Episode Transcript
[00:00:05] Speaker A: Good morning, everyone. Today I want to talk about AI technology in recruiting. In particular, I want to talk about conversational AI in the context of a company that you may know called Paradox. Paradox was a pioneer in the early days of what we used to call chatbots and developed a system for conversational AI for candidates. And the original vision was to just make it easy, easier for job candidates to get information as they apply for jobs in your company. And they were very successful at that. And Olivia, which is the name of the chatbot, became embedded into lots of applicant tracking systems and candidate websites. And sure enough, being a creative, entrepreneurial company, they built out more and more and more. And now Paradox is really an entire conversational AI platform designed initially for recruiting in all aspects of the recruiting. And you'll hear about this from Adam. There's a podcast I'm going to start from Adam Godson, the CEO. And the reason I think it's interesting is not only do they have an amazing product and they're growing like crazy, but they're essentially proving the point that an AI centric system will grow to replace a transactional system. The applicant tracking market is very old. It was around in the 1980s. The first one that I ran into was Taleo, and I'm sure there were ones before that. And the purpose of the ATS was to track applicants. It wasn't to engage candidates, or to find candidates or to assess candidates or any of that stuff. It was basically an ERP tracking system. And then companies started to add all sorts of things on top of it. The presumption being just like in HCM, that we'll take the transactional core system and we'll add advanced features on top of it to make it more and more useful for candidates, for recruiters, for hiring managers, for internal candidates, and so forth. I think AI sort of proves that that business model is now about to be reversed. In other words, if you start with the candidate in this particular case, or start with the needs of the recruiter, and build a system that automates all of these messy, unproductive, transactional things, you have to do, scheduling an interview, doing an assessment, answering a bunch of questions about what to wear to an interview, and when to show up and how to get in the front door, you can then build the transactional stuff later and end up with a system that is radically better and radically different from the transactional system that had AI bolted on top. And that is basically what paradox has done. I don't want to get into the comparison of paradox versus all the other atSs. But when you talk to the paradox clients, which I have, they are doing things with paradox that they never could have done with their traditional ATS platforms. So let me leave it at there and let you listen to Adam a little bit, and I'll wrap it up at the end. Thank you. All right, everybody, I've got an interesting conversation we're going to have with Adam Godson, the CEO of Paradox. Adam, why don't you introduce yourself and tell us a little bit about what your role is at Paradox first, and we'll talk more about the company.
[00:03:26] Speaker B: Sounds great. Josh, great to be with you. And excited for a great discussion today, somewhat newly, the CEO of Paradox had been in the recruiting industry a long time. So 20 years building recruiting processes, sat in the seat as a recruiter, had the experiences from that. Built systems, have used just about every piece of recruiting technology you can imagine. And the last four plus years, I've been running product and engineering at Paradox, and then I've been the CEO here for about three weeks.
[00:03:55] Speaker A: Well, congratulations on being the CEO. We can talk more about what that's like.
But first, let's talk about Paradox. I think a lot of people think, or don't even know maybe what paradox does, and they think it's a chatbot and sort of a little tool, but it's much, much more than that. Can you tell everybody a little bit about the history of the company and where you guys are?
[00:04:15] Speaker B: Yeah, absolutely. So conversational AI for recruiting is the right angle there. And you're right that there are many conversations that happen in recruiting, and some of those are from where we start was sort of simple questions and answers on a website. All those questions that went unanswered about what do I wear to an interview and what's it like to work there. We started to answer those with conversational AI starting in 2016, and from there realized that conversational AI is actually a transformational technology, and that could help automate so many of the things that are small conversations that happen in recruiting processes. Hey, can you move your 02:00 so I can do the interview at two? Hey, do you have happen to have this kind of certification or license? All the things that just take time and energy from teams and being able to automate that. So today we serve clients in a variety of ways, all the way from having our own conversational based applicant tracking system that's used lots for volume recruitment at restaurants and retailers, all the way to being on top of major systems around the world, workday, SAP and oracle, and helping clients in hundreds of countries, schedule their interviews, screen candidates, and get what we call the boring stuff done so that people can focus on conversations with people and not software. And that's actually the paradox for which we're named.
[00:05:34] Speaker A: Well, let me go back. Let's talk a little bit about the technology, and then I want to talk about all the different applications and use cases. So when you guys started, I know you weren't there at the very beginning, but you've watched the technology for a long time. I guess it was kind of called AI, but it was kind of not the AI we know of today. Right. So what is this?
[00:05:53] Speaker B: Yeah, that's right. It's interesting. So from 2016, there was sort of this early era of chatbots that used natural language processing, which, when you look at it, is a subset of AI.
But today, you wouldn't really recognize that in many ways as intelligent. It sort of trained models that mimicked human conversation, but a lot of the time, we all use them in customer service or other applications. You get a does not compute. Right? I can't answer that, etcetera. And we have some of our clients that still use those models as well. Thinking about the way the world has changed in the last several years with technology, and particularly large language models has really begun to change how conversational AI is done.
One of the core principles for us, though, is self determination from companies. Some companies are very comfortable throwing the anchor forward and say, let's use large language models and let's use generative AI to make all sorts of interesting advances. Many other companies have lots of lawyers and lots of it, security and privacy folks that are going to want to run things through the wringer, and we welcome all types in that realm. But thinking about companies are going to be at different stages there. Some companies are going to focus on really deterministic models of. We wanted to always say this when answered this question. Others are going to use different methods to be able to generate answers.
[00:07:17] Speaker A: If I understand the way this worked, Josh explained it to me. I believe the original model was more deterministic. It was a whole bunch of tables. If they asked this answer, this sort of like that, I guess more sophisticated. The newer one is actually generating answers from content. Is that correct?
[00:07:34] Speaker B: Yeah. We use a technique called retrieval augmented generation to be able to give a framework to that. So it's really just giving structure to those responses to say, this is the answer, and if we have the answer, we will give that answer, and then lots of flourishes around that as well. So thinking about what's the context of this conversation.
Is there a point to insert empathy? Can I use an emoji? And emojis are actually really hard because if you get them wrong, you get them really wrong. And so the quality has to be really, really high there. But it's an important thing for us to do as well because those add the feel to the conversation. And we measure our conversation quality in lots of ways. One way we do that is when people say thank you. And you wouldn't say thank you if you knew it was not a conversation subconsciously. And the feeling we want to produce is someone is consciously knows they're having an automated conversation, but subconsciously feels like it's a real person. And that's how we, one of the ways we measure that.
[00:08:33] Speaker A: Well, after talking to many of your clients, I know it's a very, very sophisticated system. But again, just one more thing on the technology.
Given the fact that llms are sort of a commodity already, very easy to get your hands on one, and you can ask it questions and it sort of answers reasonably well, what makes paradox different from me taking all my recruiting docs, sticking them into chat GPT and asking it a bunch of questions?
[00:09:03] Speaker B: Yeah, it's a great.
[00:09:03] Speaker A: Now there's a lot of, I really want you to explain it to people so they understand the sophistication of what you guys have built.
[00:09:09] Speaker B: Yeah, absolutely. So I think the first thing is going to be about context and about the specialization and the pre training of our models. So we use open source, large language models and we have significant effort into pre training them. We've been pre training those with our tens of millions of conversations we've had over the last seven years, all in the context of recruiting. To be able to use that data to form the pre trained model that is specific to this context. If someone asks what are the benefits?
We know they're probably talking about employment benefits related to a job. And we can talk, we can break that down into 30 different other subsections of follow up questions like, are you talking about insurance benefits or time off benefits? Whereas if you ask someone, just that string of words may mean any number of things. So the pre training of that model matters a lot. The second is safety is being able to save your candidates, most often when your users from the trouble of having lots of embarrassing moments, potentially of a model that's not built for purpose. And so whether that's you asking it to talk like a pirate or asking, can I take my dog to work? Okay, great. Here's the pet policy. Can I take my alligator to work and being able to.
[00:10:22] Speaker A: Can you ask Olivia questions like that?
[00:10:25] Speaker B: You absolutely can. Yeah. I'll share some of that with you. To be able to give the right context, and that's where the magical difference is. Right. Where in an old NLP world, the can I take my alligator to work? Gotta understand, or here's the pet policy now be able to say that. Say, well, we appreciate the question.
We know that an alligator is not inappropriate for work based on the underlying language model. And so I think those are the two key things, is safety and then the pre training of the model. That makes a huge difference.
[00:10:54] Speaker A: Well, let me ask you about a third one. The retrieval part, the search part. In our system, which is not a conversational system, so much, the big key is searching and finding the right content to answer the question.
In your case, you've gone from a candidate conversational bot to an ATS, essentially, and more.
What is the data? How do you manage all the data behind this? I'm curious. And does it remember who the person is when they come back and ask more questions?
[00:11:26] Speaker B: Yes. Is the answer to that. So remembering the person and the context of that conversation, and that's been one of the best advancements based on large language models, is being able to understand the full context of that. And so the data is managed in our systems to be sure that we can understand the context of when someone's answering a question as well. So even my earlier example about benefits or time off, it might matter where in the process you are to that as well. What are the benefits to this job? I might give you an answer with a level of specificity in the recruiting process. And once you're an employee, I might give you a different answer that says, here's where to go, find more information about that, or here's the provider to call, because you're already in our other systems in that. So having the context of not only the previous conversations you've had, but also where this person is in a process. Have they scheduled an interview? Have they already had the interview? Have they already been hired? All that matters to the type of conversation we can have.
[00:12:27] Speaker A: Okay, so there's essentially a workflow management system going on behind the scenes here that's keeping track of what this person's been through. And then when somebody buys paradox, I've never done an implementation. Do you coach them through? What do you have to do to teach it about your employment practices, your hiring practices, your culture? I mean, there's a million questions recruiters have to answer. I don't know where that stuff's even written down, to be honest.
[00:12:54] Speaker B: Varying levels based on what companies have. But our baseline where we start is we asked our clients to onboard Olivia and about half our clients change the name to some other Persona, but to onboard like an employee. So if you were onboarding someone, you have all the information that you would have about. About all the things to work there. Where's your. The handbook and the manual and the benefit policies? All those things we ingest that we use AI to rip that apart, put it in vector databases. But importantly, with the retrieval augmented generation model, we also allow for review of that from our clients. And so if there's going to be an. For what's the culture like, we're going to have an answer that we generate through LLM. But we also want someone to be able to have the approval to make that different. If you want to have that answer change, then the base answer.
[00:13:46] Speaker A: Does the client get a tool to see how the answers are running and then go back and tweak them? Yeah, yeah, exactly. Right.
[00:13:53] Speaker B: So we're going to make the base answer from the LLM and then the client's going to be able to review those and then also change any of those, update those as well.
[00:14:00] Speaker A: Okay, so in a sense, what this is, it's a very sophisticated domain, knowledgeable chat system built on AI, not a chatbot. In a sense, it looks like a chat bot, but it's very sophisticated in this domain.
[00:14:15] Speaker B: It is. And that's why we like to use the term conversational AI, because that world has changed what it used to mean versus what it means today. And then I think some of the things you mentioned, too, about workflow actually matter a lot to their ability to get work done. And that's one of the things that is so important in our space right now, is, yes, everyone wants to know what recruiting is going to look like in five years, but they also have problems today.
How do we both have that vision for the future, but then also be sure that we're getting work done today, we're improving outcomes, we're getting interviews scheduled, candidates screened, getting things done.
[00:14:54] Speaker A: Well, the companies I've talked to, that use paradox have reduced the workload to recruit by almost orders of magnitude. Really. I mean, it's just sort of astounding.
And I know you acquired an assessment tool. You've built out an ATS. In a sense, you're getting into candidate relationship marketing. Tell me a little more about the advanced stuff that's in here. It's more than a chat?
[00:15:18] Speaker B: Yeah, that's it. Well, and I think the idea of where the future is going is that recruiting is going to break down into really two types of people. There are going to be people that design processes and people that manage technology and really think about how to make technology, especially conversational AI, but also tools that help with assessments and augmenting decisions and other things. How do we get that work done? And then there's going to be people that have a lot of conversations and convince people to join.
And I think that's really good for the future of talent acquisition in that today, I think there's a lot of people that wish they were having conversations all day. They convince people to join. They're actually spending a lot of time clicking in their ATS or doing administrative work. And I think the future breeds a lot more clarity for that to say. People are going to be convincing others to join, and the competition is going to be fierce for talent in the future. And then we're absolutely going to need people that are real recruiters. Right, right.
[00:16:16] Speaker A: And not people doing administrators, human centered recruiters. Right, exactly.
[00:16:20] Speaker B: People clicking around so much of the administrative work will get abstracted away. And paradox is on a mission to help do that in lots of ways. And I think one thing that's unique about our toolset is we can work with just about anyone, whether you're a large multinational company and want to put us on top of workday, or if you're running restaurants in a city and want us to help with the entire process. So there's a tool set that's built flexibly well.
[00:16:44] Speaker A: So I've talked to McDonald's and compass group and a bunch of your clients. They tend to tell me things like the number of recruiters went from this big to this big, much reduction in the number of especially things like coordinators. So what happens to the recruiter's job when you turn this thing on?
[00:17:01] Speaker B: Yeah, mostly they spend more time with people and not software, and that's it. And I think the administrative folks that are oftentimes running inefficient processes, sometimes those are outsourced processes to low cost countries, sort of following that continuum of centralized, outsourced automate. What we find is people spend more time with people. And the people that are recruiters that want to be doing that really enjoy the extra time they have spending talking to people rather than scheduling interviews, other things. And we're able to abstract, delay lots of administrative work. The other thing is an increase in velocity. So you mentioned lots of the results that we tend to get is taking time to hire from 17 days to three days or 60 days to 35 days. So cutting out waste and waiting in the process is something that we've done.
[00:17:50] Speaker A: A lot of research with talent acquisition leaders and they want to be more strategic, but they are not able to because they have so much of this administrative stuff to do. So let me talk about just the high volume for a second, and we'll talk about the other roles in the high volume applications like FedEx and others. I get the feeling that the system almost does everything. Can you actually use paradox to apply, screen and get hired?
[00:18:16] Speaker B: Yes, you mentioned FedEx. We've been at various conferences in the last year with them as well. And the challenge they came with is we want to be able to give someone a conditional offer in less than ten minutes and they can. So it's a series of integrations. Right. Have you worked here before?
[00:18:32] Speaker A: Great.
[00:18:32] Speaker B: Let's check that record to see. You'd be able to write, will you submit to a background check? Can we do a pre background check? And being able to get that offer with just that initial conversation and then have the follow up background check and all the compliance pieces, etcetera.
[00:18:45] Speaker A: Very cool. And then McDonald's has the, you know, would you like a job to go with that hamburger today kind of use case?
[00:18:52] Speaker B: America's best first job, right? So, yes, being able to just accelerate that and take out the waiting, that's what.
[00:19:01] Speaker A: So let me just make a comment on that for the people listening. You all know that the job market's getting tighter and tighter and tighter. So for these high volume, high turnover roles, this kind of a technology is absolutely going to be essential in those kinds of situations because it's so hard.
[00:19:16] Speaker B: To find people, it is a competitive advantage. And we've seen that with our clients having staffing rates significantly higher, being able to open new restaurants, I think most of us in the last three or four years have gone to a restaurant that was closed or a section that was closed because there wasn't staff to staff it. So it has become a business imperative. And some of the clients that we have that are the most aggressive and the most successful are where people are close to money, outsourced services and restaurants and places where you can't make money if you don't have people. And they tend to care a lot about making money and then therefore getting people and being able to automate and act with speed because it's strategic for them.
[00:19:58] Speaker A: So in the non high volume recruiting, high value or whatever, everybody's valuable. And I know General Motors and some other companies. How well is it suited for those kinds of candidates, those kinds of positions?
[00:20:10] Speaker B: Yeah, it often depends on that experience, design of what people want to automate in there. I think some areas that we found that are almost always automatable. One of those is interview scheduling. So we scheduled about 16 million interviews last year around the world in over 190 countries. And it's one of those things that I started doing automated interview scheduling 14 years ago, and people said, well, isn't that impersonal? And I think what we've learned about interview scheduling is people want it to be right and fast.
[00:20:43] Speaker A: Just find a slot and just stick it on there.
[00:20:46] Speaker B: Yeah, exactly. And so being able to use conversational AI as the right tool to do that, sending someone a text immediately, and the medium matters a lot, even for us. We see email take two or three days, SMS or WhatsApp take less than five minutes to be able to schedule.
[00:21:03] Speaker A: This can really save time in all areas of recruiting, not just high volume.
[00:21:08] Speaker B: Exactly. And then be able to answer questions, what do I wear to the interview, where do I park, what do I need for security? Those kinds of things. And being able to reschedule those interviews, move a time for a hiring manager if they need to do that, rather than cancel the whole thing.
So that's one of the key areas, areas that we know that suits just about every company. And there's a whole bunch of other things that come out on a company by company basis that they need help with in the recruiting process that conversationally I can help with.
[00:21:35] Speaker A: Let me ask another more business question. My sense of the market is that you are so far ahead in this particular domain that you're really becoming the leader or the definer of the market. I don't know if you agree with me. I don't want to be over optimistic for you because you're the CEO now. You can sort of tell me what you think, and that means you move beyond recruiting and you think about other things you can do with this.
How do you see yourselves in terms of this market? Do you feel like you're creating it? Do you feel like you're following it? And where do you want to go with the technology?
[00:22:11] Speaker B: Yeah, it's a good question. And I think we've gotten to a place with our market share that our clients are leading the way for us there, and they're telling our story as well. They're saying we got great results with paradox. We've been able to save millions of dollars in advertising or higher, 70% faster and lots of different things. And so that's becoming the standard in many ways there.
We think this is just how hiring is going to evolve for everyone, that hiring is going to get faster. Friction is going to come out of all these processes everywhere, and we know that people are going to catch up and they're going to chase us and they're going to find that. So we've got to stay ahead and innovate. And that's true in talent acquisition and our products and talent acquisition. But we do think that there are many ways in the employee lifecycle that conversational AI matters significantly as well. We've been asked since the beginning about, Olivia does such a great job in their recruiting process. Can she help in the first 30 days, the first 90 days, some of those critical moments, especially for frontline workers, and we started some of that work to help understand that as well. How can we be of help there with onboarding in those critical moments to be sure that we're getting people in touch and doing the same, answering questions, of course, and then doing the same thing, which is being sure that we can take care of the administrative work so people can spend more time with, with people.
[00:23:33] Speaker A: Well, in a sense, almost all these aspects of employee experience. What are my benefits? How do I find so and so? What is the policy on such and such? Are all, all those internal issues are very similar to what you're doing for candidates, I think.
[00:23:47] Speaker B: Yeah, that's right. And I think the focus for us has been on the frontline worker in that people that have traditionally been left out, where they don't have an email address, they're issued on their first day, they don't have a sharepoint to go to even if they wanted to. And so what they do have is they have a phone, they have a contact they can have in their phone. And just to be able to ask a question to how do I get a new hat? When do I work with, let me.
[00:24:10] Speaker A: Throw you a curveball. You may not have thought about this, or you may have. So one of the things that's big right now is internal mobility, finding people inside the company that might be able to fit this role using a talent marketplace, using talent intelligence, whatever it may be. Where do you guys see your role in that process?
[00:24:26] Speaker B: Yeah, I think as a communication tool. Absolutely. I think one of the key parts that is missed there is awareness. So we run into that a lot in, again, distributed context where there's a new manager role open at the store, but people that are at stores 2 miles away have no idea, and they're not going to apply for it because they don't know. And so being able to solve some of the basic awareness things there as well, I think there's a lot more that we'll continue to think about from an AI perspective about what are the rules of engagement. Right. Can I send a message to your employee without you having to wanting to stab me? Those types of things that are difficult, that come at a programmatic level as you think about internal mobility.
[00:25:08] Speaker A: So the recruit. So I just didn't understand this. So a recruiting function could send out a message, in a sense, to all internal employees in a given domain about a job that was opened and then they could communicate with Olivia to apply. Is that correct?
[00:25:22] Speaker B: Yeah, absolutely. Absolutely. You know, being able to do some job search, too, being able to do some light coaching, like what kind of jobs do I have available? What do you want to do? Those kinds of things. Some of this is, you know, my early career as a recruiter, I literally used to, in manufacturing, I used to go put a big sign by the highway, said, now hiring, we put it out there, open interviews, come on in. And people would come in, they would talk to 22 year old me, and I would say, hey, I'm Adam. What do you want to do? And in many ways, that's manifested in Olivia today, where she shouldn't have to understand what our terms are. What are your skills and what do you want to do? Let us figure out whether you should be in procurement or sales or marketing in many ways. And what are the words we use to describe those positions and then be able to make that match and that suggestion, and we'll go from there.
[00:26:11] Speaker A: Well, given how important your product is and your technology, the big ATS providers, including phenom, to some degree, Oracle, SAP, workday, isims, all the others. Do you see them as partners? Do you see them as competitors? Do you see them as frenemies? What's your relationship with all that incumbent stuff that everybody has?
[00:26:33] Speaker B: Yeah, I think that all of those things, to be honest, important partners for many of them, of course. I think our lens is actually on solving problems, and it's about the client and another industry person I had a couple of conversations with. They said, I don't want to integrate with that because they're a competitor. I said, that's not a very client centric view. Why does the client actually care at all about your relationship between two companies? Help them solve a problem, and good things flow from that.
We work with all of those companies in different ways. We actually have integrations with all of those companies. And for us it's about how does the client best solve this? And sometimes, and many times it is about keeping workday and SAP and Oracle for that continuity and that platform. They get around around data and also understanding those tools can't do everything well. And so we see a lot of situations. You mentioned a couple of them earlier where paradox is really driving the experience for the frontline worker on top of workday or SAP and workday or SAP do a great job in the corporate environment and many other places, but paired up by having both, they're able to optimize for both those scenarios.
[00:27:40] Speaker A: You know, I love that I've talked to your management team so much. I think you guys really have this abundance mindset that there's room for everybody here and if you constantly work on solving problems, you'll always find a market. So one more sort of market question, then we'll wrap up. You know, we're talking to a lot of companies about various AI tools because of our own product and we keep hearing about the Microsoft copilot and you know, a lot of companies have it. I'm not sure they know what they're going to do with it yet. It's pretty good for finding documents and stuff, but it has promised to do a lot more. What do you think the impact of that will be on recruiting at this point?
[00:28:17] Speaker B: Yeah, it's interesting with every feature that we make, maybe full stop, but especially with AI, we have to go through and think about ok, in the future, where is this going to be done and who is going to do this? Because I don't want to replicate or replace anything that's going to be done in Microsoft or Google. I think that's where something like a job description generator is a good early example for us is everyone made one, us included.
But the thought really is where are people going to write job descriptions? And the answer I think is actually in the flow of work where they already do that today. And for many of them that's in either a system of record or it's in the documents like Microsoft or Google that they do that today and where they have those, we have to think about what is the best tool to do that for us, thinking a lot about the assistance that people are going to have through work generally and trying to do specialist things, things that they're not likely to do, things they don't have the data set to make them really intelligent and make them great at and thinking about how to stay for us, do things that we can be exceptional at, where we can help recruiters in the flow of work, not just because we can or it's easy. There's absolutely going to be a place for ubiquitous copilots that are from many providers like Microsoft and Google. And so we've got to figure out what's the hardest work, what's the specialist work to do, and that's the work we're focusing on.
[00:29:39] Speaker A: That's been my conclusion, too. You have a very, very deep, domain specific, highly refined conversational AI system that may be connected to the co pilot at some point or connected to another system, but I don't know that they're going to compete. And the same thing we're finding with Galileo, Galileo is very good at what it does. It doesn't do everything. So that's kind of an interesting direction. What else do we need to know about this market that you really would like people to understand when they think about recruiting and conversational stuff and AI?
[00:30:10] Speaker B: Yeah. I think the connection between problems of today and visions of the future are making the job of a practitioner today really clouded. And everyone got an unwanted second job last year, which was to be an AI expert, and everyone wants to know what they're doing with AI. And I think that the focus for us on product productivity is number one. I think I've heard you say this, Josh, which is. I love, which is, it's a search for productivity, not a search for AI. Everyone is thinking about that. So for us, it is that relentless drive for how do we use our conversational AI toolsets and our products to drive productivity and recruitment processes. And I think just that focus on design and the relentless focus on how do we get work done is the right lens.
[00:31:01] Speaker A: Bravo. Okay, thank you very much, Adam. That is great. We'll look forward to hearing more from you in the future, and thanks for your time today.
[00:31:09] Speaker B: Yeah, thanks, Josh. Appreciate you having me on.
[00:31:11] Speaker A: Okay. Obviously a very interesting conversation and lots of things to consider. Let me just give you a couple things as takeaways. Conversational AI is complicated.
It's not simply chatting with a piece of content. There's context, there's definitions, there's rules, there's curation of content, there's managing the understanding the expectations of the user who's using the system. So there will be no generic chatbot that does everything for everybody? Not in my opinion. Number two, the focus on the user and the user experience first and the transactional application second is the future of AI. It's the future of all of our lives. It's the future of our consumer life, and it's the future of our HR technology and our employee technology. And third is focus. Paradox has focused like a laser on the needs of job candidates, of recruiters, of internal job candidates looking for opportunities inside of the company, and then slowly getting into other applications like onboarding. And so they really understand the small, intricate use cases around this. The same thing goes for you. And whatever your initiative may be, a leadership AI, an AI that helps with benefits, an AI that helps with understanding best practices, and vendors like us is going to be different from an AI that helps you change your vacation days. So even though these may all be embedded into one master platform, over time, maybe it'll be the Microsoft copilot, maybe it'll be something else. These specialized systems are very, very powerful. And the fourth thing I'll just tell you is the ROI of a highly refined system like paradox is massive. If you look at companies like FedEx, others of the many paradox customers, they have reduced time to hire by ten to 20 to 30 times, from 30 days to five days reduced, reducing the workload of recruiters by 70% to 80% so they can spend more time doing strategic things. So these are really groundbreaking opportunities. And I just am just a big fan of paradox because of that. So have a great weekend, everybody, and thanks for listening.
