Is there a single right way to run a home care agency? We sure don’t think so. That’s why we’re interviewing home care leaders across the industry and asking them tough questions about the strategies, operations, and decisions behind their success. Join host Miriam Allred, veteran home care podcaster known for Home Care U and Vision: The Home Care Leaders’ Podcast, as she puts high-growth home care agencies under the microscope to see what works, what doesn’t, and why. Get ready to listen, learn, and build the winning formula for your own success. In the Home Care Strategy Lab, you are the scientist.
Miriam Allred (00:01)
Hey everyone, welcome back to the Home Care Strategy Lab. I'm your host, Miriam Allred. In the lab today I am joined by Dave Dworschak the head of growth at Phoebe Dave, how's it going?
Dave (Phoebe) (00:11)
Good Miriam. It's good to be back. I'm excited for the conversation.
Miriam Allred (00:15)
Good to be back. It's been a minute. We were both looking to see how long it's been. Last time I interviewed you was January and here we are, November. So it has been a hot minute. A lot has changed, a lot has happened for both of us. So excited to reconnect.
Dave (Phoebe) (00:27)
Yeah, no, absolutely. I think it's been we've we've almost like doubled the lifetime of our company since we since we talked last. So this is great.
Miriam Allred (00:35)
Yeah. Let's start with a a little Phoebe update, you know. ⁓ you all obviously have been like sponsors and partners of the podcast for a long time, so people are familiar with you. ⁓ we've got a lot of people listening that are customers using you. Give us a little like behind the scenes update of where you guys are at.
Dave (Phoebe) (00:51)
Yeah, sure. So for for folks that maybe heard heard me on the podcast here in January, you know, quite a lot has changed since then. But just as a quick introduction for anybody that's new, you know, I'm Dave. We're Phoebe. We build AI teammate for the home care agency, which or sorry, the home care industry, home care agencies, which we'll be talking about today. ⁓ you know, since January we've we've grown a ton. I think the last time we talked, our team was, you know, maybe half a dozen or so, maybe a little bigger give or take. ⁓ and now we're
close to 30 people. So team is growing, customers are growing. I think we maybe had a again 100. I didn't check the facts, but maybe 100 or so agencies last time we talked and ⁓ as of last week we're live in over 800. So we've been you know growing quickly and are still focused on you know home care and home care scheduling as the primary thing but we've been really expanding ⁓ now that we've we've spent so much time really focused on just getting the hard things right and you know doing the hard things well we've we've started
to expand into some other avenues outside of scheduling, which we can talk about. But you know, by and large, we're working with you know with home care agencies to solve the pain points in and around, making sure the right caregivers get to the the right shifts at the right time on time every day.
Miriam Allred (02:02)
Awesome. Yeah, we'll talk a little bit more about like your focus on scheduling towards the end. ⁓ but we'll kind of like weave that in throughout the conversation. I wanna just ask you like a variety of questions about like where AI is at, what people should be paying attention to. I've talked to a lot of agencies lately about like builder buy. So I just wanted to like bring you on to ask you like a variety of the questions that are bubbling up in my own mind, but also with a lot of the conversations that I'm having. People are asking me these questions, so figured I'd ⁓ fire them off to you. So
Just starting kind of like broad stroke, there's a lot of noise in home care right now when it comes to AI. You and I were both at a variety of conferences this past fall. There's AI companies popping up left and right. And so it's just getting as noisy as it's ever been and will probably continue to even get worse. but what's worth paying attention to right now? Like, is there too much hype on any one thing, or are there things people should stay away from or should be looking into? What's what's your take on the landscape right now?
Dave (Phoebe) (03:01)
Yeah, that's a great question. I'll like maybe start home care specific and then we can zoom out and zoom back in. ⁓ but I I think the the biggest thing
you know, is there's a lot of new companies coming onto the market. And I think that as new companies come onto the market, the the reality is is like when you start a tech company, it's like, well, let's just like tell everybody we do everything, ⁓ and we'll see what people actually want and then we'll kinda narrow and figure out our focus from there. And by and large, that is the right strategy for most ⁓ technology companies. But I think with with AI and people being able to build so fast, you've got a lot of companies coming onto the scene that are just like, you know, we're gonna do everything and we're gonna say we do everything and then we're gonna actually try to do it. And
I think what what
ends up translating to owners like evaluating I AI is like, you know, I just want like one company that can do everything for me. And I think people are focusing a lot of attention on like how many tasks can the AI do. And they're not putting enough attention on, you know, whether the AI knows enough, you know, and has been built well enough to actually make the right decision to do the actual work that you're you're kind of hiring the the AI to do. So that's like one thing that stands out is like as as new companies crop up, everyone's like, well these guys say that they do everything. so I'm just gonna use these these
guys for everything and the the reality is is if somebody's been around for for six months they they should not be doing everything if i if i zoom in like a little bit more ⁓ you know i think there's there's a lot of kind of hype coming out with the the ehrs ⁓ or scheduling systems kind of building internal ai and don't get me wrong i actually think that's a fantastic thing ⁓ i think a lot of operators will benefit a ton from from that ⁓ but i i think that
there's too much kind of focus or hype on like how do we just keep everything in one system. ⁓ and I think that there's been times in the world when when that's been true. But I think, you know, one of the beautiful things about AI and large language models is the you can give them context to many tools and they can work in many systems. ⁓ so I think the the kind of like
The hype of like, hey, I just can like use the AI in this one tool. And every tool's doing this. It's actually isn't even home care specific. ⁓ every tool anyone logs into nowadays has its own chatbot on the homepage. but the reality is like most people shouldn't be using any of those things because the the tools only have the context that's in that tool. But like a home care industry or home care agencies business is not just the scheduling system. It's like the CRM, it's the phone lines, it's the text messages, it's the team conversations happening in Slack and Teams, it's the group chats and your iMessage. So there's
There's like so much context that I think the hype of like let me just wait for my provider to build the AI, ⁓ I think is is is a pretty interesting topic of conversation, at least today, and that could change.
Miriam Allred (05:42)
Yeah. Somebody said this to me recently and I was like, wow, that's a really good analogy. Was think of like Microsoft has co-pilot. You know, Microsoft's this huge conglomerate. They have a ton of products. They have the co-pilot, but how many people are using co-pilot and like maximizing it? Everyone's out still using chat and Claude and Perplexity and all these other tools in that, like, you know, all of the EMRs in the space, they would be silly not to be building this AI technology, but there's all of these other tools, these one off tools that like
are better because the tech companies, you know, it's their sole focus. And so I like that analogy of like Microsoft and Copilot versus like the point solutions. And that's kind of like likened into home care with the EMRs. ⁓ the last like three to six months, I feel like we've hit some pretty major strides with just like improvements of these models. ⁓ what are some of the advancements that you've seen firsthand when it comes to voice agents, processing speed, APIs, interoperability? Like what have been some of the
The key advancements that you've seen like firsthand the last, I don't know, three to six months.
Dave (Phoebe) (06:46)
Yeah, I think the the first one that probably stands out, it kind of goes to what I what I one thing I just said, but I'll expand on it where like it we're getting to the point where you can connect your LLM, whether that's, you know, you're using chat or Claude or whether you're using something that's built for your industry, like Phoebe, you know, you you can connect it to to many different tools. and you're you're able to
Okay, I'll I'll say it this way, like six months ago, twelve months ago, it's like to make the AI effective, you had to be just incredibly good at prompting and you had to know exactly what to say to the AI. ⁓ it's getting to the point where if you just can get your AI connected to all of the context, you actually don't have to be that good at asking it what to do anymore. ⁓ which is gone from, okay, now I have to like study to figure out how to be good at using AI, to now I just like talk to it like I talk to my ten year old and it it actually works quite well. ⁓ but it only works well if you're able to like be working in an AI that has
Kind of cross-system and and cross-channel context. and I think that gets further enhanced, and this is more of a technical advancement, but we've gone from like browser use and computer use, ⁓ where like your AI is kind of using the computer for you or using the browser for you, being a thing that like was slow and it kind of worked, and ⁓ you know, it was almost more of like a gimmick and and like a flashy demo to like the the
like AI can use your computer and use your tools and like log into something it's like never seen before and know exactly or figure out exactly what to do faster than you can probably figure out what to do. ⁓ which opens up like an entire new door of possibilities for what you can start to use AI for because you're no longer constrained to like, well, I have to have the AI connected to all of these APIs and to get connected to the APIs, the company has to provide the APIs and then the software tool I'm using has to build to that API. Like we're getting to the point where it's like
If there's an API, great. If there's an MCP server, great. ⁓ if there's a connection directly for the AI to talk to the tools using, great. But if not, it's actually kinda fine too. It can just use the browser, use the computer and do that on your literal computer while you're working and it doesn't really even get in the way of what you're doing. It's it's pretty fascinating.
Miriam Allred (08:49)
So explain that a little like simply but technically because pushing back against just some of these tech startups, not like PV specifically, but just pushing back a little bit. Every every tech website that I go to right now, one of the first things on their homepage is like interoperability. We connect with your entire tech stack. But I think that is misleading in that you might be connected.
Dave (Phoebe) (08:52)
Mm-hmm.
Miriam Allred (09:14)
in kind of like a light version, you know, I I just don't feel like there's full interoperability, but a lot of the tech companies are touting that there are. But explain to me what the home care agency, what the provider does to connect their tools to like truly be integrated.
Dave (Phoebe) (09:32)
Yeah, it's it's tricky with like an industry like home care because if unless you're doing like fully private pay, non-clinical, where you're not having to deal with kind of HIPAA regulation, like it's probably a bad idea just to like connect your, you know, twenty dollar chat GPT subscription to your to your tools because like you're not working in a HIPAA compliant environment. ⁓ not to say that chat can't be HIPAA compliant, it very much can be. all the models will offer BAAs and can be can be HIPAA compliant. you just have to spend a lot of money to to make that happen. but the the
Interesting thing, and I'll try to break this down. Like the the interesting thing is like a few months ago, even, or we'll call it back like 12 months ago, a lot of people were like, Well, we integrate with all these things, you know, all the you know, we're the most interoperable. And really, it was just like, let me just like put some logos on the website and then see where I get the most hits, and then we're gonna go integrate with those tools. And that's honestly what most tech companies do. ⁓ but the browser use and computer use has gotten to the point where like you you don't really need to even have an official integration in most cases to be able to.
Do things in the system. And there's been things like ⁓ RPA or like robotic processing automation for a long time where you can do like web scraping, but that this is like a totally different thing because it's like the AI is kind of using the computer, using the browser. And maybe the simplest way to explain this is like most companies with their APIs, which is their connection point they're giving to other companies to get data in and out of their systems, they they are limited by design, which is
There's lots of good reasons for that, actually. but what that means is like a company can have an integration with, you know, this software system, but they're kind of at the mercy of like, well, this software system's APIs only let me do tasks A, B, and C, and it does not let me do tasks, you know, X, Y, and Z. So now with advancements in things like, you know, browser use and computer use and being able to give your agent a computer and give your agent an email address and give your agent a phone number, like the ability for an agent to kind of work in a tool like a person, you can now say, Well, I'm gonna use the API.
For A, B, and C and on times where I can't get it done, rather than kind of pointing back to the human to say, like, we can't do that in the system, the AI can just use the browser, use the computer to do the tasks that have never been possible before.
Miriam Allred (11:43)
So I'm just personally starting to dabble in this. So let me understand, make sure like I understand this right. So like you have a tool like Claude, for example. I just told you like I'm kind of migrating from chat to Claude because the Claude models are just so good. And every time I like get these responses, I'm just like wowed. ⁓ Claude has Claude code, codecs, whatever, where I can you give it access to your computer, therefore, all of the files, all of the information, and then
It's kinda like like screen monitoring where if I'm in a browser, it's watching what I'm doing in that browser and then it can learn from my my engagement, my activity. Is that what's happening? Is that what you're referencing?
Dave (Phoebe) (12:23)
You know the the the
It doesn't have to do that anymore, I guess would be the simplest way to explain it. Like you used to have to like train the models on like, well, I've got to have enough data of, you know, enough users using this tool so that my AI agent now knows how to use the tool. Like the models have gotten so good that like the, you know, your agent can just say, like, hey, go look in this tool to see if you can find this information and it will just go figure it out. And it could be the first time it's ever seen that tool before. and it will still do pretty well. ⁓ this is actually interesting, but we've
Chatted with a few Phoebe users lately that are just, you know, we've got a broad spectrum of Phoebe users, some people who are like very AI entry level, and some people who are, you know, just as advanced as you know we are that we're building this and they choose to use Phoebe for a lot of good reasons, even though they know how to build stuff with AI. ⁓ and we've talked to some users lately that like they they actually don't see, they don't look at Phoebe very much anymore. They just talk to their agent and their agent uses Phoebe, which is great. We don't, we don't mind. ⁓ you know, come use the Phoebe interface. It's ⁓ you know, honestly easier for us.
some regards if we you know it's like the agent will figure out you know something that might not be the best placement of a button better than a human will figure out something that might not be the best placement of a button. So the shorter answer to that I I guess would be, you know, you you don't
You don't like give it access to just like read your screen all the time. ⁓ there are companies that force people to do that, but if you authorize like Claude to do computer use or Codex or ChatGPT to do computer use, it's not monitoring what you're doing all the time. It's just using your computer or your browser for you at the point in time that you're asking it to carry something out.
Miriam Allred (13:56)
Got it. And just to clarify, the agent that you just referenced, like Phoebe has built an agent that's kind of like a monitoring agent. Because I don't know, I've seen the you the word agent in like every tool that I use. And it's kind of like, wait, it's almost getting a little like confusing, agent, ⁓ the term a little bit. So you're referencing the agent inside of Phoebe that's similar to like using Claude, where it's like you're giving it access and then it's monitoring, and then you can prompt it as like its own agent.
Dave (Phoebe) (14:10)
yeah.
Yeah, so so most of our users, you know, as a human talk to Phoebe the agent, ⁓ which has access to your EHR, your voice system, your, you know, Slack channel. you can talk to it in iMessage, like all these different things. So ⁓ most people that use Phoebe today are using Phoebe the agent, but we have some users that use their personal agent and just have their personal agent go talk to Phoebe and ask Phoebe to do things. so we have some users now that are kind of on the more advanced side of AI that they're like, Well, I just have this.
know Grokbot would be an example where I just talk to Grokbot all day and I ask Grockbot to go do stuff in Phoebe for me or go, you know, tell me what's going on inside of the Phoebe automations, rather than having to go look at a screen or read the notifications or like, hey Grokbot, hey Grokbot, go check out, you know, what's going on with Phoebe in the last couple hours and let me know what's happening. Does anything need my attention? So people are using agents to talk to other agents would be the it's it's all kind of confusing to be honest.
Miriam Allred (15:19)
Okay. G for those that don't know what Grokbot is, explain what that is if that's the first time they're hearing it, Grokbot.
Dave (Phoebe) (15:26)
Yeah, so y basically it's like XAI's version of a of an agent. So you've got like Muse from Meta, you've got dot that chat GPT released. ⁓ gosh, I don't even know if Claude has released one yet. You got all the open source models if anyone's kind of was following following all the ⁓ you know, hype maybe six, twelve months ago where you had kind of
ClaudeBot was the original name, but then they they changed names because of Claude. But it it's basically like a personal agent that you know has access to your LLM that then has access to all your tools. So GrocBot is like XAI's version of you know Muse Meta's Muse or ChatGPT's dot.
Miriam Allred (16:06)
It. We're just confusing people. But this is but this is a good just like reality check. No, it this is good because again, it's like there's names and labels and agents like all over now. And it's like hard to keep it straight. ⁓ on the model side, like I just reference personally, like kind of migrating from chat to Claude and seeing like the benefit of it. ⁓ are there any models you think are worth using right now or paying attention to right now? Like think of all the people listening to this, they're using like all of the primary tools, but like
Dave (Phoebe) (16:08)
I know, I know.
Miriam Allred (16:35)
Would you be pointing anyone in any one direction or encouraging people to use multiple for different things? Like what's your advice there?
Dave (Phoebe) (16:44)
Yeah, it's it's a great question. It it probably depends on like your level of use. Like, you know, for me personally, like I will switch between like Claude and you know, OpenAIs, you Chat GPT or Codex Claude and Claude Code from Anthropic. Like I'll switch, you know, sometimes every few months because the the model's really, you know, it's gonna be better over here for a few months and then better over there for the next few months.
I mean, right now it's, you know, depends on what your job is. But if you're, you know, doing software engineering, you're probably using Claude and Opus, you know, here on October thirtieth, ⁓ twenty twenty six. And, you know, it it could be a totally different answer, you know, next week or next month. But I think for the vast majority of people, ⁓ most of what you're doing if you're using, you know, kind of your, you know, your
Claude or your OpenAI, or if you're like a Gemini user, or if you're you know an X AI user or like whatever, you know, a meta muse user. Like it's usually for most people, you can just kind of pick one and stick with it. But you will find I think we were talking about before we started recording here, like for you, you're like, okay, wow, Claude is like actually outperforming what I've been used to in in chat for the last few months. And it's because you you're probably using, you know, the newer model, which is is now outperforming in some and it's it's all up for debate, but it's it's outperforming.
performing in in some ways than the the current latest model from from OpenAI, which again as of today is Astro but will be different, you know, probably by the time most people hear this.
Miriam Allred (18:06)
Yeah. ⁓ let's go into kind of like the builder buy conversation. I want to hear who you think, you know, home care is like fragmented and vast and like all these agencies are at different stages and different like tech capabilities and stuff, but like who should be building and who should be buying and why?
Dave (Phoebe) (18:27)
It's a great question. my
One sentence answer is e everybody should be building something. ⁓ regardless of whether you're also buying, like everybody should be building. ⁓ I think if you start with, you know, if you're still at the point where you're not really sure where you're gonna bring AI AI in, ⁓ you know, you can be building all sorts of kind of like say micro automations that I would say f focus on like building the things internally where like real time mistakes won't cause your business to fail. ⁓ so if you think about things like analyzing data, preparing for meetings, you know, you know
Of ideating on kind of follow-up drafts. I I I almost said writing emails, but I'm I'm not a big fan of people using AI to write for them. It's probably not the best decision. but you know, ideating on what you should be writing, ⁓ or kind of building small tools to address like workflows in your agency that if the workflow fails, it's not gonna cause a problem in real time. so when we think about building versus buying, I would say, you know.
outsource where like reliability or real-time reliability is like paramount to the success of your business. So things like recruiting, you know, interactions with caregivers or scheduling with clients or scheduling with caregivers or, you know, picking up the phone. These are things where if like you make a mistake, the system goes down, you know, you don't have the right guardrail set, the AI doesn't have the right context, it's it's literally going to cause you to, you know, lose clients, lose employees, lose revenue, lose business. ⁓
Sometimes worse. so that's really my big distinction right now. And I I I think there will be a time in which that shifts. I don't think that time is quite any time soon, but I I do think everybody should be building something and then outsourcing the things where real time re reliability is a must have.
Miriam Allred (20:07)
Yeah, great answer. And I completely agree. Two other like just advice from my end, what I've seen, like wherever there's paper and wherever there's dual entry, like a lot of agencies, there's still paper and there's still dual entry. And it's like in the era of AI, those two things should become obsolete. If they haven't already, those two things have to become obsolete in your business in the next 12 months. Like there's just no reason for that to still be happening. And so those are kind of like the two areas where I recommend people double down.
You've worked with, you know, hundreds, if not thousands of agencies at this point. I feel like a lot of owners are the AI champions and also kind of like the most tech savvy at a lot of these, especially small to mid sized businesses. What's your take on people developing more AI champions and power users inside their organization? Like how essential is that in growing an agency right now?
Dave (Phoebe) (21:03)
Yeah, I I mean at this point my my general ad
advice to any owner would say AI should not be a side project of a single enthusiastic person in the company. I think we've gotten to the point of advancement in AI. And again, come great if you want to come talk to Phoebe but also if not, you should be doing this anyways. Like we've gotten to the point with AI where you should be doing something and it should not be a side project of like a single champion in the company. That said, like I do think champions are still important. You you almost with always any sort of technology or really even process adoption, you you need like a kind of directly response responsible individual
Individual who is going to be like overseeing progress and making sure that things are still moving forward. So I think having a champion is important, but the the way to get that to kind of disseminate out to the rest of the company is, you know, if your champion does not equal the owner, if it's not the same person, then like the the owner first and foremost needs to be modeling the behavior that the champions are trying to get across the business. Because I mean, this is not telling anyone anything they haven't heard before, but if you as an owner were a manager or leader are not
you know, doing the things you want your team to do. They're they're just not gonna do them, or the vast majority of them are not. So important to have somebody responsible, somebody that you know you may designate as like, okay, Dave is gonna be the the guy that like he's gonna stay on top of you know the news, he's gonna stay on top of development, he's gonna tinker, which is important. Like give your champions time to to tinker with building things because you know everybody's busy at work all day and you know things get
improved when people have time to to work on improvements. But at at the end of the day I think the, you know, have a champion, but it it can't be a side project. It's gotta be ⁓ resonating within the or whole organization at this point.
Miriam Allred (22:45)
And here we are too like younger tech forward people. I don't want to people to think like we're disconnected from like the reality of the office team. Like we know that the traditional like home care office team right now is like heavily middle aged women like running these these offices. And so what I would say there is find people that have an interest or a knack for this or are tech forward and like, you know, lean into that and develop them. But if that's not realistic, that's okay.
Be thinking about your next hires. Like who are you bringing on and h I I would advise like it's really important to start bringing people into the fold that are tech forward, that have tinkered in these tools, that are comfortable with kind of like innovation and moving quick on this stuff. Because again, it might not be realistic for your current team to become these champions. Like that's best case scenario. But if that's not realistic, be thinking about your incoming hires and making sure that a part of the job description, a part of the conversation is just like
tech forwardness and innovation and capability and familiarity with a lot of these tools.
Dave (Phoebe) (23:48)
Right. No, that's that's a great point. I mean, it's important to bring people into the company. If you don't have folks that are willing to go out and kind of like be on the edge, like you need to start figuring out how to bring those people into the business. and I will say as like a a counterpoint at the same time, I mean we we have users of Phoebe, you know, like thousands of, you know, people in home care offices at this point using Phoebe every day and this spectrum of, you know,
computer skills is is all the way from end to end. ⁓ and we we meet schedulers every week still or coordinators or administrators or you know whatever their their roles are that come in and they're like, there's no way I'm gonna learn this. And it's like, well just spend a few hours with us. We'll meet with you a couple times a week for the next couple of weeks, make sure you get comfortable. And almost everyone always tells us, well it look took me six months to get using, you know, to get used to using my last software platform. And we we have gotten to the point where because AI is doing so much of the work for you in the software,
Like we we like to paint Phoebe as like you know, Phoebe is a tool that you use and but but ultimately it's like self-driving software. Like what used to take you 15 clicks to do something now might be, you know, a simple voice prompt to the screen ⁓ as you're kind of chatting with Phoebe Live. So the barrier of entry to learn AI tools ⁓ is lower than I would think most people think. ⁓ if you're talking about something that's designed for this type of user base, it's still pretty high barrier if you're trying to figure out like, well, how do I, you know, connect all.
my different tools to all these different things. It takes a little bit of work, but if you come into like a curated environment, we we're we're finding folks of all different skill sets and levels and backgrounds and even languages, right? We have users that talk to Phoebe all day in Spanish now and and you know Phoebe you know talks back to them in Spanish and translates all the results to English in real time. So ⁓ it can be pretty powerful to come in and and do this regardless of your current skill sets.
Miriam Allred (25:35)
Yeah, that's a really great point. If anyone can learn the EMRs of the space, this AI stuff is like child's play because we're used to like all these dropdowns and menus and clicking, you know, endlessly. It's like really this is a lot more simple than the the tech prior. So that that's a really great point. ⁓ you've been working with some of like the larger agencies, like maybe in the enterprise space. Are you seeing larger teams dedicating a position to AI or innovation or interoperability? Like, are offices getting to the point where
There's someone that is dedicated to these functions?
Dave (Phoebe) (26:09)
we're definitely seeing it in like the the real like enterprise companies where you might have a kind of a one company or organization that operates five, six, ten different brands of agencies. Like those companies are definitely putting not just like a classic like chief technology officer, chief information officer in place, but like pe like dedicated like head of AI. ⁓ so we're we're definitely seeing that at the like true enterprise level. ⁓ I would say seeing less of it at the at the kind of S and B and even mid market level.
I'll be honest, I d I don't have a strong opinion one way or the other about it. I I just have seen this as an observation. We've we've signed on several customers lately where like the buyer was a consultant working for the agency and they're like an AI consultant to come in and evaluate all the AI tools in the industry and then help them choose which one is right for their business. And this actually kind of surprised me. ⁓ but over the last six months we probably, you know, sold to a third party consultant that was kind of buying on behalf of the agency, like I mean, maybe maybe even dozens of times at this point.
So that's maybe more sensible, right? If you're a team of five or ten internally, you don't necessarily need a full time AI guy, but if no one internally is gonna be that like on the edge champion, it might be worth, you know, the cost to say like, hey, I'm gonna bring in an expert. But you gotta be careful with that because there's so many people out there shilling their expertise on like, hey, let me come set up AI in your company and they charge you ten grand and all they do is like set up a couple of COD automations and walk away.
Miriam Allred (27:39)
Yeah, that's a that's an interesting point. But I think it just shows like we all feel me me myself included, like imposters in this of like I don't feel qualified to know which tool to use and how to use it and where to deploy it and where to like trust it. Like I feel like everyone's in that camp of like, I know I wanna be using these things, I'm tinkering with them, but it's a whole nother ball game to like turn over your operation to an AI. And so
people are seeking like outside help of like help me analyze and help me see where I should like place my trust ⁓ because everybody feels like an imposter. ⁓ let's get technical on like where AI is thriving, where automation is thriving, and where the human touch is still so essential. And I will continue to talk about this to the ends of the earth because I just still think this distinction is so important. But I do feel like it is a moving target. I feel like
Where d the technology is quite literally improving. The processing speed, the quality, the context, it has improved significantly. And so the the goalpost of what should be automated versus what should be human is changing. So in your own mind, we could talk about scheduling and then we could, you know, maybe zoom out to recruiting or sales, but like in your mind, where are some of the lines being drawn as to what
AI is capable of and good enough at versus where the human should still be involved.
Dave (Phoebe) (29:07)
Yeah, I'll I'll I'll try to answer high level first and say if if the if you are kind of thinking about where to have the human involved versus where to just like let the A AI go, ⁓ you know, look at the purpose of the interaction. ⁓ if I think about like
purposes like logistical coordination or reminders or like routine questions that you have to ask somebody every week or ⁓ you know routine information you have to collect to make sure that you know what to do next week. You know, all of all of those things you should be looking at quite differently from things that are more relationship driven or things like reassuring your clients your caregivers or resolving conflict or making sensitive decisions about, you know, what type of care, you know, is is going to be received for a client of yours. Like those are the areas where you you don't
you know, I I would encourage you to use I AI to support you in those decisions, but not, you know, not to kind of let your hand off the wheel. ⁓ if I drill down to like Phoebe specifically, ⁓ we have, you know, most of our customers now, ⁓ after two to three weeks of using Phoebe, get to the point where we kind of go full auto, is what we can what we call it now. That's kind of a lame way to say it, but we'll we'll, you Phoebe can go full auto, which means that it doesn't have to get approval or direction to, you know, identify that someone's calling out or even receive the call out directly.
And then you know, start reaching out to caregivers, find the right caregiver, validate that caregiver is a good fit, and then actually just put them on the schedule and let the client or client point in contact know that they've been scheduled. So you can kind of go full end-to-end auto with Phoebe. And you know, we we feel confident in that because Phoebe's like memory layer and context is so rich and so good that like if the caregiver wasn't a fit for the shift, Phoebe would not have recommended and then reached out to them in the first place. So we're kind of doing the pre-bedding up front, which is different than kind of the mass blast. I think we probably
talked about on the on the last podcast but the technology has gotten so good that it used to take us months to get somebody to to full auto and now we can arguably probably get there in days if we wanted to but we we we haven't quite allowed that ⁓ be because of how fast Phoebe learns and gathers context. so what happens here is like Phoebe is now
taking on the logistical coordination, ⁓ the routine interactions, the routine asks, the clock and clock out reminders, ⁓ you know, checking in, reviewing care plans, just kind of we have agencies that so many customers like pre Phoebe, they would just call the caregivers, you know, every time it was a first interaction with a client and just read them what was in the EHR. so it's like those are things that the AI can actually do quite well for you. And and what that opens up time for is more of those ⁓ kind of other side I was talking about, which are
like relationship driven things, reassuring caregivers, resolving conflict, answering sensitive questions, ⁓ you know, helping a caregiver talk through a challenging shift that they had. and like the beautiful thing about, you know, Phoebe in in AI is, you know, you can have Phoebe handle all all of those, we'll say logistical interactions and open up the time for the, you know, coordinators, schedulers, administrators, owners to actually talk to the caregivers about, you know, how was your experience at your first shift and those things that they everybody wants to do, but at the end of the day, like
Most days just didn't get done before.
Miriam Allred (32:13)
Yeah, one way that I think about this, and this isn't universal, but it's kind of like inbound versus outbound. Like I see and talk to more agencies that are letting AI handle all of the inbound because you think as to why caregivers are calling in, paychecks, schedules, you know, a question of something they forgot, like oftentimes they're checking in about like relatively low yield tasks and topics. And so
what I'm seeing is a lot of like, yeah, inbound call volume is AI as the first l line of defense. But then what you just said and thinking outbound, a lot of the outbound calls are more relational. Like we want to check in with the caregiver. We want to, you know, talk about their experience. We want to ask them personal questions about how they're doing. Like there's kind of this like inbound outbound that I see is is kind of a distinction. Again, I want to say it's a hard line, but thinking about the inbound calls versus the outbound calls.
And starting on the caregiver side, you and I are aligned on our resistance and hesitancy to like deploy this with clients and families. But I think on the caregiver side, they're you know, we're at the point where it's it's a safe bet to get them interacting with AI. And one step bef one step like removed from that is also after hours, you know, after hours and weekends.
I think it's really a safe place to start and tell your employ your caregiver, your workforce, that, hey, we're gonna have agents answering our after hours and weekends because they understand like the nature of a of an office. Granted, 24 this is a 247 business, but I think it's easier for them to digest, like, at 8 p.m. I'm not gonna get a human, I'm gonna agent, I'm gonna get an agent. But if it's important, obviously it can be escalated to a human. But I think starting there, like
Dave (Phoebe) (33:28)
Right.
Right.
Miriam Allred (33:53)
There's just kind of this like crawl, walk, run approach. And I think like starting with after hours and weekends and then working towards inbound, you know, and keeping outbound in the human side, like that's a an approach that like I see and would take myself.
Dave (Phoebe) (34:09)
Yeah, no, I I totally agree. And I think the the the inbound side, like Phoebe, you know, for the last six, seven months, ⁓ we have been kind of answering inbound calls for agencies. So, you know, we typically find the setup is during business hours, Phoebe's gonna answer every caregiver call, but it's gonna let, you know, anybody that's a non-identified caregiver, you know, clients, prospects, you know, families, what have you, all route to the the main line and ⁓ just get picked up by the office as usual. But then after hours and weekends, Phoebe's often picking up a hundred percent of the calls.
And the I would say the the key learning for for us and why we were so intentional with rolling this out is, you know, eight out of ten of the inbound calls that come in, you know, are things that AI can handle very well. But you also have to have your product or your AI able to, you know, be very good at identifying like when do I need to transfer something? Because the last thing that you want is, you know, a family member or a client or a caregiver calling in about a about a distressful situation that's happening in the home. And then
all of a sudden, you know, they they can't get out of the AI loop. And ⁓ you know, we we have spent a lot of time making sure that Phoebe is incredibly good at identifying very quickly and then at getting clarity if it doesn't it's not able to identify very quickly, like why is the person calling? What are they calling about? Who are they? And if it's something that is time sensitive, is critical in nature, is, you know, relationship sensitive or is somebody in distress, like we're very, you know, very quickly getting that person to the right human that should be taking the call.
Miriam Allred (35:37)
Yeah, a hundred percent. Share your semi hot take about like intake and the client side. Like why are you hesitant to go after that side of the market?
Dave (Phoebe) (35:49)
Yeah, you know, I I shared this a little bit at the beginning, like, you know, companies are coming in and saying we do everything. And it sounds really awesome to say, like, we do intake and we'll schedule your assessments and we'll kind of vet your, you know, families and, you know, potential client, you know, family members that are calling in to determine if you should take them and the the reality is right now, and I do think this will change eventually, I don't think that we're there yet, is when you look at like the front end intake process and if I was out looking for care for my parents and I've got, you know, two numbers to call and I call one and I talk to an AI and I'd call another.
And a person tells me about the agency, like I am a I'm I'm fully in I'm fully in with AI, but I I'm I'm I'm in it every day. I I use AI nonstop, but like I'm gonna talk to the person that I'm gonna talk to the agency that's gonna let me talk to somebody. so I think the the front-end screening of like intake is still at a challenging place industry-wise where everybody thinks that they want it because they're like, I never want to miss an intake call. But it's like the trade of like not picking up the phone one time in the month at 9 p.m. to not miss the intake call. ⁓
to like just putting AI in front of all of your intake calls, it's it's still not worth it in in our opinion. And that said, we are working on some stuff like in and around, you know, intake and assessments that that we're expanding into. But there are other ways to use AI to help you with the intake process and help you with automating aspects of the assessment process. That does not mean having AI be the the thing that's responsible for determining if you're going to schedule an assessment.
Miriam Allred (37:16)
Yeah, I I completely agree. And I take that same like personalized approach. If I were calling the agencies and I talk to an AI about my mom that needs care, like there's just no way. Like again, I believe in all this stuff, I buy into it all, but like there's just no way that that would be a turnoff for sure. But here, you know, we'll have another conversation in a year from now and like things might be different. ⁓ but today as it stands, it's like I I completely agree with what you just shared. any other examples of like
Dave (Phoebe) (37:35)
Right.
Miriam Allred (37:46)
over automation right now of where people maybe you've seen firsthand in recruiting or in other areas of their operation where you think people are like taking this too far. Like have you seen that yet?
Dave (Phoebe) (38:00)
Yeah, I mean we we see a lot of customers that will come to us and say, like, I want to
I I want the AI to be able to let me completely disconnect me and every person that works for my agency from this time to this time. and that there are AI companies out there will Yeah, you can do that. but we'll say, you can't do that with Phoebe. you know, you've got to have somebody who can pick up the phone in an emergency because if you've got a caregiver calling in and they've got a client that's fallen and is getting in the back of an ambulance, like we we I mean, I I don't even want to say like we don't want to deal with it because the AI can handle the conversation, but it's like you should step in.
there. ⁓ so I think that like folks taking it too far just saying like I'm just gonna completely disconnect you know from from parts of my business I I don't think that we're we're there yet and I think it takes a lot of
care and attention to design a system that knows when to loop the person in. because, you know, at the counterexample to that is you build a bad AI tool that like every time it thinks there's a problem, it's just like ringing the phone of the owner at 10 PM. And it's like if you do that 15 times in 14 out of 15, it was, you know, crying wolf, it's like that's not going to be a good product experience either. so you've got to figure out and make sure that you're setting your AI and and picking your vendors in a way that are going to be able to make the distinction and make the distinction incredibly well.
I think that, you know, I said this earlier, and this is maybe a lame thing that's not home care specific, but I I think too many people are using AI to write everything for them. ⁓ and it it's just the the receiver of a message that was like totally written and not even read by you, that is just like, you know, pure slop. It's just like disrespectful to the person receiving it. ⁓ and it actually makes things harder to understand. And I think the
It's part of the AI company's marketing, right? It's like you open your email tool rot now and like you almost can't open Gmail without having a draft made for you that you didn't ask for. but ⁓ so it's like it's being forced on you. But the the reality is is like ⁓
It's it's mostly making it where the person receiving your information is like, you know, not gonna be as crisp on the thing that you were asking for. And then secondarily it ⁓ you know, writing is thinking, right? So you you you you write to kind of get clarity on your own thoughts. and if you're letting AI do that for you, I I do think it's taking it too far and again, there's a world in which that changes, but I don't think that we're there quite yet.
Miriam Allred (40:24)
Yeah, I'm glad you brought that up. It's a major turnoff when you can just tell what's written by Ana AI and what's not. ⁓ and as like a content creator, making this personal for a second, like as a content creator, it's like I'm getting tired of social media. Like literally s people have reached out to me, like, Well, you're less active on LinkedIn, you know? Are you okay? Like, what's going on? I'm like, I'm just kind of tired. It's so sloppy. It's so much like garbage and so much noise out there. And I'm just like kind of tired of it. And so for me, it's like,
quality over quantity, like spend time writing and creating the things that ⁓ matter and mean something. And if I'm sending yeah, if you're sending messages to families and caregivers, it's like those messages matter. Like, you know, it's so obvious when it's AI versus not. And so like take the time to do the things that matter. And oftentimes writing a personal email, answering the phone in an emergency, like those are the things that matter. And you should not even be thinking about automating that stuff away. to your s to your first point though,
I've talked to a couple of new owners getting into this space that are on kind of like the younger tech forward end. They have come to me and said, like, hey, can I build an agency like without an EMR, like straight AI, like end to end? Like the people are pushing the boundary of like how far and how lean and how tech dependent can they be? And I'm like, whoa, whoa, whoa, like
That's a misconception of like you feel like you can just like take some shortcuts right now. Like a lot of home care is still manual, it's still personal, it's still boots on the ground marketing. Like there's no shortcuts. AI can't solve and automate away a lot of the just like the effort that this business takes. And I it's almost like misleading of like, man, I think people wanna try and like cut corners where they shouldn't be. And that's discouraging.
Dave (Phoebe) (42:09)
It it is. I mean, and you know, we for for for what it's worth, I mean, you you probably c could ⁓ do about, you know, eighty percent of the work pretty easily to, you know, run an agency on pure AI without an EMR. But, you know, it goes back to what we were chatting about a little earlier of like you you know, you don't wanna be unless unless you're like, you know, have the capital to hire the people to focus on the technology exclusively, ⁓ you know, you don't wanna be responsible at this point for being the
you know, something goes wrong and now all of a sudden, you know, your business can't operate. ⁓ so that's that's still the distinction I would draw. We hear, you know, in tech world all the time people are saying, like, I canceled my, you know, two million dollar contract with s you know, so and so database or, you know, so and so CRM and I've rebuilt it myself and, you know, silently, you know, right now most of those guys are back on the off the shelf software a few months later. ⁓ and again, don't get me wrong, we spend probably
you know, twenty, thirty percent of our company time, ⁓ building internal tools from scratch that, you know, you used to buy off the shelf, but your your core system of record is probably not the right decision.
Miriam Allred (43:16)
Yeah. Yeah. Exactly. Kind of a one off question. Think of like the the tech adoption like bell curve and I'll kind of state this. I'm curious where you think home care is right now. ⁓ there's like innovators, early adopters, early majority, late majority, and then the laggards. I'm curious where you think home care is right now. Like are we at what stage of that bell curve would you say the industry is?
Dave (Phoebe) (43:40)
Yeah, I, you know, maybe, you know, controversial, but ⁓ I I think we're we're kinda still sitting in that like, you know, early adopters, you know, part of the bell curve. Honestly not not even in the early majority. You know, one could argue and maybe I could argue that we're we're in the early majority at this point, but ⁓ it seems that way when you're on the internet, ⁓ because it's like that's all anyone's talking about is AI and their home care business and ⁓ the but but the reality is just like
Most people don't just do their home care business on the internet, like they're doing it out in the in the real world. And we still talk to owners every single day that are like, I've never heard of Phoebe. I've actually didn't know that anything like Phoebe existed or, you know, I, you know, my my kids use Chat GPT. I've never tried it. And like that is changing, don't get me wrong, but we we definitely are in the early, in the early adopters phase or past, you know, early majority phase of like general AI use. But that this is
the number of people that pay for AI is still insanely low. ⁓ so like even though everyone's exposed to it and using it, the number of people who are like really adopting is still very low. So I would say like maybe crossing the kind of chasm from, you know, early adopters to early majority, but I I think we're still in the the early majority side or early adopter side of things.
Miriam Allred (44:55)
Yeah. Hear you explaining that, I I think I agree. ⁓ and you also think of like this technology relative to a larger timeline, like we're still in the early stage. Like, look how far we've come in the last two years. But if you think about five years from now, eight years, ten years from now, like there is still so much to come. And so I guess if you think of like a I'm thinking of like kind of like a ten year timeline, I really feel like we're still in like kind of two to three. And so being in the early adopter phase, like from a ten year timeline.
actually makes a lot of sense. So I I completely agree. And I actually love that statement of like if you're on the internet, it feels one way, but if you're not on the internet, it feels a completely different way. I think that's super, super accurate. ⁓ that being said, do you feel like agencies are getting left behind? Like it's really competitive and it's really saturated. And so like think of think of a market. You're down in Florida. Like think think of a market that's like really saturated. Companies that are using AI, like
Dave (Phoebe) (45:37)
Yeah, for sure.
Miriam Allred (45:54)
At what point are we gonna see people getting left behind or people closing their doors because they're not keeping up?
Dave (Phoebe) (46:03)
Yeah, it's a it's a great question. I mean, we I would say we're not at the point where like the the people who are still lagging ⁓ are fully being left in the dust. I think that there's enough to improve still with AI, ⁓ where like you can still run a competitive business ⁓ if you're not kind of
Sipping sipping the juice of AI, right. ⁓ you know, you can still hang. ⁓ you know, we're at a point with Phoebe where we very confidently tell people like, you know, bring Phoebe in, we can get your schedulers from, you know, you know, twelve to fifteen hundred hours, eighteen hundred hours a week of management to, you know, three thousand, four thousand, five thousand. But we're we're kind of still sitting even from a product perspective and ⁓ you know, Phoebe's adoption, you know, can get you two to three X kind of more ⁓
work done or or we'll get you two to three X more time out of your your current scheduling capacity. But if I fast forward to let's say 12 months from now or 18 months from now or 24 months from now, like I think that is gonna two to three X in itself. So we're gonna go from, you know, Phoebe making you two X more productive to 10X more productive and then eventually 20 X more productive. So ⁓ I'm kind of mixing topics here, but even if we're in the like later phase of the curve than I think we're in, the technology itself is still in the early
early, early stage of what's actually going to be possible here with with AI. and I think that
The answer will change by the next time we talk. I think if your agency has not made moves within the next, you handful of months, like you're gonna reach a point where you start to get left behind. ⁓ we have examples of customers on Phoebe that, you know, basically were growing 10%, you know, year over year for the last three years. And in their six months on Phoebe, they've doubled the business. So things like that we're we're seeing happening. ⁓ and those are folks that come in and they just like fully embrace it and they fully dive in. So we're not, you know, be clear, not all
of our customers have doubled since they've started Phoebe, but we we've seen that like if you just kind of like lean into this new way of working and lean into the technology, that it is possible. And the percentage of people who have figured that out is still very small. So once you know that shifts into being a larger percentage of the market, people are going to need to get on board.
Miriam Allred (48:14)
Yeah, you're talking about like growth margin. I also think of like net margin, like net profit and profitability. Like again, the like AI is like a productivity multiplier, getting your team to be 2x more effective, 4x more effective, 10x more effective. That is improving your bottom line and improving your net profit. Therefore, there's a lot of agents, we talk about this all the time, like that just plateau that hit these walls and they just like stop growing. The companies that are using AI, they're going to be able to like exponentially keep growing.
And doesn't necessarily mean they're gonna put the bad actors out of business, but it's like they're just gonna keep getting further and further ahead. And so maybe just like the gap in home care is gonna become wider. Like it's already a pretty wide gap, I guess I would I would say. ⁓ but maybe with AI that gap is just gonna continue to widen and then time will tell if it's actually gonna like put people out of businesses or just put people like w even further ahead. That's maybe I'm just like thinking out loud, but like maybe that's one way of thinking about it.
Dave (Phoebe) (49:11)
Yeah, it's very true. And and the industry in its own right is growing, right? So you could we could get to a point where like purely based off the fact that the kind of
AI adopters are embracing the new way of working and embracing A AI. And like those are the ones that are capturing the new market share. And it's not necessarily killing the businesses that are slower to adopt because the the actual market is growing. So, you know, I I think we're gonna see this this very interesting thing where the the pace of how far people are falling behind is is not gonna be realized as quickly as it will in other industries because you know there there aren't many industries that are gr growing at the pace that that home care is growing. So that's gonna it's gonna shift.
Yeah.
Miriam Allred (49:53)
Yeah. What are mistakes owners, operators are making when they're implementing AI? Like you have helped implement AI in these businesses. What are what are the mistakes they're making or what are the misconceptions that they come in with that you want to just like say out loud right now so people stop making them or stop thinking certain ways?
Dave (Phoebe) (50:12)
Yeah, I will say one from a selfish perspective that we've learned a ton about. I would say like, you know, b buying for the sake of not wanting to get left behind versus purchasing for an actual bottleneck or pain. we have brought on some customers that we've, you know, ultimately made recommendations to pause with as we're doing our kind of ROI reviews because it's like they're at, you know, not they're they're they're basically like, you know, 2x capacity on their human team for what they need for the volume of care they're providing. And, you know, folks like that.
Come in, they're excited, they're excited to see their ROI, but then it's all of a sudden that the people on the team are like, Well, I actually didn't like really need any help because I'm I'm only like busy half of the day, anyways. and a lot of owners, for good reasons and businesses, like they want their business to grow and they want those people to work for them long term. So it's not like, okay, we'll just get rid of get rid of an employee. Like some people do that, but you know, a lot of people don't. So I would say when you're when you're implementing AI, you you want to make sure that you are not just buying the the tool for the sake of buying the tool. You've actually sat down and figured out.
Like, where are the bottlenecks in my business? ⁓ and this goes back to what we were talking about before. Like, everybody should be building. So, even if you're not at a point where you know using a tool like Phoebe is going to solve your core pain, like figure out your core pain and figure out how to solve it. And that might mean building something yourself. And then as your agency grows, you grow into the the kind of realm where it makes sense to adopt, you know, something like Phoebe. I'll say one more thing that has also been a big learning experience for us that has changed the way we we sell and how many people we say yes.
S2 is, you know, for folks that, you know, want to come in and just kind of leave their team fully out of the decision and then just kind of like top down force this. it it's it's not gonna work. ⁓ so I don't get me wrong, I am I I can be a proponent of like leadership making decisions and kind of like enforcing those, you know, down to the company, but it
for for this to work well, and this is gonna go for any AI, not just Phoebe, like you've gotta have buy in from the people on the ground. so just being thoughtful to not show up like surprise, we have a kickoff with Phoebe today. I forgot to tell you about it. Like that's probably not gonna go very well. And it's it's surprising to me how many times that's happened. ⁓ but I I think we've course corrected on that in the way that we sell.
Miriam Allred (52:27)
Yeah, one more that I'll add to those great points is having worked at, you know, Home Care Poles Careswitch, like that's no secret. Like I was building these tech products as well. is that the product that you buy, the day that you close, the day that you implement, these products are moving fast. And there are a lot of changes, a lot of new features, a lot of new functionality. And it is a little bit hard to cat to keep up. I will not deny that. But go into this knowing that you are going to grow with these platforms.
So provide feedback, but also we were talking about this before. These tech companies put out product releases, feature updates, like make sure you are you're in the mindset and training your team to keep up with those things. Because again, you might buy something and be like, this is great, and then we get used to it. But this technology is moving so quickly. These these features and functionalities are moving so quickly, which is a good thing. And again, it's hard to keep up, but stay up to date with the product and the changes because
That's a part of this whole process. Like you said, people are you know, signing up for these tools maybe out of like FOMO, but it's like, no, sign up to solve a problem, but then understand that the way in which we solve the problem is subject to change in a good way. And so stay up with the platform, read the product updates, know that in a good sense, like these these companies are evolving and the the tech is evolving and
you can get left behind in the tools that you're using. Like we talk about this quite a bit holistically of like underutilization platform and product underutilization. Like think of these EMRs. Like people are using them like fifty, sixty percent. Same thing with these tech companies, with these tech tools, but it's like make sure you are staying up to date and evolving and growing with the platform itself.
Dave (Phoebe) (54:15)
Yeah, no, that's a that's a fantastic point. And it's like classic software statement, but you know, it you know, historically, you know, you measure it, eighty percent of the value, you know, comes from twenty percent of the features and you know, the the pace of development.
For software and AI tools or using AI has just exponentially increased the amount of new feature updates and new value points that a software company can provide. So this is on us too. I mean, we we are still learning, like we're building so fast and we're shipping new value points so fast that we we have not figured out how to make sure that every customer is kind of aware of the new value points because the the the 20% of the product that you're getting, 80% of
value from like I think that's changing now because the the the product can be s kind of so specific and tailored to your specific business versus like everybody kind of using the exact same thing and clicking the exact same buttons ⁓ that you know to to keep up with the the pace of the product it it requires real work right now both on the part of the software company and on the part of the customer ⁓ and we're trying to figure out better ways to to help the customer not have to work so hard to keep up.
Miriam Allred (55:19)
Yeah.
Yeah. And that's just back to like, you know, to be a good partner like takes one to be one. You know, like as agencies, these tech companies like want feedback. And so I just, you know, I'm an advocate for of like be vocal and provide feedback and ask questions along the way. Don't feel like, you know, you're buying something off the shelf, but it's like they want feedback and they're growing and evolving and like be be vocal. ⁓ just a couple more questions about Phoebe as we close here. I want you to more explicitly state, ⁓
As to why Phoebe has stayed in the scheduling lane. Like you and I both know how fast this industry's moving and how many people are coming on the scene. And I've even said, like, hey, are you guys, you know, where are you at with recruiting and with sales and all these different functions? Because there are so many companies saying they do it all. But I want you to share why Phoebe is doubling down and staying focused on scheduling intentionally. Like, can you just answer that directly?
Dave (Phoebe) (56:11)
Yeah, I mean the the the reality is, and I you know, I've probably said this to people before if you've ever talked to me directly, it's like it's very easy at this point, like incredibly easy at this point to make a product that does something like Phoebe does, ⁓ and handles
The situations well, let's say 80% of the time. ⁓ so it's not like, okay, I built my 80% and I just have 20% left to go. Like that 20% ⁓ is insanely difficult. All of the edge cases, ⁓ all of the you know, things that need to be a part of the product and learning experience ⁓ to make you know something not just be 80% right, because if you you know mess up with 20% of your caregivers or your clients, like your business is not gonna do very well. So we've stayed focused in you know scheduling because it is the hardest.
part of the business and it's also the operational heartbeat of the business. And you know, we feel very passionately that ⁓ the about that about that final 20%. I would say we're getting to the point where we're comfortable starting to expand. We've got, you know, Phoebe now, you know, making tens of thousands of of
phone conversations a day. I mean Pho Phoebe's talking to tens of thousands of caregivers on a on a on a daily basis at this point. And all of that data input is what allows you to solve for that, for that last twenty percent. ⁓ and if you think about things like the the way an agency runs effectively, right? Is they're able to
focus on or they're able to do a very good job at understanding like who to call, when to call them, when to deploy them, where to deploy them. And to do that right, there are sometimes thousands of inputs every single day that need to be updated in your EMR to make sure you're keeping on top of that. And the reality is is like 10% of it makes to the makes its way to the EMR, you know, 10% of it makes its way to a notebook or spreadsheet or you know a sticky note. And the other 80% just lives in the heads of the people who do the work every day. so
Do building a system that like can take that 80% of the context that, you know, historically just lives with the scheduler who has to be on call twenty-four-seven ew even when they're not working and and getting it where the technology can real time recognize what to remember and then actually remember it and then actually use it in its execution of, you know, the scheduling work that it does, it's it's just incredibly difficult. so we've got a sizable team of, you know, the best engineers we can hire that that work on this day in and day out.
Miriam Allred (58:31)
Yeah, I like it. And I respect it. And I guess one way to think about it this is kind of extreme, but it's like if you can s if you can truly solve scheduling, like to the best of your ability, like everything else is relatively easy. Like there's billing complexity, there's, you know, assessment clinical complexity. But I I agree with you in that, like if you can really build something strong enough to handle the edge cases of scheduling.
everything thereafter would be like relatively easy. And so I respect like the time and dedication and the intentionality of focusing on scheduling ⁓ with the the plans to expand. You know, it's like it's just a matter of time, but it's like let's get really good at one thing. And I back to what you said a minute ago about agencies buying technology to solve a problem. Like if you buy a solution that solves 10 problems, like like there's only so much value you're gonna get out of that. And so like
Buy something that does one thing really well and you will be impressed by like the result of the singular thing rather than like the lack thereof across multiple things.
Dave (Phoebe) (59:35)
Right. Yeah, and in that said, I mean, I feel like I should take the opportunity to say like we we've stayed focused on scheduling for for such a long time ⁓ for good reason. but we are getting to the point where we're feeling in insanely confident about the the quality of the product and still bugs from time to time, but you know, we we work through those and we improve and ⁓ we get them solved. But we we are, you know, now actively moving towards other avenues of the business. And we talked a little a little bit about or hinted at some of the stuff we're doing around the front side of, you know,
client assessment and you know care plan generation and intake and then there's the more midstream, you know, areas like recruiting and credentialing and onboarding and kind of ongoing compliance management. These are all things that we're now actively building. ⁓ you know, using a lot of the inputs that we've gotten from staying so focused for so long. ⁓ and we'll be rolling those things out ⁓ with with some customers actually pretty soon.
Miriam Allred (1:00:27)
Yeah, super exciting. Any last like words of caution to people that have not bought something yet, but that are like researching and maybe in the sales cycle? Like any words of caution or questions that they should be asking as they're talking to several tech companies?
Dave (Phoebe) (1:00:45)
Yeah, I mean I you know, I I w we were a young company with no customers once, so I almost hate saying this, but it's like I do think we're at a point in the industry where there are enough people who've been here for long enough ⁓ that are AI native, that kind of have been building AI from day one that like you you should be asking people, like you know
How many people are using this? Like, what are your results like? What are the KPIs that you track? You know, tell me about a specific agency that you were able to take from point A to point B and like what were the differences and how did they get there? Like you should be asking specific questions about
performance. and I am, you know, not a person that that necessarily thinks that like a new company can't come in with no context and build a great product. Like I I actually think the best products are built by outsiders who are actually just good at solving problems and not necessarily like insiders who have kind of bias on certain industries. So there are is plenty of room for more people to come in and build great products. ⁓ so I I'm very sure that more competitive folks will come on the scene, but
I would heavily encourage owners right now to be making sure that if you're implementing a product that is in a you know part of your business that would serve as a point of failure, like scheduling or, you know, intake assessments or even recruiting to some extent, like you should be making sure that the the company can can give you some statistics and proof that they can deliver. that's that's probably the the main thing. I think that ⁓ you know caution, caution wise, you know
Folks should be should be asking like what the support experience is like, again, with specifics. ⁓ like who do I call when I have a problem? How do I get a hold of you? How fast are you gonna respond? Am I gonna talk to somebody who works for you, or am I gonna talk to somebody that's kind of, you know, outsourced? ⁓ you know, how do you actually run? Yeah, or or or an AI. I probably said this last time, I'm not sure if I did, but we still today ⁓ there is no AI between us and our customers. There's no AI responding to
Miriam Allred (1:02:30)
Or an AI. Is your support team a full AI? ⁓
Dave (Phoebe) (1:02:44)
There's no AI answering emails, there's no AI picking up the phone. Like that's just like if if you're if your ⁓ if your customer service experience when you have a question is to go talk to an AI chatbot, ⁓ that's probably not the company you should be working with.
Miriam Allred (1:02:59)
Mm-hmm. Which sounds funny, but that is very true. And so I I say that because I would ask that to every company that you're talking to. Like how what is your support process? And is it humans or is it AI? Because they're all running so lean. Most companies are running so lean that they're dependent on AI as well. ⁓ just last question in closing, like anything you want to tease for what's coming in the new year and the next like three to six months, something that you're excited about or something that Phoebe's working on that you wanna tease or talk about?
Dave (Phoebe) (1:03:26)
Yeah, sure. I mean I I talked about some of the stuff we're expanding into around like
you know, recruiting and credentialing and onboarding. we've got some like active effort there, also enhancing kind of some of the client side interaction that we're doing within the scheduling realm. ⁓ I also mentioned we're starting to look into what is more of like the I'm not gonna say intake because I'm still, you know, as we talked about, not quite sure that, you know, the the AI should be, you know, scheduling your assessments for you in the absence of talking to a human. But ⁓ we by the time folks are listening to this, we'll have announced that we've acquired a a ⁓ competitor called Sage, Sage Care for
folks who are familiar, we're very excited about bringing in the team. These guys have been focused on the client intake side, you know, ⁓ intake automation, ⁓ assessment automation, care plan, you know, automation, ⁓ some AI receptionist work similar to what Phoebe does, but ⁓ very excited about, you know, bringing this team in because they have spent so much time learning about like, okay, what do people say they want versus what, you know, do people actually use and what are people actually willing to tolerate on the client side. So as we've gotten the scheduling to
a point where we're, you know, now, you know, getting the the you know data to say we should start expanding. one of those ways of expanding is to bring in some experts who've done nothing but think about this for the last few years. So ⁓ excited for that. ⁓ I think by the time folks are listening to this that'll be public news, even though I told you today it's not quite quite released yet, but the the the ⁓ the deal is closed and we're we're working on ⁓ making it public.
Miriam Allred (1:04:48)
Yeah.
Yeah, super excited. Well, I like you, I'm just like, you know, I'm not building these things day to day, but on the edge of my seat and just like trying to keep up with where everything is headed and how fast we're moving. But I'm just excited, you know, like fired up as to what's possible with home care, like what the industry's gonna look like a couple of years from now. I feel like we've come a long way the last couple of years and things are just really starting to like get exciting. And so I get excited about like where this is all where this is all headed. the the last thing I was gonna say as just kind of like a piece of advice as obviously
you and other companies are like, you know, selling products. But I would also advise owners, operators listening about listening to this and thinking about this, like see these guys also as like consultants, ⁓ and like tech partners. Like, granted, you know, it's not it's not a secret that they're not gonna try like sell you something, but go in asking them like hard questions and like ask them questions about your business and about your processes and about automations. Like use these guys as more than just ⁓
you know, like sales reps see them as like consultants and people that you can bounce ideas off of because like we said before, everybody feels like an imposter. You know, it's like intimidating to talk about these things and know where to deploy these things. But really like we're all just getting educated and learning together. And so see these as opportunities to talk to these folks to like learn about your business and ask good questions. And even if you don't buy their products, like
It's a learning opportunity for everybody. And so just like kind of like that mindset shift of like I would be talking to every company, you know, like I would be asking all of them these these same questions, but using it as like market research for your own business and like gathering intel whether or not you're ready to buy now or in a few months or even in a couple of years.
Dave (Phoebe) (1:06:38)
Yeah, that's a that's a great point. And genuinely we we welcome that. ⁓ you know, we've we've started to expand our sales team. You know, like last time we talked, I was the only one out, you know, selling and we've got a half dozen or so folks that are out, you know, having sales conversations every day now. And we have conversations literally every day internally of like, Hey, I talked to this prospect, you know, this is what they're dealing with. ⁓ should I try to push it? And like
Usually if the rep is asking, you know, me that and I'll sit down, I'll look at it and I'll go over the notes and we'll say no, but you should tell them like it's not a good fit and this is why and like this is maybe the thing to look out for in the business, that would be a good point to kind of re engage the conversation and you know the you know, that's that's hard for to do as a tech company, right? 'Cause oftentimes a good sales guy can get the the wrong person to sign the contract. But we're we're not in the business of that and we're we're not afraid of telling you, you know, hey, this isn't the right time.
Miriam Allred (1:07:27)
Yeah, I completely agree. Well, Dave, this has been awesome. Thanks for joining me in the lab. Thanks for being a great partner and friend and advocate for all this technology. Again, like I I think it's just a really exciting time and I'm excited to be partnered with people like you that are pushing the boundaries. But honestly, just like helping good people do better every day. Like honestly, that's the mission of technology is how do we how do we leverage this stuff to help with the operations, to help with the dual entry, to help with all of the manual tasks so that we can literally
improve the quality of care and the experience for our caregivers and I believe that Phoebe's doing that. And so thank you for being a great partner and being here and ⁓ awesome conversation.
Dave (Phoebe) (1:08:02)
Yeah, likewise Miriam, I appreciate it. And I would encourage everyone to think of ⁓ make sure you're not a afraid to reach out to Miriam either. ⁓ she she she knows a lot and maybe more than she leads on and has been a great kind of partner and ad advisor to us as well. So ⁓ it should go both ways.
Miriam Allred (1:08:17)
Awesome. Thanks, Dave.
Dave (Phoebe) (1:08:19)
Awesome.