Helping 1M people get a new job, with AI Apply. Yoann Pavy talks growth and marketing tactics from the Depop days to today’s world of AI.

0:26Yoann Pavy:Welcome to Curious Customer, the consumer podcast. this show is for anyone curious about the world of consumer from insights trends companies brands and design we'll uncover the hidden stories from the people at the forefront of consumer industries the show has a wide range of guests from marketing leaders to startup founders so whether you're in the world of brand building or the startup ecosystem or simply curious about the world around us then this podcast is for you my name is matt i hope you enjoy the show yuan thank you so much for joining the show it's great to have you here so you are a you're a growth and marketing guy you've held a bunch of senior roles across companies like deliveroo and perk box and you were the cmo at nude head of growth at depop as well so actually worked across growth and marketing at a good percentage of the uk's billion pound consumer companies and now you're doing a new one, right?

1:21So you're in the founding team of AI Apply, which is an AI powered job search platform. You launched that about two years ago, and you already have north of a million users. So it's safe to say that that's absolutely flying. So maybe let's start there. I'd love to just start with AI Apply's category itself of the hiring market in terms of what's going on here? What's happening? Why are people finding it so hard to find jobs? What's happening at the graduate part and in summary what's broken there yeah but first of all thank you for having me Matt that's great good to be here that is a great question I think we're in middle of a storm it feels like in the job market with a lot of things happening across industries parts of it because of AI parts of it because of the economy and wanting to like I don't know save costs or anything yeah it feels like it's getting harder and harder for people to find jobs even if there is more flexibility since COVID with remote working, which seems to be trending backwards back into the office now anyway.

2:22But it feels like there is less job for more people looking for it. So in that space where we are trying to help people find the best match for jobs as soon as they go live, to help them apply for this job as fast as possible in the best tailored way possible. And there's different things that we need to think about because markets are different. US companies are in slightly different ways than UK companies and French companies, etc. So there's like some nuances there, but we basically try to save people time and give them better matching in their search. Because we know the biggest pain point, whether it's a tough market to find a job or not, the biggest pain point is just applications.

3:09when you want to find a job you need to apply to a certain amount of them and you say you'll do a few tonight and then you don't do tonight you do it next week a lot of people procrastinate in that phase so we try to speed up the process to the interview phase basically so yeah that's what we do i joined a bit more than a year ago so the product was already out there and in the early phase with four people now actually so very small funding team but yeah lean and mean yeah yeah already growing fast yeah so can you walk me through the consumer journey what does that actually look like when somebody realizes that they need to change job or maybe get their first job what's that whole process and then we'll get on to the what's actually happening in the background yeah we're trying to get people to have what we call like a job application kit so cv resume a cover letter and follow-up email in like seconds so you find a job description you like or we find jobs that we think match your criterias and we'll give you everything you need to apply to that job in the space of a few seconds that's effectively our entry point to what people are looking for and then on the outside of that we will also help you with mock interview preparation using ai we'll have an ai tool app on mac os and iphone that would also give you live answers during interviews or during prep interviews for you to have like the best kind of guidelines to what you need to say a lot of people also stumble upon their words and what they want to say when they being asked a question during an interview so having like a guide or bullet points or like numbers that are directly pulled from your experience back into your into your screen so you don't miss a point that's very useful for people too and lastly we um we also have very advanced automation on auto-applying to jobs.

5:00So we do the job application kit, but we're going to also do the application for you. And there's another massive subset of people who really like that. All of these products are kind of puzzled together in subscription. And some people take the whole thing. Some people take some different add-ons. And that's how we play around with our business model pricing and to have the best unit economics to grow. When you think about the timeframe at which somebody wants to be successful, which is presumably finding a new job. I don't know what that is, a period of six months or something. What's the success rate that they have within that window?

5:33Yeah, that's in the pipeline for us. We've grown so fast. We don't have like a clear daily, weekly number on like success rate for like interview and stuff. That's something that we want to put into the tool actually, like some sort of like advanced user dashboard type thing where like we can tell you like we've applied to 100 for you. You've got an interview rate of 5%. The benchmark is 10. maybe you can do these things to like fill that gap i think there's like a ton of value in having this kind of information from back into the the user using the pool of users that we have but typically it would take for the best ones i would take a few weeks to get a few interviews and an offer like under a month and some people would take a few months but that is a representative of the job market like some people will struggle a little bit more to like find a job because everyone's unique with unique experiences unique personalities and there's definitely like a part of intangible which is always going to be difficult for us that to cater for everybody not everybody is equal on the job market obviously so some people are differently more successful than others what we can do though is whilst at the beginning resurfacing some of that information back to the user to help guide them into what would be like above benchmark type of conversion to an interview and then to a job offer.

6:48Got it. Okay. Okay. So what's actually happening in the background? So you've taken my profile, you've applied to 100 jobs for me, you've created that cover letter, then we've done this sort of mock interview, so I'm getting really comfortable. But what are you actually doing? Are you scraping jobs boards and job sites in order to do the application? Yeah, in short, that's what we do. That's where my technical knowledge is limited, but it's a lot of scraping across everything that's available on the internet. So we've got at this point more than a million jobs onto our job board. We've got a public job board as well within the tool, and that is refreshed every day worldwide.

7:23And every week we're looking for new sources of jobs and new other job boards, APIs, et cetera, that we can plug into. And obviously when you do that, you will tend to have better coverage in, for example, the US. So as a growth person, when we're trying to open up pockets of new markets, becomes a bit difficult when, let's say, we don't have enough Spanish jobs, for example, and our users in Spain are complaining because we don't match them enough with enough jobs. But that will come with us growing and going through the pain of adding more jobs, basically. But yeah, it's a lot of scraping. Okay.

7:59And so I'm interested on this. You've almost flipped the way that the hiring market typically works, which is B2B first, helping the hirer to find the right candidates. You've turned it on its head, right? So you're supporting individuals that want a new job or a first job. Can you help me understand that a little bit and perhaps also dive into some of the pressures that that might put onto the hiring manager? Just to give an example here, because the speed at which you're able to apply and high volume for all of your customer base, does that mean that they are one CV of maybe many more, right, on the desk of that hiring manager?

8:31Yeah, which I think is already a problem that's being solved by AI on the other side too. There's already solutions out there for recruiting agencies, recruiters, and just recruiting teams, internal recruiting teams to filter CV with some layer of AI. We already know about ATS systems that go through criterias. And if you don't meet certain criterias, you're automatically discounted from the recruiting process. There's already solutions on that side. So we, and I say we, I wasn't there at the beginning when that idea originated, but I think the idea going through customer first is that also you, it changes everything that we can do from a marketing point of view, right?

9:12Like a B2C approach to this is going to be very different than a B2B approach to, I think my preference of always B2C is more interesting. We can do more and more cool stuff on the B2C side. And I think the B2B side will come for us eventually. There's no reason why we shouldn't solve the problem on the other side, provide solutions that auto-filter applications or applicants that we send. And then suddenly it would become like a full-on marketplace where we can have different pricing for B2B, B2C and fulfill the supply and demand on both sides. I think if we do our job using AI and, you know, enriching like the data that we get from our users.

9:52When they're first on board with us, it shouldn't be a volume issue. Like we should do enough matching, enough great matching that you shouldn't have thousands of applicants overall. And it should be really just the best ones. Right now, it feels a bit like the Wild West because, yeah, there's a lot of companies who are trying to do what we do either on the B2C or the B2B side. And what I see on the B2B side when it comes to AI, There's a few in the UK that are starting right now, like raising money, etc. Like you said, it's more like the traditional, let's sign some B2B, some companies to give us their role.

10:30And we'll try and find them the best applicants using some like advanced AI chats kind of features. That is interesting too, but it feels more like a bonafide recruitment agency, basically, as opposed to real AI scale innovation. yeah can you talk about some of the future looking product pieces here as well like how you how you think ai might play out in this category and i'm interested in how it might be able to go beyond the cv and beyond my skills and experience in history and actually think about things like character and culture fit and referencing is it able to do that type of stuff yeah yeah so it already is and we're thinking about solution to do that too like we want to go beyond the cv imagine a world where a recruiter could have more information than the two paragraphs that you're able to squeeze in one page which is very restrictive like with with effectively i think cvs are kind of dead we have to play along those lines because that's how recruitment works still to this day but imagine a world where ai could scroll everything that's online from you and look at all the project that you've done all the podcasts that you've recorded all the youtube videos that you posted and give a more rounded kind of profile of who that person is obviously i'm saying that it opens up massive kind of worms of who are you online what have you done in the you know however 20 30 40 years you've been on this planet and posted on x if you had any controversial texts and stuff like obviously this will be a discussion but i think it's an interesting one to have for sure that's a sort of black mirror type future there yeah i think so i think to be honest i don't think we're very far from it i think we're already kind of in it even though it's not fully admitted or enabled by technology but i think it's going to very quickly be enabled by technology everyone is already googling your name when you're at the second stage third stage interview maybe not the first stage but further down the line when they do the curtain people check if you have an instagram account what you've done on google the stuff you said on twitter they're already doing that.

12:34It's just like AI will be able to do that for you and categorize it at scale for like thousands and thousands of people. And I think if we find the right interface to do that stuff, I think it will be great for both sides. But I think for recruiters, I think we'd really enjoy having those kinds of extra tools. So we still, I think there is the the notion of finding a job i think the bigger play for us potentially is to grow beyond applying and finding a job is the first three months into a job i'll do it further increase the utility of our tool to become like your career co-pilot from your first the first time you apply to a job all the way to the end of your career like across seniorities and i think there's like some important key stages further down the line that thinking about it from a purely from a business sense would like reduce our churn keep our customer longer i think that's those kind of steps is how you become a unicorn eventually by becoming more and you call it like verticalized as your retention driver and you've helped people through the hardest part of all of that journey as well but yeah my big takeaway from all of that is i need to go and check what i used to say on bibo instead of 2005 so yeah yeah talk to me about this like i've done my podcast for a year and a half two years like the ton of stuff i've said freely every week as uh probably scary i don't i don't want to think about it yeah well you're not looking for a new job right and and on that point okay so you've you've been building this business as you said four employees only three of you for the majority of the business's time talk me through the the growth journey right so zero to over a million users now what have you done how have you done that with so few people Yeah, so I think growth and marketing and building a brand has changed massively over the last few years.

14:24A lot of that is down to our online channels have evolved from being connection-based or human graph to interest graph type of algorithm, which opened up a massive opportunity for distribution for small brands. before to get seen, you'd have to spend six, seven figures on brand activations and whether it's online or offline. Whereas now you can find arbitrage to get the maximum amount of impressions for very little investment. And some people call it going viral, but I think it's one way to say it. I think some people think it's a joke or it's not real. It became kind of a meme, like telling your markets and make it go viral.

15:11I genuinely think these things can be engineered. And there is so many examples today in the consumer space, whether it's subscription apps of like 17, 16, 18 years old building companies, when literally them and their friends getting hundreds of millions of views on TikTok and Instagram and getting to crazy numbers. I think that's what's changed. And that's what we're doing. I think the notion of paid marketing in that kind of discussion can also be targeted as being like we've done all of this with zero spend or whatever i think that's wrong i think paid marketing is the close seconds nailing what you're doing on mostly google and meta is even more key than it used to be i don't really subscribe to the it's getting more expensive let's find other ways like it's only getting more expensive if you don't know what you're doing.

16:07There is still very profitable and you can still build businesses on month one profitability with meta ads. You just need to find the right product market fit or channel market fit or content that fits your products, that fits the channel. That kind of equation is very important. But yeah, that's what we've been doing. We've been working really hard on scaling on organic distribution as well as our paid distribution on the main channels that everyone knows about. So there's no secret. Like it's the same usual suspects. I just posted on X this morning, actually. Like I looked at the last 30 days of our organic impressions for AI applying.

16:47We got 52 million impressions in the last 30 days. Organically, that's not even ads. 52 million? Yeah. So can we talk a little bit about that, the organic side? What are you actually doing? What's your experimentation process? And how do you realize that the signal is actually something to really lean into versus just another signal that you should ignore? So it's never a signal that you should ignore. It's the exact same equivalent than having, I don't know, Nike paying millions of dollars on TV to show up and continue further build their brand. It's the same principle, but for new businesses that don't have millions of pounds to spend on traditional channels or TV and things like that.

17:30So what it does is that it builds repetition and top of funnel and awareness and brand. Because if you do that over time every month and you get good at it and you're getting more and more and more impressions, it fuels everything else. We talked about algorithm, like I mentioned algorithm earlier. I think having these engaged first touch point, engaged people into your content, whatever the content is, we can talk about that in a second, what type of content I'm talking about, what you're doing in paid afterwards. Like it will fuel those engaged audiences that Meta algorithm will look at to try and tailor like who to show the ad to.

18:10Same on TikTok, same on Google. So I think it's super valuable. I think you can put a proxy monetary value to it. It's very hard to track the ROI to these things. That's the key kind of like counter argument to when I mention stuff like that. Some, it was crazy to me, like some marketer would push back like CMOs and stuff. They're like, what's the ROI on that? Like what do we do? It's like, it's like, what's the ROI on TV marketing? It's really hard to track. You have to be confident in how much does it cost you? And do you think you're finding good arbitrage to give you a proper number to base it on?

18:45that 52 million impressions that we got last month or our blended cpm for that was 0.05 p that's absolutely ridiculous yeah it's ridiculous right like it's ridiculous like your cpm on tv or your cpm even on meta ads is going to be like especially if you're targeting the us is going to be in like the 20 30 40 dollars sometimes so it's insane like it's insane distribution for very low cost why would you not do it like that that should be the that should be the playbook and i wish i had the money to also do tv and like kind of blend all of these things together but it's like right now you do you do with what you have and you're trying to find the the arbitrage to to do it at your scale so the content i'm talking about can be different things for us it's like mini product demos where we would show the product in like videos that will be less than 10 20 seconds with like some shocking factor or you know like i caught my girlfriend or i caught my friend using this you won't believe it i'm simplifying it but this is kind of like the type of content i'm talking about and i feel like this type of framework for many many product sometimes the problem is the product is not built to be visually impactful in those type of content so you have to think about it from this point of view so when i think about what we're doing on the product i would actively think Think about what would that look like if someone films a computer screen right now to shot a TikTok, Instagram Reel.

20:12Will it have some sort of visual cue that would make what's on the screen? I want to comment or I want to send it to my friend or I want to say something about it because it's shockingly interesting, like it's funny or whatever it is. Like you need to find those kind of angles. And I think you can, engineer is a big word, but I think you can think about it this way and find your formats that way. And I think that is the source. I think that is still very early. I think there's very, very few companies doing that. The ones who aren't doing that very well right now, they're not really hard to find because marketers like to brag.

20:47And marketers and founders that are all into that space, especially in the US, are massive braggers. So if you find them on X, a good example is the founder of Cluey, Roy. like massive controversial guy would talk about all that all of that stuff is in the open there's no secret but um sometimes i wish it was a little bit more secretive and people would just not just see up about it like just the first day they find an edge they just talk about it just to get the clout and the kind of followers of the back of it because i think everyone is kind of building their personal brand of the back of that too but yeah i think that's the that's the playbook so once you have that that kind of format or few formats like that it doesn't have to be one like you can find different angles with this it's literally a numbers game you do know that film um moneyball one of my favorite films so it's like you know that the the term when there's the the scene where they're picking the players and it's like what did you pick what you what did you pick him he's not famous or that he gets on base it's the you need the volume of people getting on base to be able to win games and when it comes to that strategy it's finding creators that can have good content, not necessarily followers, but that you can train into doing these formats and then post every day.

22:02And then you multiply this by 10, 20, 50, 100 creators that would all post on new accounts to try and go viral every day with that strategy. That's how you find the scale with this. So it's not easy. It's kind of tedious. It takes time to build those kind of processes. It's not AI-led. like some people want to think everything is these days it's very granular it's very manual but once you start finding piecing those pieces together finding the right people teaching them the content then posting every day and then tidy up that process then it's a numbers game like how many more can you find can you build some sort of like pyramids marketing pyramids scheme within them so like the best one teaches the others the others recruit others and that kind of thing that's the playbook called this um so yeah that's a lot of my time fair to say that organic or going viral on tiktok say that is much more of a science right it's about finding that right content and then driving repetition of that as a sort of a summary of what you've said and then the second piece is actually thinking about your product as an acquisition tool rather than just the moment once that user has actually downloaded and enters into the product which it kind of takes me back to to depop so so maybe we can talk a bit a bit about that and depop had this really interesting social inherency where people would go on they would have their items of clothing and then they would actually promote that on other social platforms come and look at my wardrobe and buy so you had this sort of really interesting growth mechanic there can you draw on a couple of things from your depop time which which created the success that it became so i see as very different times actually i think it's very different so what you're describing here is there was a social element to a marketplace, which was very new at the time.

23:45And yes, drove a lot of growth because people wanted to sell their products. So they posted on social. But remember, it was a time when there was no TikTok. And Instagram was still very forward follower led. So not so much like it is right now, which is very much like TikTok, which will see stuff that you're interested in before you used to be seeing stuff that you were following. But there was this mechanic that was product-led growth. By the time I left, we were tapping into the social content element like that, but mostly it was mostly for ads. So we were, I think at the time we were the first one to actually do ads with sellers in very raw form.

24:32So like raw interviews, like it looks organic, like it looked more organic than the organic stuff that you would see on feed. It was very role, film with like cam recorders, like it was in the 90s. That was the, I think that was the edge. Lots of product growth and innovative paid formats that was very new for marketplaces at the time. What I'm describing now is very different because AI apply doesn't, at least yet, I'm thinking about it, but I think in certain categories, I think in the case of someone looking for a job, it's not necessarily something that you're going to share or brag about.

25:08Like it's for a lot of time, it's a period of struggle you're really trying some people for some people are successful some people it takes more time it's really hard to have those kind of mechanisms where people would feed back into social platform and share things so instead what i described before is kind of the equivalent to get like an edge and a ultra distribution is to find people are not necessarily users to the platform they're just really good at creating content and they're not necessarily our followers they're just really good at creating content. Because in 2025, anyone can create a new TikTok account with zero followers, zero video.

25:46All it takes is a few days of warming up the account. I won't go into much details, but you do have to warm up the account a little bit. And literally second video can go to 5 million views. That didn't exist when I was at Deep. So that's the clear difference. I wish that this existed when I was at Deeper because we probably would have grown 10 times the size it is now and probably wouldn't have sold to Etsy. But it's very different. I think it's a very different world. I think, interestingly, I haven't seen the AI replaced or replicated in some ways with a 20th of the employees and a hundredth of the investment and get to the size that they are like, you know, super fast.

26:27But I think it's coming forward at every industry. I think pretty soon we'll probably have like a new Vinted or Depop that is like AI first with like two guys crushing it, you know. maybe if you see it make sure you send it over to me i want to um i want to just quickly talk about consumer trust and your point in there on depop with this very authentic kind of camcorder hell type content that you were creating we're in a very different world now so we've got a lot of ai generated content coming out you're seeing influences i mean there have been some cases of backlash where it's so clearly ai generated content but almost now actually or maybe weeks away we're going to see stuff that it's very very hard to decipher between authentic imagery and non-authentic ai-led created content so i'm really interested in your view on the consumer perception towards that and and also within that the if you look at apps and app stores of old people would typically look at the star rating right and then actually we all realize that product people can gamify the star rating because if you have a good experience the product you surface that rate me out of five you get a five star so the trust kind of reduces like what is the next phase where are consumers going to go in future for actually really understanding and starting to trust a new product it's really hard to predict the future what i can tell you right now is there's definitely a big mixed bag of sentiment around ai being either like the most terrible thing that ever happened or this most magical thing that ever happened there's like this big like most things in society these days there's like two big extremes you're quite violent on both sides i think we're in the space what we openly say to our customer if you go on the website you'll say like ai didn't take your job but he can help you get one that's our tagline at the moment so we we're leaning towards it by saying to people look we know we're an ai company we know that there is fear about ai replacing you with your job instead we're giving you solutions to actually find one i think in other industries it's probably gonna live the other way where it'll probably be better to not mention your ai led still use solutions that are ai led but not necessarily talk about it openly and it being more background thing it's just adopted there's a lot of talk about that about ai being like um what big data was a few years ago where everyone was talking about like it was like a badge of honor, I think that would naturally probably fade away.

28:57We do think about it internally as well. Being called AI Apply is openly and really fully embracing the term and surfing with it. But who knows? We might have to change or dial it back down or find other ways as things evolve. But I feel like right now it's just a big divide. I see in the comments in our ads, there's people loving it and being very passionate about it. And also knowing that on the other side, recruiters are using AI to filter candidates. So they're like fight fire with fire. Almost some people made the analogy of Terminator as well. Machine against machine, you know. I like that.

29:32I like that. But yeah, who knows? And see six months, I mean, or like two, five years. I have no idea. We'll see. Yeah. Okay. So today you're still seeing skeptics, but you're also seeing the people that are adopting it for utility purposes, like finding a job. They're seeing really exciting success. Just to follow this thread then, answer engine optimization. So chat GBT usage and people are searching for a consumer product or a need and then something gets surfaced. And a very high level understanding is that a lot of that is actually powered by Reddit, by community forums where the engines feel like there's really high trust.

30:06What's your advice to brands, to consumer products today that want to be more present and want to sort of build that trust in the background? and maybe what do the consumers not really know that's actually happening there? I think it's like the next big agency industry being created right now. Like every new agency that's surfaced on my radar is one of those chat optimization engine agencies. I think at the core, it's the same principles as it was for SEO. It's a lot of being mentioned in the right places, in trustful places, and volume is important, apparently, from what I've seen. So it's really hard for a new brand, same as SEO.

30:44to be surfaced as an answer by Chargipity or others. But it's not impossible. So take us as an example, being just a couple of years in and really one year in scaling, we are starting to see some traction in being mentioned and getting customers through it. Same as SEO, you'll never mention in isolation when Chargipity will give a solution or give like a few options. There's yet to be a paid solution, which might come at some point where you want to be the first one to be mentioned and maybe they'll start monetizing, same as Google did, to be surfaced. I think right now it's just a mentioned game and you said it, like Reddit to be the main source for LLMs to look at what people are interested in talking about and kind of like guessing the answers that people are looking for certain questions.

31:35So we are experimenting. All I can say is, as you can imagine, there is a very legit way to do that and there's a very quickly, very non-legit way to do that and automate that stuff, which I think in the last few months, Reddit has made big changes in their guidelines and bots and automation tools that would infiltrate subreddits. So it's not easy to do that. And that's definitely not something that is working for us at all. But there's probably some organic ways, especially if you're really just starting and you're looking for your first like 100, few hundred customers, there's a very legit organic way to leverage that on reddit get some customers get some exposures and get some mentions by tragipities and others as a new brand really helpful yuan because we're a consumer podcast i'd like to ask everyone if there's a consumer brand product or a band or an artist that you want to shout out then please do yeah yeah i'll share i really like music as an artist I discovered literally a few weeks ago that was so good that I booked tickets for his concert next February in London already there's a guy called Rusowski his name is definitely a nickname there's a guy from Madrid and his album is called Daisy it's like a yellow with a gorilla on it it's like a yellow square with a gorilla in the middle and it's amazing what genre so so good it's kind of like I don't know it's kind of pop hip-hop it's what it's like the guy's an instrumentalist and he did like a tiny desk on npr and he plays literally every instrument as this falsetto voice and is the most crazy outfits i don't know it's one of those new artists when like if you see him you're like there's something special it's fresh it's new so yeah i highly recommend having a listen yeah i get myself a ticket i'll see you there yeah yeah come true you want thanks so much for joining the show really appreciate your time amazing thanks for having me bye mate

33:43and that's it folks thanks so much for listening as mentioned please do provide any feedback or ideas of future guests via the website curiouscustomer.co and please do click subscribe if you did like it because it really helps improve visibility in the algorithm and makes the podcast more discoverable amongst a sea of great content. So that one click does really help. My name's Matt. I hope you enjoyed the show.