0:00Jordan Metzner says Ryz Labs does thousands of interviews a year and has done thousands since launching Ntrvsta. Sam Nadler calls it a labor of love that took a long time to build and is having a big impact.
2:35Jordan Metzner says interviews take 30 minutes to an hour and were the big time sink after Ryz Score automated resume scoring. He says the AI interviewer offers unlimited availability and scalability, reduced bias, dynamic conversations and many languages.
5:18Jordan Metzner says candidates find the avatar patient and calm once they adapt, and each receives an email with their score afterward. He argues AI is the only way to scale interviewing at the top of the hiring funnel.
6:41Sam Nadler says candidates can interview within minutes instead of waiting days to schedule, and humans still interview later in the funnel. Jordan Metzner says Ryz Labs has actually grown its recruiting team since adopting Ntrvsta.
9:30Jordan Metzner demos a dermatologist candidate scored 89.5 overall, with chaptered video playback, multilingual subtitles, playback speed, per-question evaluations, a full transcript and whiteboard analysis. Sam Nadler notes the avatar asks follow-ups based on previous answers.
14:40Jordan Metzner says most of the team uses Ntrvsta inside their applicant tracking system; it integrates with Greenhouse, Lever, Ashby, Workday and Salesforce. Assessments come from an uploaded or ATS job description or manual entry, with optional whiteboard mode.
16:50Jordan Metzner says Ntrvsta uses ChatGPT and other AI tools for scoring, translation and feedback, plus text to speech, speech to text, avatar generation and video streaming to create a natural, meeting-like interview experience.
18:00Jordan Metzner expects AI interviewing at pretty much every major company because it levels the playing field and removes biases. Sam Nadler says Ryz Labs increased its number of recruiters by 25% in the last thirty days.
20:27Jordan Metzner credits Microsoft's leadership for adapting to AI by leveraging its data center business and its relationship with OpenAI. He says cloud demand for Nvidia chips keeps growing and expects Microsoft to keep rising.
22:20Jordan Metzner says Figma was priced at 33, opened at 85 and traded around 102, about a $41 billion valuation, after losing its $20 billion Adobe deal. He suggests a direct listing might have captured more upside.
AI Interviewer Avatars: How Ntrvsta Screens Candidates: the key questions from this episode, answered with quotes and timestamps.
Sam Nadler [0:00]: This has been a labor of love that's taken us quite a long time to build, but is having a big impact on our organization.
Jordan Metzner [0:08]: At Ryz Labs, we do thousands of interviews a year. Since we've launched Ntrvsta, we've done thousands of interviews. It's been incredible use case, and I think we'll dive into it a little bit more.
Sam Nadler [0:33]: Hello, everyone, and welcome to another episode of built this week, the podcast where we share what we're building, how we're building it, and what it means for the world of AI and startups. I'm Sam Nadler, cofounder of Ryz Labs, and I'm joined by my friend, business partner, and cohost Jordan Metzner. What's up, Jordan?
Jordan Metzner [0:52]: Hey, Sam. How's it going? Happy to be back.
Sam Nadler [0:54]: Doing great. I'm especially excited about today's episode. Like always, we're covering a tool we have built. However, this tool is not some small internal tool we built over a small period of time. This has been a labor of love that's, taken us quite a long time to build, but is having a big impact on our organization. We're gonna dive deep into that tool. I'm gonna let you present it, and then we're gonna do some quick really hot news topics. As always, AI is crazy. There were, lots of earnings, these past couple days, so the market's on fire. So there's lots to talk about today.
Jordan Metzner [1:36]: Yeah. So much to talk about. Again, another hot summer week with incredible earnings across the board with AI just driving it all and just complete market frenzy. So, yeah, lots to talk about and super excited to jump into our product of the week as well.
Sam Nadler [1:50]: Yeah. And before we jump into our product, everyone who's listening, please don't forget to like and subscribe to Built This Week. New episodes every Friday where we're covering what we're building and what we think about what's going on in AI. So let me queue you up really quick.
Jordan Metzner [2:07]: Great.
Sam Nadler [2:07]: We have a pretty big product launch that is literally launching today as this episode comes out. The product name is Ntrvsta, which most people probably don't know what that means. But why don't you kind of give us the background, how it was built? We'll we'll I have a series of questions. Hopefully, that is is interesting to the audience, but let's kick it off and dive in.
Jordan Metzner [2:35]: Yes. So Ntrvsta is a super interesting product, that we started as an internal project here at Ryz Labs, and we've now, gone on to launch it as an external product for other external customers to use. And, like many of the products we build, it started with our own internal problems and trying to find some solutions for them. And I think by the time this episode's out, Ntrvsta will be on Product Hunt as well as a few other places online. So if you find Ntrvsta on Product Hunt, please give us an upvote. But just to give a quick background on Ntrvsta and the problem case and why we built it, at Ryz Labs, we do thousands of interviews a year. I looked at the numbers and realized that, obviously, every interview is incredibly expensive for us. And, obviously, if we interview someone and we hire them, then, that's a successful interview. And if we interview someone and we don't hire them, then, maybe we've wasted some time there. And, in general, interviews take anywhere from thirty minutes to an hour.
Jordan Metzner [3:29]: And, you constantly spend time in interviews where you know, like, possibly that the candidate may not be great, and yet, you have that time dedicated as a out of respect for the candidate and whatnot. As AI has been growing across our business in general, I think it was, obvious to us that AI and interviewing were gonna be a critical piece. And, once we built, the Ryz Score app, which I think we talked about on a previous episode, scoring resumes automatically, kind of that was a huge time killer for our recruiting team. I mean, it saved them a ton of time because, we're able to run through the scores and use that as leverage to move much faster. But we realized, obviously, the big time suck is is interviewing. And so it started off with just, a really simple idea of, how can we use AI and AI avatars to help us interview. And since we've launched Ntrvsta, we've done thousands of interviews. It's been incredible use case, and I think we'll dive into it a little bit more. But, it's it has some really great things that humans don't have, which is, unlimited availability, unlimited scalability, high reduction in bias. All the scoring is is pretty evenly fair.
Jordan Metzner [4:40]: Large amount of expertise, so, the AI interviewer can interview across a multitude of subjects. And, in this case, it's a female Carolina. She's an incredible expert in pretty much every subject she interviews in. The interviews are dynamic, which means that, there is not a fixed set of questions, but more of a conversation, a dialogue like a realistic interview could be had in many languages. It's easily shareable and repeatable. And I think it just shows that, AI is gonna have an impact across so many different industries, but it certainly had a big impact on our business so far. And we think Ntrvsta is an incredible tool for other large organizations that are hiring.
Sam Nadler [5:18]: I think and we'll get into this, but I think it's easy to make the argument for large hiring organizations why Ntrvsta could be such a great tool. But for even the candidates themselves, why and you touched upon this a little bit in your introduction, but why is potentially this even better than the previous, historical human solution?
Jordan Metzner [5:42]: Yeah. We've got a lot of feedback from candidates that they really enjoyed the experience, mostly because they realize that it's they're they know it's an avatar. And so, while it's a little, I would say, maybe unique or different or even awkward, At first, they kind of adapt to it pretty quickly and realize that, she is very patient. She's calm, and she's there to help you. I think the transparency of the scoring algorithms really help candidates. So, after a candidate finishes their interview, they receive an email with their score. And that just drives towards kind of more buy in and ownership of the candidates themselves. And at the end of the day, what you're really looking for is try to hunt find the best people for your organization, and the way to do that is really to scale up, the top of your funnel. And that means, more candidates through the door. And that as you open up the top of the funnel, that means you need to interview more candidates.
Jordan Metzner [6:31]: And so, AI is kind of the only way that you can truly kinda scale your organization at the top of the funnel through interview process.
Sam Nadler [6:40]: Okay. And for you know, obviously, like I mentioned, this is a bit easier to define. But for large hiring organizations, why could this be so nice? For me, having being part of a, an owner and a founder who runs a staffing organization, speed is critical. And the number of candidates we're moving through our funnel is also critical in order to find the best candidates. And since we've adopted Ntrvsta, the ability to move candidates through the top part of our funnel, which is where we use Ntrvsta, is so much faster. I mean, immediately from the Ryz score directly to being able to take an interview literally within minutes as opposed to waiting three days and scheduling it, the availability of the recruiter versus the availability of the candidate, can take days, if not weeks, versus immediate access to being able to interview. Secondly, you have just an immediate, feedback on if a candidate meets the needs of the opportunity. You get scoring immediately based on several different criteria, and you can potentially prioritize those candidates in even faster than your typical kind of normal process. Anything else from the organizational point of view that just makes this such a unique tool?
Jordan Metzner [8:03]: Well, one thing I would I and maybe your feedback's important here, but, Ntrvsta obviously is an alternative to having recruiters do interviews. How do you think the recruiters have been able to take advantage of it, and has it replaced recruiters, or has it kind of, given them superpowers? What have you seen from your side?
Sam Nadler [8:21]: So there's still a human element to our process. We still do human based interviews within our funnel, so it hasn't completely replaced the recruiters. It's just significantly accelerated the top of the funnel for us. Recruiters will be able to I think they'll still exist. You still need there's still elements where they'll need to manage, like, open role requests and, get people engaged in opportunities. I wish every opportunity was, pure inbound, but a lot of it's it's outbound. But, hopefully, we can, scale up our capacity significantly without necessarily having to scale our recruiting team. So I don't think at least for us right now, there's there's no focus on reducing the number of recruiters we have. It's more on ex expanding their productivity greatly.
Jordan Metzner [9:14]: Yeah. I think actually since we've
Sam Nadler [9:16]: In fact, we're hiring more recruiters.
Jordan Metzner [9:17]: Yeah. I was gonna say that since we've implemented Ntrvsta, we've actually increased our size of our recruiting team simply because they can be more efficient. We can take on more roles and more opportunities at the same time. Okay. Let's jump into the product a little bit here.
Sam Nadler [9:30]: Yeah. I would love to see a quick demo of what the actual tool looks like.
Jordan Metzner [9:33]: We can jump into a specific candidate here, and we'll have to block some of these details because they have some personal information. But, Ntrvsta works across all different types of interview types. In this case, this is actually a dermatologist who's interviewing for an RLHF role in regards to dermatology. And, the reason I found this role, so interesting, and I selected this interview, because the candidate just absolutely fantastic. She gets a 89.5 as an overall score, but a ninety in a technical score and 85 in soft skills. And so when a candidate takes an Ntrvsta actual interview, the first thing we do is we ask them their background, where they're from, what their interests are, what their hobbies are. In this case, Sarah likes to play tennis. We take us we take a headshot of the candidate, and then we also talk about highlights and lowlights of the interview. So, we see some of the areas here where the interviewers did great, and then maybe some other areas where they can improve. And then when I was I jump into it, what I think is really exciting is that you can watch back the entire interview here, simply by either watching it in a full playback mode, but you can actually just jump to any chapter in the interview and see the candidate's feedback.
Jordan Metzner [10:50]: I chose this chapter, but
Dermatologist [10:52]: Terms of the possible diagnosis. So it wasn't always very straightforward what the diagnosis would be. As we talked about before, it can be difficult to tell between an acne papule and a squamous cell carcinoma unless you have the appropriate history or you're looking at surrounding photo damage, on the patient. How do you handle cases where a lesion doesn't fit neatly into standard categories? This can be difficult. When this occurs, I use my knowledge to make the best annotations possible based on the information that is given. Typically, you are able to tell color. You can tell the primary lesion. And if other options are indeterminate, that indeterminate is selected so that the AI is not being trained to make guesses that are not accurate.
Jordan Metzner [11:39]: Great. So I just want to show out some features here besides chapters. Obviously, we've got subtitles supporting multiple languages. We've got playback speed, so you can actually watch the interview at a faster speed, which is also pretty awesome. Here we have every question with an answer and some evaluation. So if you jump around, we have the availability for you to copy and review the entire transcript of the interview. And for some interviews, we also have whiteboard analysis. So the candidate will be asked to sketch something. In this case, a simple workflow diagram that standardized the annotation process. And here you can see the image that the candidate drew during the interview as well as some analysis there. So really some awesome features when you're going to review a specific candidate.
Sam Nadler [12:21]: And as you can see from the video, yes, I don't think it's hiding the fact that it's an avatar. But, like, pretty quickly, it becomes pretty normal. The latency is is very reasonable. Yeah. It's maybe not as fast as a conversation with two humans, obviously, but it's it's, not awkward. And I also noticed at the end, like, it's dynamic. She asks you know, she takes the avatar takes your answers and asks a follow-up based on your previous answer.
Jordan Metzner [12:56]: Yeah. It's incredibly dynamic, and that's, think, I what makes it feel so natural for the candidates as you saw here with Sarah. And, here is a selection of dermatologists, and you can see their different scores and how well they performed. And in this case, Sarah's the top performing. But I would say that, 89, 85, these are still really, really good scores that we see overall. I could also jump into some other features. Theoretically, you could have, whatever, a dermatologist job opening post this role, move people through, and, you
Sam Nadler [13:27]: Know, depending on how fast people take the interview. But within a day or two days, you could have five candidates, their scores, their full interview, and wasted no time. Before we jump into the other features, I see that we have these dermatology candidates, and they were asked specific questions regarding dermatology. We have engineer candidates. We have legal candidates. How do you know that the avatar is gonna ask, dermatology related questions and legal related questions, like, versus something completely off base?
Jordan Metzner [14:00]: Yeah. So we've done a lot of prompt engineering to really get the interview experience to be really tight to keep Carolina, our interviewer, on track and make sure that she's able to both be friendly and communicative while not being inappropriate and not falling offline. We've done a lot of work to try to break that, see if we can say have her, say something inappropriate or something she shouldn't shouldn't be saying. And we've actually found that we've had really great results so far. And, most candidates are trying to get a job, so, they also act professional. And we find that, the feedback loop is is pretty accurate on both sides, both from the recruiters as well as from the candidates.
Sam Nadler [14:40]: Cool. I wanted you actually to, if you don't mind, show me how an assessment is created via the exam engine.
Jordan Metzner [14:46]: Yeah. Well, one thing I think before we jump into that, obviously, Ntrvsta has a really nice interface here, but most of our team actually uses it with a direct integration into their applicant tracking system. So, at Ryz Labs, we use Lever as our applicant tracking system, but Ntrvsta actually integrates with a bunch of different ATS and CRM tools. So, whether you're using Greenhouse or Lever or Ashby or, Workday or even, Salesforce or something like that, you can directly integrate Ntrvsta into those tools so that, Ntrvsta becomes kind of a secondary view, and you can continue to maintain your applicant flow and move candidates through the funnel. And that's what we found has worked best for us internally as well. So, you can simply upload a job description right here. Like I mentioned, you can also leverage job descriptions that are built inside your a inside of your ATS or CRM, so we can pull those automatically as well. And then you can also do manual entry by, entering the name of an assessment and filling out some information like job description, topics to cover any specific questions or any type of certain instructions. And then we also allow the ability to turn on the whiteboard mode. So, we can decide whether or not an interview deserves a whiteboard or not, and we can turn that mode on.
Jordan Metzner [16:03]: And then once you do that, that'll create an assessment, and then you can immediately start sending that assessment out to different candidates just like I showed you previously with the dermatologist.
Sam Nadler [16:11]: Yeah. And not only will it create an assessment immediately, you can then once it creates the assessment, if something doesn't feel right, you can go in and edit it to finalize it, or you can just ship it as is. So it's a great little feature there.
Jordan Metzner [16:23]: Yeah. That's right. You can here look at the assessment, and you can make edits to the assessment. And, not to spend too much time on the product itself, but obviously, it's got user management. It's got, dashboards and analytics and a lot of other features that probably be too boring to talk about on a podcast, but, it is a fully feature rich product and allows organizations to understand what's happening within the organization when these interviews are taking place.
Sam Nadler [16:49]: And, while this wasn't vibe coded in a few days by one or two people, Can you give me a little bit of insight into how this was built?
Jordan Metzner [17:03]: Yeah. So, one of the things about Ntrvsta is that as new technology comes out, we continue to make improvements, and we've seen a lot of our AI toolset improve significantly over the software development time that we've been building the app. So, we use, AI, including ChatGPT and other AI tools for scoring, for translation, for feedback. We use, text to speech tools, speech to text, avatar generation tools, video streaming, video editing. There's still a lot of technology built into the Ntrvsta app that allows it to have this smooth end to end, user like meeting experience that that gives candidates, as natural as we can with today's technology interview like experience that that hopefully feels like they're, taking a real interview.
Sam Nadler [17:55]: Great. Anything else to cover on Ntrvsta before we move on?
Jordan Metzner [18:00]: Yeah. I think, right now, we're talking about AI interviews as if it's an oddity or it's like a new technology. But, it seems to me, especially within our use internally, that, it's only a matter of time before this becomes the norm, like, pretty much within any organization. I think we're gonna see a lot more AI interview technology implemented across pretty much every major company. And I think mostly because of the fact that it actually really levels the playing field. You know? When I worked at Amazon, you had to get trained and how to give interviews. After you give an interview, you have to score that interview. There's, four or five people coming into a room having a debrief. And I think if, the more information you have about your own organization, you're gonna get better and better at, leveraging a tool like Ntrvsta inside of it. And I think that what you're gonna find is that it's gonna remove a lot of biases to get some of the best candidates to the top, and not just, who feels one way or another. And so, basically, the cream will rise, and you'll be able to really get, more people through the door and better candidates through the door as well.
Sam Nadler [19:07]: Amazing. Yeah. I love it. It's been a huge game changer for us. Our team loves it. I think they were a little apprehensive at first, and now they love it. Now, like, our ability to take on more clients, more roles with, roughly the same previous capacity has allowed us to now hire more people, which I think is, like, the immediate fear with AI is is this gonna take my job? Where for us, I mean, obviously, it's been a very short amount of time. In the last thirty days, we've we've increased our recruiting capacity, our recruiters by 25%. So, it's, it's been the opposite in a short period of time on kind of what that fear is typically. Cool. I think this is a really exciting product. It's been really a exciting journey that, in many ways, is just getting started.
Jordan Metzner [20:07]: Yeah. Totally. So I think, again, as we see new AI technology being shipped out from the major players, I think, hopefully waiting for, like, GPT-5 to come out soon and some other new LLMs, I think we'll continue to see improvements inside of, our tools and AI interviewing overall.
Sam Nadler [20:27]: Okay. Great. I do wanna cover a little bit of the news today or this week. Microsoft entered into the I believe they're calling it the 4,000,000,000,000 club with NVIDIA. What are your thoughts on Microsoft, and where does it go from here?
Jordan Metzner [20:50]: Yeah. I think, Microsoft's such an interesting company because, traditionally have been considered so, quote, like, old school or kind of in the first wave of, technology enterprise companies. And I think it just shows kind of, that a leadership team has done an amazing job adapting AI and, leveraging their data center business and their relationship with OpenAI to build just a massive business, and it doesn't look like it's gonna slow down. I mean, there seems to be almost unlimited demand from, the major cloud providers, AWS, Microsoft, Google, and Oracle. This demand doesn't seem to be slowing down. It seems like the demand for chips from NVIDIA is continuing to gain significantly, and Microsoft's been able to capitalize on all this. And, yeah, I mean, you see the shares jump significantly, as soon as the earnings were announced, and it looks like they're not gonna slow down. So, yeah, I think, some of the old guards, the old technology companies, the Microsofts and Oracles of world, look at them. They're at are at the top of it, you know, top of the list for being innovative in AI. And, they're leveraging their ability to use their capital to be first market players, and it's paying off.
Jordan Metzner [22:05]: And I think it's not gonna slow down. I think we're gonna continue to see Microsoft rip over the next few years as, they get a lot of investment in AI. Okay. Let's let's move on from Microsoft. You wanna go into Meta or Figma? What do you think?
Sam Nadler [22:20]: Let's do Figma. It's IPO ing today. We talked about it briefly, I think, two weeks ago. It's a really interesting IPO. It is open now, it looks. I think it was initially going to list at, what, $28 or $25, but already at a 102. That's crazy on its first day, and it's a great product. What do you think?
Jordan Metzner [22:55]: I think it was supposed to open in the it looks like previous closed open rate was 33. It opened at 85. It's trading at $1.00 2. I mean, these are just, like, incredible numbers for IPO and technology companies overall. I think this might beg another reason for why not do direct listing. I read that it was a lot of secondary, but still, I think direct listing probably would have captured some of the upside here. But, yeah, Figma is a great product. We use it internally. I think, since, implementing AI, we probably use it less than we used to. But, yeah, it's been an incredible story of, the Figma having the deal with Adobe at 20,000,000,000 and then losing the deal with Adobe. And now here they are on opening IPO day, trading at 41,000,000,000. It looks like it's continuing to trade up and down. It was halted for a little bit. I don't even know if you know that, but it was halted this morning once it went out to post trading. So, yeah, the market is feverish for these types of hungry growth technology companies.
Jordan Metzner [23:51]: And I think, probably even at this price, even at a 40,000,000,000 cap, it's probably still gonna be upside here for companies like Figma. And I think it'll probably be a pathway for other technology companies like the stripes of the world to go public if we're seeing such hunger in the market. You know?
Sam Nadler [24:07]: Yeah. I was gonna ask. So, 2025 has been a pretty good year for IPOs, if not a great year. Who do you think's coming next?
Jordan Metzner [24:15]: Yeah. I don't know who's coming next. I think, we're gonna see the hottest technology companies go public. And if now is not a time for a company like Stripe, I don't know who is, but I don't know when is, actually. But yeah. I think and I hope more technology companies do go public. I think it's a good thing for the market and good thing for the tech industry overall.
Sam Nadler [24:37]: Alright, Jordan. So as always, a crazy week in AI news with Microsoft, Figma, Meta, so many more happening. Thanks for walking me through Ntrvsta. It's been a fantastic episode. Hope everyone likes and subscribes, and we'll see you next week. Any final thoughts, Jordan?
Jordan Metzner [25:01]: Yeah. Great episode six. Love talking about AI this week and Ntrvsta, Meta, Microsoft, Figma. So much going on. Seems like it's not gonna slow down. Maybe GPT-5 coming up next week, which obviously will just turn everything upside down. If we see GPT-5 beat Claude, it's gonna be another crazy week. And, yeah, I don't see any way of this slowing down. So AI is gonna continue to dominate the tech news and tech space, and, I think we're up for another crazy week next week. So, anyway, this was great to see you, Sam, and looking forward to next week's episode.
[25:35] Theme song: Built this week, breaking it down. Built this week, we show you how. A fresh idea, a clever tweak. You locked in. You built this week.
