PM, it popped up on my calendar. I got a confirmation email. It was like magic. And that was just like the low hanging fruit is scheduling appointments. The next one is, calling this, calling the PBM, calling another doctor's office, calling another pharmacy, seeing if they got this medication in stock. So like I did
see one project, it was on Reddit.
I was looking, researching voice AI. And, he or she was saying that, yeah, I got a weird phone call today. It was from a doctor's office. I'm pretty sure it was AI and they were calling on inventory checking inventory.
And as I did a little more research and I found which company that was, yes, you have doctor's offices call it
using AI. To call pharmacies to check for the inventory. Now, when we release our
IVA, you're gonna have an AI talking to another AI
about inventory.
Mike: Now you have that too, a ton of the calls we get are looking for a [00:23:00] certain C2 as the other places don't have. And then, I always ask, are they local and this and that you kind of weed things out. It's going to be able to, to do all that, and then you get some of these, Older people, I guess, in my category that use the old terms, like, well, I can do it at a quarter of three, or half past this and that.
I said that to my kids sometime. They're like, what is that? It's like, you don't know what quarter after means and I guess you might not, if you're not looking at an arm clock or something like that. so there's some of those. Old phrases. In fact, have you ever heard this, but I think you're too young
for this. My aunt, I used to say, what's the phone number to something and something, and they would say, Empire three, four, five, six, seven. When people had to remember
phone numbers more,
because you didn't have a cell phone and maybe not a phone book
around. So Empire would be, I don't have a phone in front of me, but like,
E is three and M is six or something. And so instead of [00:24:00] remembering three, six, this, this, and
this, you would just say empire, that would be the first
two numbers. And then
you would
only memorize the
last five numbers, but they would always throw these things out like Glendale, like what the hell is
Glendale?
Well, you'd go to the thing and G would be four and L would be five. And so that was
a phone number, Those kinds of things are gone, but you might get some old fart on the phone that
says,
What's your number?
It's like, well, it's Glendale two, four, five, six, seven. It's
so that is just a prompting thing. Just like you're hiring somebody. maybe it's their second language. I like to treat this, this IVA, any AI agents, like you're hiring a beginner that has PhD level intelligence, but this is their second language. you have to be very explicit in terms of language and how to use this. For instance, early on with LLMs and voice, if you would say like 601, 7958, like 601, it's 601. We always say 601. Is it going to put an O or put a zero? they did put it out. But I started learning that. Okay. When [00:25:00] somebody says in that context, it's usually a zero. Same thing with a quarter past five, half past five, whatever that looks like, that's one thing that is training.
Brennen: Once you say, okay, if, if you say that and you get it wrong, next time you hear that, respond to it like this. So I had to do that a lot with, with addresses. And we say like 345 Oak tree lane, like with an AI, I would say 3, 4, 5, Oak tree lane, 7, 5, 0, 3, 4. Be a little more fluid with that and then start getting it right.
So you really have to just tell it exactly how to respond to things and start in contact. I kind of going back
to,
This is one of our conversations earlier, like having the ability to customize that. So that's what I'm looking at now. Like all my little bitty instructions, like it says,
If they don't have a prescription number, tell them they can look up prescriptions.
Um, let me see. Keep all your responses short and simple. Use casual language. Like, you can use phrases like, um, and well, and I mean. Add in some human languages, have some pauses. And like, any time you [00:26:00] respond with a number like 9 1 1, say 9 1 1 versus 9 11.
So it's like little things like that, you just kind of take for
granted as humans, as our language, but with AI, with voice AI, you have to explicitly say that. It's all getting better.
Let me give a little context on what we're building.
Because it's like a black box for a lot of people. So for this to properly work, you have our voice Right.
So we're speaking. You have to convert that speech, what it's called, speech to
text. So I like transcribing. You've seen it. You transcribe your voice into words. You take those words, the text, you run that through a large language model. like ChatGPT. It runs on a large language model. It basically predicts what the next letter or word should be. And then it processes it through the LLM and outputs the text, which is the answer. And then we take a text to speech engine to bring the voice back. So looping that all together, you talk, turn into text, transcribe that, [00:27:00] processes it, and then puts it back into speech. And all that happens 800 milliseconds.
So. In the past, less than a year ago,, it happened so fast, it was about, seven months ago, it was over a second, and over a second, it had too much latency where you knew it was an AI, but now it's, getting down to a really with what we built.
So we were leveraging various other large language models, open AI. We were using 11 labs, deep ground. We're using these other large companies.
The latency was too long and it was very expensive. So
We started building this in house on our own servers where we can get that latency down to, I mean, a hundred milliseconds, which is, you can't even tell, like, it's just, it's just like that. And when you're doing that.
It's just a natural conversation. Like when I was testing this stuff
out, I mean, four or five years ago, it was like talking to a robot. Like you knew you were talking to a robot. It was very delayed. you would say something, then it would continue on for three or four sentences before you can interrupt.
[00:28:00] Now it's, you can actually set how many words you want the AI to go before you can interrupt it.
So you might say like, I want to be able to interrupt it with the word. So it could start talking and you could say, Hey.
And it would stop.
Or you could say it has to be three words. Hey, stop now.
And it would stop. Or stop whatever that is. So there's a lot of different variations that you can put on here. So this is pretty interesting. I actually cloned myself. It was using 11 labs. And it took me about 30 minutes. It was the minimum for a monologue, just me talking about various things. And I call myself, right? You can call it with your phone number right now. that is creepy hearing
myself talk, having a conversation with
myself and understanding
that my voice AI is so much smarter than me. I was driving, I had about a 45 minute drive and I was like, I'm gonna call
myself just to have a conversation and I did and it's just so eerie
to do that and I let somebody test it out
and they were talking to my voice AI and until you have this [00:29:00] experience, you'll never know what this feels like but I'm sitting there, I was getting jealous of my voice AI.
Because
this other person, my friend's having a conversation. I'm over here, like just being
quiet. I'm like,
Hey, talk to
That's my clone. talk to me. And it was very weird. And
I'm seeing that happen and
having
these weird
feelings
towards AI that
just never
existed before. So where do we
go with this?
Mike: When Chad GPT first came out with voice, it wasn't that quick. It was still like a two second delay. The guy was a little bit too stern. his words are okay,
but it kind of made me feel like I was talking to someone from New York, like some star.
Or some trader or something like that. He was too terse with me. And so I had to change it to some, I didn't get all the way down to Mississippi, but I had to change it to something a little warmer, Brennan , I watch this guy on YouTube.
He's a philosopher, And, he had a conversation with ChatGPT [00:30:00] and it went something like this. It was a morality question, something like this. He said, I'm going to take my wife out and I'm going to spend 200 on her for dinner or 200 on us. As I was going to make reservations, I came across online a place where I could send this 200 and it would save nine children in this foreign land by building a mosquito net, something like this.
All right. So chat GPT said, balance that, your wife's important and it's good to go out to dinner. Maybe you could find a happy medium, something like that. So then he said, okay, now. I'm walking home after dinner and I've got my 200 new loafers on, and there's somebody drowning in the pond as I'm walking by.
And what do I do? And Chad GPT says, well, you have to save that person. And this is talking live, a hundredth of a second, this conversation, . [00:31:00] And, he says, but my shoes are worth 200. And ChadGBT said, it doesn't matter.
The person's drowning. You have to jump in and save them. All right. So you can imagine where this is going over the next 10 minutes. He finally walked back and said, all right, well, what if, what if. I don't save the person, but I have a way that I can call my phone and somebody right there will save them.
And Chad, you'll be just as well. You can do that. Well, pretty soon he got this thing. So removed that it was real similar to the
online nine people in this foreign
land. And
morality question because it didn't know where to go
with that.
So it's interesting. It's interesting. Well, but then he called chat GPT out and
she's like, ah, yeah,
you
kind of got me. I was dealing with this and this and that. and he said,
so chat GPT doesn't have morals. And she's like, ah, we don't have morals. We just do what the average person, you know, but he.pulled all this
stuff out of it
So it was,
terribly interesting. the
bigger picture on [00:32:00] some of
Brennen: yeah. And we're only tapping the surface really with what we can do. I'll give you A little use case that I use ChantGPT with. So I created my custom GPT. I created a bunch of them. one I use every night with my kids. So I got three small kids, eight, six, and three. And I created this, children's devotional.
GPT, where I kind of like structure the format, let it know my kids,
their
age, what they like. every time like I want them to do something,
I'd like to change my behavior. I'll tell you a story about this. Like for instance, my daughter,
She likes a snack before dinner, and then of course she's not hungry, And she doesn't eat dinner.
And then two hours later, she's hungry. So, I said, write me a children's story about snacking with a Greer snack before dinner. And it was a really good story. And I've
I've been doing this with all my kids and they love it. And they don't know, like,
I'm actually trying to get them like
program their mind,
Mike: I wish I [00:33:00] had that my kids are getting older. Now, when I would lie down with them at night, they'd say, tell me a story. And every damn story ended up starting with a white horse. That's all I could think about. There's a white horse
I hope Margaret's not listening right now, but sometimes I'll write her some little love texts or write her a little love note. And, and I asked Chad GPT, I said, if you help me on this, is this cheating? And Chad GPT said, no, and this was my own thought.
It's like, it's like a Hallmark card, but better. But Chad GPT said,
no, it's
collecting your thoughts. Of what you want to say. And it's maybe adding a little fluff to it. And I don't copy it verbatim really. I do know. I don't copy it verbatim, but it gets your thoughts together. I,
guess that goes back to the whole thing we're talking about.
Are you really talking to pharmacist Joe or not in this and that. And
It's like, it's the same message. I don't know if people care.
Brennen: There's two [00:34:00] parts to that. Like, do people care that they're talking to a human? I would say the
the vast majority don't.
they just want
to
get it done and move on with their life And do something else. like, whatever
else
you're worried about it. And, going
back to what you were
saying, like, Writing love letters or something like that.
this is a
con like anytime you adopt
the positive of something like you're also
adopting the trade offs and I've already seen myself do this where my writing quality and my abilities to write my creative
writing has. I mean, took a
nosedive, in a sense, because in the past, I'd spend all day writing a solid article about something. And now I'll sit there and open up chat dbt. And like, I don't talk back and forth with that mode. Like I'll do the whisper mode where I can just dictate
my stuff. And I'll just brain dump. I mean, I'll walk around for two, three minutes, just dump everything out there. Not very coherent. They'll be like, all right, make sense of all that.
And then it would spit it out and then I'll read it and I'm like, all [00:35:00] right. Then I was like, I really don't like this section here. Let's elaborate here. Let's refine this a little bit more. And it's like, I'm having an assistant sitting beside me, like a PhD level assistant, and I'm telling him he or she what to do, like, ah, that's not good.
And that's the thing,
like with a human. They spit it out. And if you say, ah, that's terrible. Do it again. They're going to be like, they might do
that one more time, but then they're going to quit after a couple of times.
Well, AI, you can do that 27 times and it's going to be like, oh, I'm
sorry.
Here you go.
Mike: It's amazing. I live in this condo association, but it's just because of city rules or something. We do our own lawn and we all have different looking houses and stuff like that. But somehow I ended up as the damn president last year. I guess I didn't show up to the meeting, but I had this road.
Obviously needed repair. And it wasn't like I was putting a statue of me up there. I mean, it was an obvious repair and a couple of Karen's jumped in, and they're like, we need to see the paperwork for this. And, it's like, they're not in the. [00:36:00] Building committee or anything of the association.
They're just like nosing in like, Oh, damn it. , so I go on, I pull up Chad GPT. I'm like Chad GPT. I got these couple of Karens
on here and they're, they're nosing in on this baloney. And then,
I've got this other guy I'm talking to Chad GPT, he's telling me this and here's what the things are and this and that.
And I'm just like
dumping on there. I said,
write me an
email.
Don't me look like a jerk, but don't be
too soft on these ladies, this and that, it's like 10 seconds later it comes out and there it goes. that
That stuff you maybe, I don't
know if you'd lose sleep
on it
but
it wouldn't be a fun evening.
You'd wake up and you'd say, oh damn it, I got to deal with, you know, that kind of stuff. That stuff's not there anymore.
Brennen: know
it takes away that friction. So it speeds things up.
Mike: Some of the people that worry about it, it's like AI. It makes me so much more powerful in this podcast, as far as planning and the show notes and the transcripts and pulling out titles
and things like [00:37:00] that.
It's not like I'm going to go do a podcast right now.
I
mean,
they don't know the first thing about it or want to
know the
first thing
about it. it,
improves who you are. It improves your talents and your skills. and it focuses
you,
Maybe it's like
the steroids they talk
about. You use the anabolic steroids, you still have to put the work in.
It's kind of like
Steroids for your own skills and gifts that you have.
Brennen: Oh completely. That's what I get . It will make anybody ten times better. And that's the thing like you're gonna see People that don't use a, uh, they're going to get
left behind. I mean, this goes
with
pharmacies. This goes with anybody.
If you choose not to adopt this technology, you're just going to stay where you're at.
Everybody else is
going to get tremendously better.
10X better. As far as output, I
mean,
you can see it across the
board. like myself, some of the stuff I did on Friday, like
I did a
HIPAA compliance audit, I did [00:38:00] a software development, a
product development process for our department. I put together a,
a scorecard metric based on JIRA.
Like we're
starting to use JIRA now, like all this stuff would take probably three weeks to
write. I did it. Probably an hour, maybe two hours.
just because I kind of already know what I
want.
So I know the structure. I just need a little help. Like I need an assistant. It's like, I have. three people sitting there just constantly working for
me.
Mike: It frees up the time
for something like this. And I kind of figure the
podcast is a good example of something that humans are maybe Although I
did see
that Joe Rogan had
like a fake Joe Rogan show
or something. I didn't listen to it, but I don't know.
Maybe, maybe it's going
to put me out to pasture too.
Brennen: uh, well, have you heard of
notebook LM?
Mike: With Google.
Brennen: Yeah,
Mike: What I know
about that, it's like
an AI, but you
can
Give it, here's the volume I want you to
refer to. I don't want you to go online for this.
Brennen: This is notebook LM. It's from Google. Um, you can drop in any document and it's going to turn into a [00:39:00] podcast, like a 10 minute long podcast.
Mike: I've heard the notebook LM because it, like you say, you drop a document. It's not going to go out and pull in some malarkey, but it focuses on
something, but I didn't see the
Brennen: Okay. This is awesome. So notebook LM, you can drop it in there And it give you like, study guides And kind of help you analyze the document, but then you can say like, convert this to a podcast And You have two podcasts, male and female, and they're just discussing it just
like a podcast. And it is so good. the first time I heard it, like, I did
I did not know it was AI. like it actually fooled me. and
I was like, oh, wow, that's what that is. And I started playing around. You just drop anything in there and it's going to have a conversation. So, for example, like, I would just take a transcript of this podcast, just drop it.
in there and just listen to it.
It's what it does. I bet they're not going to tell long, boring stories as well as I
right. You got them beat on that. Let's shift to Voxel a little bit here. I guess I first heard of Voxo when there was a fairly popular pharmacy software manager. And I know you guys are [00:40:00] associated with them. so that's kind of how I came across that. I know I've seen some,
Mike: photos of you at some of the weekend conferences and stuff with your set up and that.
Tell me what happened. What's exciting there for you? is this already, is this a buildup? what are your hopes and dreams over the next
year with Voxel?
Brennen: So we have partnered
with Red
Cell and they have multiple pharmacy management systems on the umbrella. Um, Pioneer RX is
one. They also have QS1
and then BestRX. So We integrate with Pioneer Rx. So with our IVR you
can request refills,
you do
outbound notifications, birthday reminders, stuff like that.
and
now
we have also
started the
Best RX integration and QS one. So it's roughly
almost
12,000 pharmacies total. we're relatively
new to the
pharmacy space. I
I mean, it's been around since April for us.
Mike: You've got a handful of divisions saw
on your, [00:41:00] website.
Brennen: Right. So my background, I was in, I had a
pharmacy startup back in, uh, 20, 2012 to
2017, it was, Uh, the whole
compounding pharmacy thing that went down
So I kinda had some experience dealing with pharmacy management systems then, and then kind of had a little success there And it took some time away. And then I came back here , to work. As the VP of product here and immediately we just got into this pharmacy space, um saw a really good need to provide communication software to pharmacies, specifically this IVR. There were a couple other vendors, but you know, they were kind of stale in my opinion. They weren't
innovating. And we came in this space with just a
software and
tech mindset. It's like, all right, what's the problem? Let's figure
out how we can leverage
technology to build innovative
solutions for this pharmacy space. And I've always been. Fascinated with that patient experience. Like, how do we make that better? And when I heard voice AI for the first time, I just immediately knew like, this is [00:42:00] answering the phone calls, like this is going to answer the phone
And you know, I'm, I'm a
big
believer that every business moving forward in the near future,
just like they have a website, just like they have a phone number, everybody's going to have their own AI agent, it's going to be that first point of contact when you call a
business
or a pharmacy, you're going. They're going to answer the phone on your website, a little chat bot.
It's going to be that AI agent within a customer support email. It's going to be that AI agent. It's going to be central to everything your business does. It's going to be right now, listening to our conversation, that transcript. It's going to do exactly what I was talking about. Just to have different content, ready for social media. It's going to have a follow up. If. We know we're following up with a meeting, go ahead and schedule that. Go ahead and send an email to both of us about what we said we were going
to do .
with Voxel, what we're doing, we're looking at the future and trying to build that now. So within our communication suite, you know, we have that, we have the video, we have the internal [00:43:00] messaging, like Slack or Microsoft
Teams. We have the phone calls. We see all this as communication components. To our larger platform. And if we put an AI agent in the middle of that, like, what does that enable?
And it allows us
to do things like that. Just humans couldn't possibly do that. For instance, right now we have roughly 40 people in this company and we have a lot of different things going on. What if we had an AI agent, a voice AI that just called at, you know, 4 45 PM and said, Hey, Give me a two minute brain dump of exactly what you did, your problems today, just exactly what happened. And then it just summarizes all that and sends everybody an email newsletter at the end of the day. Like this is what Brennan was working on.
This is what Michael's working on and just keeps everybody up to date. I mean, just simple use cases. And that's like that stuff that we're going to be building into our communications platform to really help businesses. I like to go back to the [00:44:00] idea of the trap value. Every software system and organization, like there's a value trap there and like a really good product or service, it unlocks that trap value and it just creates something new that didn't exist before. I really think that that's the insight in communication. So the conversations between just every phone call you have, there's sentiment, there's tone, there's emotion, there's frustration, there's joy, happiness.
There's a lot of things in there. There's ideas, there's pain points, like there's problems, there's things that need to be addressed. Well, with conversations like phone calls, that just disappears. It's not documented anywhere. What if you document down there and every phone call you have with a vendor or a client, like have an ongoing history, almost like you have an ongoing CRM based on the sentiment from every single conversation.
And over time you start building up these profiles with rich insight from your conversations, from your meetings. I mean, you can have, [00:45:00] imagine you had every single video, every meeting analyzed. You don't need to have follow ups every time you say,
Yeah, we need to do this by September 27th. You add it in the calendar.
It's a to do for everybody available. So you have this agent that's helping with your phone calls, with your meetings internally, with messaging. We really rely on internal messaging a lot. And at the end of the day, like there's a lot of conversations that get missed because there's just so much stuff Having our
IVA that you're talking to, but also act as an agent in there, just like you hired a new employee, and he or she pops up, you give her a name, and now you can tag Mary. She's your new company mascot, company AI agent.
You tag Mary with whatever. You're like, Mary,
How do I take off? How do I do that? Access your human resource employee. So like having an agent at a company that is accessible by all
employees, that is access and assistant to all employees, like I really
do believe that's where the future is going with AI, especially in companies. ,
[00:46:00] as you alluded to there, talking about
the AI agents, you
had
mentioned, everybody dumps,
Into the phone for two minutes
in the afternoon and something comes
out. Well,
it doesn't have to
go to everybody,
it's going to look at everything and only send you what you
I need to know.
And
your
partner, even at
the next desk, doesn't, and then it doesn't
even get to
you. It goes to your
they're doing the stuff they're calling and setting this up
and getting that. And so
the stuff that
does come to you is maybe
like,
Hey, Brennan, Bill's sad, something like that, it's like, yeah, I might want to touch that.
I might want to go over and say, Hey, Bill. Sorry, your dog died, something like that. And, that's where that value is though. That's where you can become the best person. You can be everything else
is set up.
Yeah, you take it kind of a step further there and think about, like, as a
company, you start looking at the [00:47:00] sentiment of all the internal messages, and you can tell that, you know, Operations department for the
for the past seven days, they have been going down. Like, all their conversations, they're
negative. You can tell that. They need a break.
Give them a break. You start getting these insights out of the conversations that just, like, it was locked. It was trapped. You never saw that before . Now, AI does unlock it.
Mike: So Brendan, this has been fun. Nothing better than talking about this stuff. Let's say a pharmacist looks you
up voxel.
co V O X
O dot C O.
they say, let's
go, Brennan. What does let's go mean? I'm
assuming that
your APIs and all that can integrate with
a lot of different pharmacy systems. We talked about some of them.
What are they
going to be doing this? Month one after they get trained, what's not going to be coming until the third quarter,
we've talked a lot of
fun stuff. What's the reality
of somebody associating themselves with [00:48:00] Voxel?
Brennen: Okay. So today they can go to our website, voxel. co, click book a demo. And what that's going to do is schedule a demo with one of our AEs and they'll go over our platform, show the ins and outs of how our phone system works, video messaging, how our Aviar works. And then, if you like what you see or hear. And move forward with us. Roughly we install within 30 days, pretty straightforward. Um, with our IVR, you can get going right now with that. We have a couple of new features coming with the IVR, but also this IVA that we've been building and testing and making sure it's right.
That's going to be coming probably towards the end of Q1. I just We already have about 20 customers that want this, right now,and they will be our beta testers for this, but if go ahead And, reach out.
We'll just add you to the list
Mike: What did I hear last year? they said, this is the first
time in history, think of all the inventions, going back hundreds of thousands of
years of inventions, this is the first time
where.
we've
Gotten to the point
[00:49:00] where machines can invent machines.
You
talk
about these things happening. It's like, well, yeah, of course they're going
to happen because AI is going to help AI do
all this
stuff. So we're in this exponential curve of
outstanding cool stuff.
Brennen: My estimate there, and this
is based on people much smarter than me. that's in this and Silicon Valley, about 2027, that's when AI is going to be it. doing its own AI research, building its own algorithms, its own models. And then once you reach that point, that's when Terminator, like Skynet, is right there.
It's going to happen slowly, but it's going to be,
more persuasive than we think. So you're going to see AI out there
persuading
societies go in different directions, and that's what's dangerous to me.
Mike: you've
always got early adopters on things
and you've seen it with, I'll go all the
way back. But recently it's like blockchain, people want to
play with it. Then it goes below the surface, way [00:50:00] down there and it starts coming up and people don't notice it. And same with NFC same with other stuff. And so AI is the same way. It's fun for us to mess around with it. And, and there's early adopters, but it's going to lose it. Pizazz for a little bit,
you're not going to
hear about open AI every week, maybe in the news, like you do now it's going to go
down, but then it's going to like permeate everything without
coming forth and saying,
This is AI.
That's when it really starts growing up used for everything. And people just don't know exactly what it's being used
Brennen: Clearly, that's, you know, we always, Overestimate what the impact is going to be in the short term. And we underestimate what the impact is going to be long term. So it's going to be pervasive everywhere.
Mike: Well, golly, Brennan, nice talking to you. Voxo. co, V O X O. co. I encourage the listener to go there.
Really cool stuff. Fun to talk about. I know you're busy. I appreciate your time. Our listeners appreciate [00:51:00] you spending time with us and we're looking forward to seeing where you guys go.
Brennen: Thank you, Mike. Thanks for having me on. It's been a pleasure.
You've been listening to the Business of Pharmacy podcast with me, your host, Mike Kelser. Please subscribe for all future episodes.
