YC Root AccessJuicebox: AI Agents for the Hiring Process
EVERY SPOKEN WORD
30 min read · 5,890 words- 0:00 – 1:15
Intro
- HTHarj Taggar
[upbeat music] Today I'm joined by David and Ishan, the co-founders of Juicebox, who recently just announced their Series A of $30 million led by Sequoia Capital. Thanks for joining us, guys.
- DPDavid Paffenholz
Thanks for having us.
- HTHarj Taggar
Why don't we start by explaining what Juicebox does?
- IGIshan Gupta
So Juicebox is an AI recruiting platform. We help companies find, engage, and evaluate talents faster and more efficiently using LLMs.
- HTHarj Taggar
Okay, cool. So what did recruiters do, uh, before Juicebox existed, and kind of what are the biggest pain points you're solving for them?
- DPDavid Paffenholz
A typical recruiter workflow, um, has a few steps that starts with searching for a candidate. Uh, typically, that is an online profile search. We set filters, like the job title, and we start reviewing those profiles. Uh, then they'll evaluate those profiles by clicking into them, it's probably the most time-intensive portion, and finally reach out either via email or via LinkedIn. Juicebox covers that entire workflow, and so we help recruiters do that entire workflow either in an AI-assisted or in a fully AI-led way.
- HTHarj Taggar
And can you share a little bit about just who are some of your customers today, and, um, what's the, the scale you're operating at?
- IGIshan Gupta
Definitely. We work with over 2,000 customers, and we work with talent teams across some of the fastest growing companies in tech, including companies like Perplexity, Ramp, and
- 1:15 – 3:00
What Juicebox Does
- IGIshan Gupta
Cursor. We also work with a lot of large recruiting agencies, as well as some of the largest enterprises in the world.
- HTHarj Taggar
I f- wanna start by just taking a bit of a step back about your journey to get here 'cause it, it wasn't, like, a straight line to success and product market fit because you guys actually did YC a while ago in 2022, and you've been through, like, a bunch of pivots to get here. So I think that's gonna be really interesting for people to hear about. So why don't we actually start with the two of you individually. Tell us a little bit about kind of your personal backgrounds, like where did you grow up, and how did you two meet?
- DPDavid Paffenholz
So I'm from Germany, from Dusseldorf, grew up there, um, came to the US for college, went to Harvard, studied economics, um, and then pretty quickly got interested in tech. Um, COVID happened at the same time, and so, uh, that's the time that Ishan and I ended up meeting each other, and I spent some time working at Snap, uh, on their growth team. Really cool experience. Ended up going back to college and starting to build together with Ishan, uh, and so that was really our journey of getting started. We first built a music app, um, a social audio app, uh, and that ultimately gave us the conviction to apply to YC as well.
- IGIshan Gupta
I grew up in India, this town called Kanpur, uh, went to high school there, and I was always interested in tech, always enjoyed building stuff. The story of how we met is actually really interesting because in, in my senior year of high school, I came across this competition online... This is during COVID, by the way, everything is virtual, where the prize is that you get to meet Stephen Wolfram, who's a popular computer scientist, and I wanted to meet him, so I ended up taking part in that competition. And David was actually one of the organizers of the competition. So I got to meet Stephen Wolfram, but the more interesting bit was I got to meet David. Uh, he was working on a music app, uh, actually a social media app at the time. I really enjoyed building stuff, so I joined in. We started working on that, then we, as David said, we worked on the music app, and at some point we were like, "You know,
- 3:00 – 5:00
From Music App to Recruiting Startup
- IGIshan Gupta
we should start a business together," and that's when we applied to YC.
- HTHarj Taggar
Well, look, I, I have to ask, you have a very distinctive name for the company. Where, where did the name Juicebox come from?
- DPDavid Paffenholz
Yeah. Juicebox goes back to our music app days. And so when we were building the music discovery app, it was a wordplay on jukebox. Um, and then when it came to... came around to launching the recruiting platform, it was, like, the night before the launch, we're trying to decide what name we would give it, and, um, we really loved the name Juicebox, even though it wasn't connected to the business, so we decided to go with something fun, and, uh, it stuck since then.
- HTHarj Taggar
Do you think it's, like, helped you stand out at all, or is it neutral?
- DPDavid Paffenholz
I think a lot of companies in the recruiting tech or maybe even, like, B2B SaaS in general perhaps have some, like, repetitive naming, uh, and so it's definitely different, um, and hopefully a bit more memorable.
- HTHarj Taggar
Can you give us, like, the quick, um, summary of what was the idea you applied to YC with?
- DPDavid Paffenholz
The idea that we wrote our application with, and it even evolved a little bit until [laughs] we actually did the interview, um, was a music discovery app. So it connected to your Spotify, uh, and then it gave you 15-second snippets of songs that we thought you might like. Uh, it was still, I guess, in some ways related to discovery and finding new things, but in a much more fun, consumer-facing way. And then by the time we interviewed, we were already thinking about, you know, how could this be... become a business? What, what could be monetizable here? And so we pitched an idea around it being a merch marketplace. Turned out that wasn't a great idea either for, for some other reasons, um, and also gave us the openness to look at other ideas afterwards.
- HTHarj Taggar
Yeah. I, I remember that from the interview. I remember, like, you're thinking, "Okay, these guys seem great, but they're... Like, they're still figuring it out," which is totally fine. Then when YC started, y- can you tell... Talk to us a bit about, like, um, how did the idea change again? And just especially, like, how were you coming up with ideas and, and figuring out things to work on and testing them?
- IGIshan Gupta
I think one of the best things YC did for us is, especially for young founders, it kind of tells you that there's this whole world of business software. If you're a young founder, the natural course is that, you know, I can solve a consumer problem, so I can build a consumer application. But then when we got into YC and we started seeing the businesses YC has funded, we realized, hey, there's a whole
- 5:00 – 7:10
Early Pivots and Finding Product-Market Fit
- IGIshan Gupta
world of problems we can be solving for businesses. So that's when we started thinking about, you know, well, what are ways we can, like, build good business software and help, help companies do better? And we realized that one, one of the things both David and I had in common is we, we had unconventional work experiences in the past. So I freelanced a ton in high school. David had also worked with a lot of companies on independent projects. So we thought... The initial idea was what if we build a better marketplace to connect people? That's where we started off, and that got us in touch with talent teams and got us into the recruiting space, and over time, we realized that while the marketplace thing has already been solved, what's a bigger pain point is the actual tools that these teams are using right now.
- HTHarj Taggar
The recruiting teams?
- IGIshan Gupta
Yeah.
- HTHarj Taggar
Right. Okay. So then, um, when did you kind of first get the idea for, like, Juicebox itself as, like, a, a tool for recruiting? How did that come about, um, uh, and how'd you get that off the ground?
- IGIshan Gupta
This was when LLM started taking off really, and we, we, we started playing around with some of the APIs. We started playing around with ChatGPT, and we realized that LLMs are able to draw inference from unstructured data, and that's a huge part of what recruiters are doing on a day-to-day basis because they're pretty much going through profiles and trying to draw inference out of it.And it really gets semantic meaning out of, out of the profiles they're looking at, and we realized that LLMs can do that very well, so this is a space we can be building in.
- HTHarj Taggar
And so this is j- so you were working on sort of like the freelancer marketplace, then ChatGPT launches sort of, you know, towards the end of 2022, and is that ... Was it ChatGPT launching that gave you kind of like the, "Oh," like, "what if we, like, took this and applied it to recruiting?"
- DPDavid Paffenholz
Yeah. I think it, it kind of fell ... Like, it was like the puzzle pieces started falling into place because we, we had identified the problem. Um, we were already building on the contractor marketplace side, which effectively meant we were being recruiters, uh, plus we were interacting with recruiters from other companies. And as Ishan mentioned, we saw the pain points, but then we didn't really have conviction on how we could solve those yet. And so LLMs really changed the game for us because it meant we could start solving for some of the workflow that was being done by recruiters rather than just the software being used by recruiters.
- HTHarj Taggar
What did the MVP look like? So when you're talking about workflow as well, yeah, what, what could a recruiter do with version one of Juicebox?
- DPDavid Paffenholz
Yeah. Very
- 7:10 – 9:20
Why Recruiting Needs AI Agents
- DPDavid Paffenholz
first version, I think we actually did a, a, an internal YC launch on this. It must have been, like, December 2022, so a couple months after the ChatGPT, ChatGPT launch. And there we took some of the most common workflows or processes that we got to know from recruiters and that we experienced ourselves and turned those into a small LLM-powered application. And so that was, like, structuring interview questions, quickly summarizing or assessing a profile, and there was, like, a couple other use cases in it. It was pretty basic, but it showed the initial promise of what a tool could do in this space.
- HTHarj Taggar
When you launched this idea, um, how did it feel different to just, like, the other ideas that you'd been sort of pivoting around and testing out on initially?
- IGIshan Gupta
I think we had the sort of, like, instant message market fit, which we had not gotten yet. So I wouldn't call it product market fit because the product was, like, barely functional, but we had this message market fit that, like, people saw the, the message of what we were trying to build, and we instantly got so many emails. We got a lot of messages, and we started getting in touch with people where they were like, "Oh, recruiting is a big pain point for me, and this would really help me." I think compared to our other ideas, like, people really related with what we were trying to build.
- HTHarj Taggar
The key difference it just sounds like this time you launched something, and it just feels like you've got, like, inbound demand and people saying they really want it, whereas the other ideas maybe just it didn't feel like you were not getting that pull. How did you get those first few customers, and just what did those first few months look like?
- DPDavid Paffenholz
We were actually just discussing this today as well of, like, the first, the first few customers were not super happy with the product.
- HTHarj Taggar
Oh, interesting. Tell us about that.
- DPDavid Paffenholz
Yeah. It was, it was still a journey from there. So we had that message market fit that Ishan mentioned, but we didn't yet have product market fit. So the product wasn't actually working. It wasn't a recurring use case. It was maybe interesting to see in the demo and to see the initial outputs from it, but it wasn't yet good enough to be something that becomes a regular part of the workflow.
- HTHarj Taggar
Is that because they ... the underlying models weren't good enough for it yet, or what, what wasn't working?
- DPDavid Paffenholz
I think a bit of both. So both the underlying models, this was like the GPT 3.5 era, were not yet good enough. Shortly thereafter, GPT-4 came out, and it started getting a bit better for that use case. But then second, we had not built enough of a product suite around it. It was still very much like, uh, I guess it was an LLM wrapper in the purest sense of the word.
- HTHarj Taggar
Yeah.
- DPDavid Paffenholz
Um, and that was kind of our V1, and it was good feedback though because
- 9:20 – 12:00
Building the Platform: Search, Engage, Evaluate
- DPDavid Paffenholz
we knew exactly what our customers wanted, and they told us what they wanted. It allowed us to go back and, and start building for that.
- HTHarj Taggar
A- and what, at this point, just what specifically were your customers trying to do with this V1, and what weren't they happy with?
- DPDavid Paffenholz
They were trying to do many things, but the biggest one and the most recurring one was finding the right talent.
- HTHarj Taggar
Yeah.
- DPDavid Paffenholz
And so a lot of the other workflow items were important too, but everything kind of came back to that search piece. Can we discover someone that we're not discovering elsewhere?
- HTHarj Taggar
What's, like, a typical search query that they were putting in?
- IGIshan Gupta
Something like, uh, looking for software engineers in the Bay Area who have experience working with large language models, uh, or who have experience working with, uh, deep learning or machine learning algorithms, things that require you to really look at a profile and kind of gauge whether or not that's true, more like semantic queries.
- HTHarj Taggar
A- and what was happening as you were trying to ... with version one trying to process those queries?
- DPDavid Paffenholz
Yeah. We started showing ... So we sh- we showed matches. I think there was some people who found value in that use case, but the depth of search that we were doing was not quite yet where we wanted it to be. And so we were still quite filter-focused. Um, we were using LLMs to parse that into filters, and then we started working on some vector search behind the scenes to help prioritize those search results. And so it worked in some cases but not consistently.
- HTHarj Taggar
Okay. So then when, when did it feel like it started taking off and felt like you had some product market fit, and just to ... just what does that feel like?
- IGIshan Gupta
For roughly, like, four months, the revenue was kind of like this, like, flatlined basically. Then we continued iterating on the product behind the scenes. We continued being close to customers, really listening to what they really want, and we, we would obsess over every single search. I remember we, we set up the Slack integration. Any, any minute you, like, type a search, it's gonna show up in our Slack. And then anytime we see, uh, a search come in, we would, like, instantly get into it and see what went wrong, what can we do better. Is the ranking not good enough? Are the filters not good enough? So we really obsessed over making our search good for the next four to five months, and that's when it really hit a point where people started using it for real use cases. Uh, the inflection point was when people, first of all, started inviting their team members and getting their team on board, and then they started integrating their other tools in the ... in, in their, like, recruiting stack, and that really gave us the confidence that we are building something that can really be a part of your workflow.
- HTHarj Taggar
And so then coming back to the, um, the round you just raised, y- you actually took a while to raise your Series A 'cause the company was actually just, like, growing, making revenue, and doing well. So maybe just talk us through as founders your mindset around that. Like, I think you could've raised a round much sooner if you'd wanted, but, like, how did, how did you think about doing that sooner versus waiting?
- DPDavid Paffenholz
Kind of the point of ... that Ishan mentioned of where we got to, that continued growth, was around January 2024, and from there on, like, the revenue curve was no longer flat. It started
- 12:00 – 14:30
Breaking into the Market
- DPDavid Paffenholz
creeping upwards, and, um, every month we were growing 20, sometimes 30%, uh, and then that started compounding. And so by the time it was fall of last year, we crossed a million in ARR. Um, it was still the two of us, or-Our founding engineer, Minchu, had just joined and at that point we raised a, a seed round, and that gave us the capital kind of continued doubling down. It was good for the company that it took us a while to, to raise that financing. Or not that it, like, took us a while, we only really-
- HTHarj Taggar
You chose not to.
- DPDavid Paffenholz
Exactly. We, we rather chose not to. And the reason for that was that the biggest issue was iterating on product. And to be honest, you can do that very well with a small amount of people. And so it was the both of us going into every search, testing it and changing it. There was no coordination with other people, it was just us making it better every single day. And I think at that time, that was what had the biggest outcome on the product, uh, rather than building a whole team around it.
- HTHarj Taggar
I remember reading your error update when you crossed a million dollars ARR, um, and it surprised me because you got there without hiring anyone. First time I'd seen that, actually. So do you think that's kind of the future for startups? Like, can they get to million-dollar ARR with less people?
- IGIshan Gupta
So I, I think that's definitely possible, especially on the product side of things. On, on the GTM roles, I think we, we still need a lot of people to have those one-one relationships when making sales. But on the product side, a few smart people really working together and really working hard and using AI tools well enough really gets you pretty far, because you need more people who can make the right decisions, but the tools make it a lot easier to grow a lot more quickly.
- DPDavid Paffenholz
Even on the go-to-market side, I think in the early days, like, if you booked a demo with Juicebox, you booked a demo with me, and there was a fairly long time where I was doing, like, 60, 70 demos a week. And so it was a pretty high volume of demos. Um, and some of them became repetitive, but it was also nice because it almost became a second nature of this is what the customer's gonna ask for next, and here's how we're gonna solve that in the product. And in a way, it was like the truest form of product development, is like we got all this customer feedback, and then it was just the two of us going and implementing it.
- HTHarj Taggar
How did you know it was time to, um, hire your first person?
- DPDavid Paffenholz
We started onboarding some larger customers. Um, there was more customer-specific demands that we wanted to serve, and so particularly we started with working with some of the large AI labs, um, which of course have a very intense hiring use case. And, um, I think that was around the time where we realized, like, maybe it needs to be more than the two of us.
- IGIshan Gupta
It would be so funny when we have conversations and people are like, "Oh, how big is your team?" And we're like, "Uh, what team?"
- HTHarj Taggar
[laughs]
- IGIshan Gupta
Like, you know, we try to, like, avoid the question a little bit.
- HTHarj Taggar
Change of topic here a little bit. Um, let's just talk about the market that you're building in. Um, it's a market
- 14:30 – 17:00
Landing 2,000 Customers
- HTHarj Taggar
I know well. I ran a recruiting startup for four years, and historically recruiting has not been a great space to build startups in. I think investors in particular became a bit jaded after funding lots of them and, and not seeing many really grow in scale. Did you think about that when you were starting this company at all? Like, were you concerned about starting in the recruiting space? Were you excited to start in the recruiting space or, or did it not matter?
- IGIshan Gupta
Not really, I would say. We, we didn't think about, like, investor sentiment too much. We really cared about, like, is this a real problem? Do we have a real solution that we can build for it? And we saw one, so we just went for it.
- DPDavid Paffenholz
And I also think it was, like, a... In a way, the fact that we didn't know all of the context behind us helped us.
- IGIshan Gupta
Yeah. [laughs]
- DPDavid Paffenholz
'Cause I, I don't know if we would've made that decision otherwise. And then it was also fortunate timing with LLMs really changing what's possible. Like, you can actually do the work that a human otherwise does, um, through software now, which is just kind of really unique in this point in time.
- HTHarj Taggar
Yeah, so l- let's talk about that more, 'cause I, I, I couldn't agree more. I think, uh, recruiting is a space where AI is gonna completely explode the size of the market. It's a very exciting time to build. Maybe tell us what are some of the, the specific features you built at Juicebox, um, that use AI which, which just wouldn't have been possible to build two or three years ago.
- IGIshan Gupta
There, there's probably two things that come to mind. One is our autopilot, which is a, a feature within Juicebox that pretty much looks at your entire talent pool one by one, goes through every single profile, and does really a very, like, deep analysis of whether or not this person is a good fit for the role you're hiring for. That would've very much not been possible without something like LLMs. And the other piece I would highlight is our agents that we launched a few months back. Uh, agents are basically autonomous workers within Juicebox that are able to do a lot of what the top-of-funnel sourcing looks like. So you kind of have this cal- calibration process where you iterate with an agent. The agent shows you profiles, you give feedback on those profiles, and then once the agent is calibrated enough, it can go out there and start reaching out to people for you. And those are really two of the main features within Juicebox.
- HTHarj Taggar
Yeah, I think it's, it's, it's a classic example of where because the AI can do the work, the amount of money you can make per customer is gonna go way, way up.
- IGIshan Gupta
Yeah.
- HTHarj Taggar
It's not just a tool to enable recruiters, it can do the actual recruiting. Knock-on effect of that is the role of recruiters is definitely gonna be impacted by what you're working on. Um, so maybe tell us a bit about how do you see the role of recruiters evolving or co-evolving alongside Juicebox, and then what's your advice to founders who are building companies and thinking about hiring recruiters
- 17:00 – 19:20
Working with Perplexity, Ramp & Cursor
- HTHarj Taggar
and the talent function, and, and how should they adapt to this new world?
- DPDavid Paffenholz
Yeah. I think there's some parallels between what's happening in the recruiting space and in the sales space, where in the sales space there's been a large tech stack for kind of automating different things and, uh, making different flows or intent signals more advanced. Um, in recruiting, that's still a lot newer, and we see that, uh, for a lot of our customers, we're the first solution that they find in that, in that space. What we've then seen some of our most sophisticated customers do is actually have people on their teams who are focused on deploying agents into their workflow. And so they might manage multiple agents at once that all focus on different roles. And so suddenly you have the capabilities of a single recruiter are, are multiplied, and they can work on a lot more roles, they can find a lot more qualified candidates, and they can have a lot more output than they would otherwise. It also enables small recruiting teams, so be that a founder or a hiring manager who wanna do some of their own recruiting, to do that too. And so I think that's something that we'll hopefully see gr- grow a lot, is people who are not traditional recruiters acting as a recruiter too.
- HTHarj Taggar
Yeah. But that was... Definitely when I was running, um, uh, Truebyte, the general issue in the industry was the more recruiters you hired, the less effective they became because the overall conversion rate on LinkedIn messages, response rate, reply rates were going down. And so it seems like this could potentially be a solution to that.
- DPDavid Paffenholz
There's kind of two parts to that. One, you can control your messaging. You can go direct, and, uh, you can get in touch with those candidates directly rather than going through another platform. Um, and then second, that work that is still very time-intensive of, like, finding people and messaging them, um, essentially gets reduced by, by orders of magnitude.
- HTHarj Taggar
Okay, so one thing everyone's talking about right now is the AI talent wars and just how competitive it's gotten. Like, we have Mark Zuckerberg offering hundreds of millions of dollars to top researchers and engineers. What's your advice to founders, um, who wanna hire the best engineers a- and be able to compete with Meta for them?
- DPDavid Paffenholz
I think there's two things that founders can do, um, and every founder can do them. The first one is on actually being involved in that recruiting process. It's often lost in the Zuckerberg stories, but to me, one of the most interesting pieces is that he's actually texting and calling these researchers himself, um, and not the huge organization that he's built. And every founder can and should be doing the same thing regardless of, of what stage they're at. And, uh, the second part is the, the meaning behind that work. And so what is your company actually trying to achieve and how... why is that something
- 19:20 – 22:00
Lessons Learned
- DPDavid Paffenholz
that an engineer should care about or, or why should they contribute their time towards that work? That answer will be unique for every company, and it's also their way to stand out. The best founders that we see recruiting on Juicebox treat their recruiting very similar to their sales process, and so they have automated follow-ups. They quickly respond to candidates, be it positive or negative. They schedule the next step, and they can get a candidate from first email to, like, offer made within a week, and that's a fast process that requires a lot of dedication. We have to remind ourselves of this often, that, like, recruiting is our number one priority, and, um, and it's, like, an active choice that a founder has to make.
- HTHarj Taggar
Yeah, and I think it makes a big difference to candidates. It's like the whole sell of a startup is y- we move really fast. And so if you can show them that during the hiring process and they can trust it with four weeks to get through, like, the big tech process, um, you can really make a difference in closing.
- DPDavid Paffenholz
Yeah, and it's also the first impression they'll have of your company, so, uh, I think it really matters.
- HTHarj Taggar
You know, actually, something else that's a hot topic, um, actually, um, even internally at YC, is just, um, for hiring engineers in particular, how should you think about letting them use AI tools or not use AI tools? Um, I just heard that Meta are now letting engineers interviewing use AI tools. Do you have, on the front lines again, any thoughts on that, where the industry's going, and maybe how do you at Juicebox think about that for your own engineering hiring?
- IGIshan Gupta
Our interview is, is split up into two parts. Uh, for, for one half of it, you can use AI tools because that part really tries to assess whether you can build something from scratch, that that's really a culture we have internally, is we try to hire engineers who can, like, build things from scratch themselves without, like, existing frameworks and a lot of, like, existing infrastructure. Can you, like, really build something and think from a product perspective? So for that, we pretty much allow AI tools. There's also part of our in- interviews where we don't allow AI tools, which is more logic and reasoning based. And the reason for that is LLMs are, like, surprisingly good at that. Like, LLMs put one short easily, like, questions that are hard to reason through for, for people often. But those kind of questions, while they might not be directly ap- applicable to, like, engineering roles at early stage companies, I think they are still a good signal for how people generally approach problems, like, how do you approach a hard reasoning problem? How do you break it down? How do you think about sub-problems? Those kind of things are really important and I think continue to be important in the recruiting process.
- HTHarj Taggar
Tell us a little bit about the Juicebox or product roadmap. You have sort of multiple products. It's a, a product suite of recruiting tools. Um, how did you decide to build the products that you have at the moment in that specific order, a- and what might be coming next?
- IGIshan Gupta
The thing we do most commonly internally is we follow the recruiter. We look at what they are spending the most amount of time on o- and where can we
- 22:00 – 24:30
The Vision for AI in Recruiting
- IGIshan Gupta
s- save that time for them. So the first thing we built was our query builder. That's, like, the initial search process of setting the right filters. We noticed that this takes a lot of time for recruiters. How can we solve that? We built out, like, the query building part of it. Then we saw that people are really spending time going through every profile individually, so we built our autopilot, which kind of does that for you, and that's also the reasoning why we built the agent workflow because we saw that the way recruiters operate is they look at profiles one by one, and they kind of, like, share feedback on it and align internally of what a good profile looks like, and that is similar to the way our agent also operates. So our core philosophy is looking at what is being done manually, where can we save people the most amount of time, and building that out. In terms of our roadmap, that kind of translates into our agent doing more and more things. So we want to get to a point where you can really go from a job description to a first call booked with a qualified candidate where they can talk to the hiring manager directly with as little human intervention as possible, and that's what we're trying to build towards.
- HTHarj Taggar
Tell us about it, just if we zoom out from the product roadmap, just the, the vision for Juicebox. How do you see it growing into, like, a, a big, huge company, a- and what has to go right for you to make that a reality?
- DPDavid Paffenholz
A recruiting team or company, when they think about solving for more headcount growth, right now the default answer is growing their recruiting team or augmenting their recruiting team with a recruiting agency or additional headcount through an agency. Our goal is for that answer to be augmenting their recruiting team through Juicebox, where Juicebox becomes their main way of, of scaling the recruiting function. Um, we still think human recruiters are gonna be a big part of that, um, but we also think that the Juicebox agents will be able to do a lot of the work currently being done manually.
- HTHarj Taggar
Yeah, I think it's a prime example of where recruiting roles have, like, some work that people really like, like the meeting people-
- DPDavid Paffenholz
Yeah
- HTHarj Taggar
... and convince them to join and getting them excited, and then the work that they really hate, right? Like the trawling through, like, hundreds of profiles online and sending lots of cold emails. And so, yeah, it's a classic example where you free people up to do the human side that they enjoy and just, like, strip away all the automated work with AI.
- DPDavid Paffenholz
Exactly, and I think the side effect of that too is that we can actually help recruiters level up in the things that they do really well because we're taking away some of the work that, uh, they don't want to focus their time on. Uh, and so the time then spent on the candidates and candidate-facing side increases in quality, and, uh, there's deeper candidate relationships. There's longer term candidate relationships that just weren't possible previously.
- HTHarj Taggar
Okay, so maybe just to kind of close out here, like, you guys hit product market fit, fantastic round, um, learn a lot along the way. What's some of your, um,
- 24:30 – 25:50
Advice to Founders
- HTHarj Taggar
one bit of closing advice for anyone who's aspiring to be a founder and get to the stage where you guys are?
- IGIshan Gupta
I think find a good co-founder. That's extremely important. And, uh, second, be okay with the world of pain.
- HTHarj Taggar
[laughs]
- IGIshan Gupta
Uh, just stay alive. If... A lot of companies we're now seeing be very successful in AI, including us, started a long time ago, went through this phase where things were not working, but they stuck with it. They kept iterating, and they stayed alive. So I think that's the most important part. Would you-
- DPDavid Paffenholz
Yeah
- IGIshan Gupta
... do you agree?
- DPDavid Paffenholz
I, I agree with what Ishan said on the co-founder side.
- HTHarj Taggar
[laughs]
- DPDavid Paffenholz
Um-
- HTHarj Taggar
Right answer [laughs] .
- DPDavid Paffenholz
[laughs] And, uh, I think we also, like, we fight a good amount in the sense that, like-
- HTHarj Taggar
Yeah
- DPDavid Paffenholz
... we express our opinions, but it all goes on this, like, deep basis of trust. Um, and I think that's been really valuable and what I've appreciated most about building the company together. And then I think the, the kind of my added advice would be you can probably go one step further than you think you can before hiring. And so seeing how far you can push it can be a really effective thing, and it should feel like a little bit like everything is breaking when you bring on that first person, and that's a good way to test for it.
- HTHarj Taggar
Cool. All right, well, that's all we have time for today, guys. Thanks so much for sharing the story. Um, I'm really excited for the round. Congratulations again, um, and excited to see the future of Juicebox.
- DPDavid Paffenholz
Thank you.
- IGIshan Gupta
Thank you. [outro music]
Episode duration: 25:51
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode Tm5NbHUblkA
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome