Building AI Apps: From Idea to Viral in 30 Days

I think there's going to be thousands of mobile apps built with AI, but also have AI that are going to be a million-dollar-a-year businesses. And I found, uh, a founder called Sherry Jen, and she vibe-coded this app, uh, built it, built it in public, got thousands of customers to it, and I think she's a first-time founder. So I'm trying something new. I brought her on the podcast and I said, "Hey, what can you teach us about how you built in public, how you vibe-coded this thing, um, and just your your thought process for for building this company?" And I think a lot of people could copy her blueprint. That's why I wanted to bring her on and just learn a lot from it. Uh, one of my favorite parts of the episode was where she showed us how she uses ChatGPT image generation to actually create memes for her startup. So, uh, interesting episode. Um, grab a notebook, cuz you might want to take some notes. Thanks for coming here. Thanks for being here. And I can't wait to see what you build. Have a creative day, friends. I'm rooting for you. [Music]

We got Cherry on the pod. Sherry, I'm actually super excited to learn from you. If people stick around until the end of this episode, what are they going to take away? Number one, sometimes you can just take a first step to getting started. Whether it is 3 hours of vibe coding a prototype just to test something with users or writing your first like, "Hey, I'm building this app and this is day one." That's that's all you really have to do sometimes. And that compounds more than you would imagine. And Sherry's out there. I noticed her cuz she's getting thousands of downloads for her app. She's building in public. She's vibe coding. So, she's a testament to it. Um, Sherry, you brought a little presentation. I want you to go through it. I want you to teach people. I want people who are solopreneurs, who want a side hustle, who have a business but just want to AI it. I want them to come away with feeling a little bit stronger, a little more excited, a little more fired up and have some tangible product design, consumer psychology, build in public know-how so that they could grow their app, consumer app. We're going to be mostly talking about mobile apps faster. So, let's get into it. Sherry, thanks for coming on.

Awesome. No, I'm super excited to be here and I've, you know, been a fan for a while. So, I'm super excited for the conversation. Uh, so, uh, just to do a little intro, so right now, um, you know, I've been a founder for four years now and currently the product I'm working on is called Peak. Um, been building in public on X for it the last, uh, I guess four months now. Um, but basically the idea is we want to revolutionize the way that, you know, personal finance has really been, um, from dashboards to like more dynamic UIs that actually make it feel much more enjoyable and bite-sized when it comes to making some of those decisions. But I'm not going to go into like, you know, the nitty-gritty of the, you know, product just yet. Um, I actually want to talk a little bit about the process actually from from like getting from zero to, you know, something that's launched on the app store. Um, and you really, you know, can just get started with a vibe-coded prototype to start testing with users on day one. You do not have to have something that is perfect and beautiful. And I'll share some examples of what the very very early versions of of Peak look like.

What I what I like about Peak is it's a peak, no pun intended, to a whole new wave of consumer mobile apps. Like there's like a thousand like peaks, peak for X ideas, startup ideas that can come out of this. So I want you to go continue and and sort of show show what the product is.

Absolutely. Um, yeah. So I like you know talk a little bit about like you know some of these like principles that you know we were thinking about when when designing the product and then I'll walk through like two evolutions of you know what it looks like and how that translates. So um you know when it comes to like how humans process information, right, uh and just like actually how LLMs also process information, if there's too many tokens, um you need to have information chunked in ways or else people get overwhelmed, right? So I'm just looking at the, you know, we call it sauce on this on this podcast, the sauce on on this slide here, so from a product perspective, um we show only one takeaway at a time on screen, something like cards. I think that's that's really smart, like basically what what all the things that you've described here: progressive disclosure, um single purpose, conversational UI, the third thing is the daily reports that can be read as a part of a check-in or narrative story that's easy for the user, and then the last thing is um helping nudge users with relevant information at a time and provide practical first steps. These four behavioral science uh principles could be applied to any product.

Absolutely right. And and that's why like um I like the point that you made earlier around how um we're not here talking about how to build a finance app. We're actually talking about how consumer apps that are AI enabled that help people in some part of their life, whether it's to be healthy, be productive, be financially secure, right? They all they all ultimately follow some of the same principles, right? Um, but of course the data that gets piped in and you know the subject area that that could all be different, but it's all the same at the end of the day. And even when you like, you know, create a podcast or, you know, you you want to create content, it's the same thing, right? Because you're you're there to like bring like here's the audience's brain and understanding how they process information and you're bringing it down in a way that they can understand. Same thing, right, that is applied here. So I can maybe move on to actually showing you what we started to test um how these kind of you know like start to manifest in in a finance product. So this is a very these are very very early versions of something I did in Next.js. Um I I basically uh used Cursor and Vercel uh to code this out. As you can see, uh it is very very rudimentary and no pretty designs. But the main thing that I actually spent a lot of time testing with users is this area. We called it AI templates. Horrible name, but later on we call it check-ins. But this is the part where it's it's it becomes quite different. Um, and this relates a lot to firstly this whole context switching fatigue principle. So when you click into these, you basically ask like a single question, right? So instead of having to like bring in you know information from all different places like some you know Reddit post someone made here or like oh this you know WhatsApp conversation I had with a friend who knows finance. It's all in one place so you can kind of like step through the problem without really really getting distracted. Um I can maybe try to show you what the like the chat I actually have it pulled up. It is uh very scrappy. I can just maybe switch tabs. Um, so can you see this? Like I have like the console log and everything on the Oh, okay. So the way that like this really simple one worked was start conversation. Uh, it's actually powered by Claude 3.5 on the back and it was this really simple interaction that I used to test with users. Um, yes, food spending. Yeah, this is cool. Um, this is this what's cool about this and the lens I'm using for for everyone is this could be applied to a million niches. There is and I'm curious like have you thought of other niches where this could work like peak for X.

Yeah. I mean I think that uh like nutrition is definitely like eating and uh eating diet health kind of all all sort of related, right? um now I'll tell you like a habit I wish I had that I do not have because it takes too much effort um I want to be more specific about like the types of food that I am eating to like reach my goals, right? whether it's like oh I to like build muscle for this like competition I'm doing in the summer or maybe for others it's like oh I want to cut, right? um now getting like uh everything tracked and then being able to like personalize the goal then being able to like plan the meal and then figure out where to buy the meal because you know you have to go out and get groceries, cook whatever and then like give continuous feedback around how the plan is going, right? because it's iteration like you it might not be perfect the first time. Like I would love that, right? I mean, if someone's building that today, like I I would love to take a look, but it's the same kind of concept of like, you know, you you have to like capture not just the tracking and the numbers part. It's like there's so much more so many more steps to ultimately getting to you taking an action and then seeing the results of the action. So, I think this applies to so many different so many different areas where um you know where there's some level of like improvement that we're achieving for as human beings, right? Like totally what would it be for you, Greg? Like is there something right now?

Yeah. Like honestly, right now I want to get really good at tennis. Yeah. And um there's a lot of systems that I have to put place in my life to get good at tennis. So I need Peak for tennis. Yeah. Yeah. It's like um essentially like you know we have really busy lives, right? Like I think a lot of our executive function um basically this prefrontal cortex, right? Is spent on like you know our day-to-day jobs, right? Like for me it's like you know running a company. You know for you it's like you know running your stuff as well. For the stuff outside like we we it's often times we're just like too tired to think, right? there's like a a a quote to how much we can like plan and think and do all of that. And so that's why that happens a lot, right? We just we just don't have systems in place even though we're fully capable of those systems, right? So in a like mental resource constrained environment, how do we help take away that thinking that planning portion from the human? So they get you just get told, hey, if you want to get good at tennis, here's a 10-step part to your first stage of mastery, right? And this is all you have to do every single day. And maybe one of the frictions of getting good at tennis is like, I need to find somebody to play with who's at my level. Like I think tennis is really fun, but like I don't want to play with someone who's like so good that like I'm just boring them cuz like I suck, you know? So maybe like part of that is like a social matching platform where like you have scores where like you know you kind of know someone's like at a similar level as you, right? And then you get to you're in the same area, you can go and then get matched and you play somewhere like that reduces the friction of like oh how who do I know in my circle who's like at my level? So it's it's all about reducing that friction having like AI. So like basically work for us so that we can become like better versions of ourselves, like you can be like a great tennis player and I can like finally have the diet that like I really want to have that I definitely am not following right now.

Amazing. Cool. Let's go back to uh the slides. And so you vibe coded this prototype in less than three hours. And then what next?

Yeah. Well, so I then started to just like do user testing. So I uh I I I uh had this thing I call a rule of six, which means talk to six people that you think is your in your ideal customer profile. Too less is like you don't have enough sample size. Too many it's like you just don't move fast enough because like you know people are doing this on their schedule. So I just started to talk to people. Um, I pick people who are like in their 20s because um I also think uh 20s is easier when it comes to habit formation around finances because you're just starting to earn a paycheck. You're just starting to make like big financial first decisions. So, I just talked to a lot of 20-some year olds about money and then they started to use the app, right? I just like didn't really guide them. I was just like, "Hey, just like start using it and like let me just see where your eyes go to." And um it was quite interesting that I started to see a lot of parallels which um come back to a lot of the hypotheses that we had around the behavioral friction people have around managing money. So I put here like two like you know quotes from the users we spoke to. Um one person said they you know have used dashboard apps, right? by their bank, you know the Chase apps whatever. Um, and most of these apps feel overwhelming. Lots of numbers and info, but like I don't know what to do with it. Um, then when they started to talk about the experience using the the chat and Peak, they were like, "Oh, this feels like it's helping me actually get to my answers and it feels more dynamic than this like dashboard that I know is being shown to like thousands of other people, right? where like again because it's so standardized you're making me do the thinking around how to interpret it versus like you know the conversational AI doing a lot of that work.

Yeah. So the rule of six, you get feedback, you start getting insights, these insights by the way relevant for anyone building an AI app. What's next?

Yeah, you start to build. Um, so, uh, we basically took this, um, and spent about one and a half months actually building it out before we launched in the app store. And, um, for anyone who's launching a mobile app, uh, you sometimes get asked like, should you do TestFlight, should you do like the pre-orders, like do you do a waitlist? Um, we just like sent it and we're like, you know what, like let's just get it out there, right? And um you're never going to feel ready and it's fine. There's going to be bugs, but that's okay. That you don't get real feedback until you have real users. So this is what the app looks like today. So obviously a lot more uh you know design elements in here that are way more colorful than it was created. But you'll start to see like there's actually parallels, right? This this page essentially was this, right? But we like added more um you know cards that felt like a lot more readable. We added these like stories basically which are like you know nearly daily check-ins. Now uh this is a mockup so I realized uh you know there's some dates that look kind of funny but uh basically these are like almost like Instagram stories, right? Um so we uh so on this screen um we did something where we wanted to like take what was familiar with two people in other apps. So like these like little circular icons are familiar in Instagram, right? Because it feels like this, oh, like there's a little story, right? It's a narrative going back to just Stalt's principle of like let me give me something that feels narratively relevant to me, right? Organized information and storytelling. So we created these like daily stories, right? That look like what you see on Instagram. Um, and then down here, these cards we actually modeled off of like a lot of wellness apps, right? like meditation apps or like uh journaling apps where like you click and then they'll have prompts of like in the vibe checks question there's a vibes checks card there's a question like hey like you know um what's what's my financial standing like right now, right? it's like a vibe check so um we used a lot of the you know things that make people feel very calm um and applied it to money where people don't always feel calm, right? and some of our users actually told us that um They feel like they can open up Peak a lot more because it doesn't feel as scary as like their banking app like with red numbers being like, "Oh, like you spent so much more this month." Like here, like they're like, "Yeah, I feel more comfortable because of how friendly it feels." And it doesn't feel judgmental. It feels like it's trying to help me. It's like here's where you are and we're going to get you to where you want to be.

Yeah. So if you're building an AI app, think about well I think you got that insight from the rule of six and and getting you know so step one like in a world where anyone could vibe code an app and the hard part isn't really the software. It's really about the insights that you can generate from your target audience or community and then thinking about okay what are their pain points? Um, you know, one thing, by the way, you can also do is you can go and simulate what people's pain points are by just going to like ChatGPT and being like, "Hey, like what are the most common pain points with an app like Mint.com?" And that's called like synthetic user testing. So you can do synthetic user testing and real user testing, which by the way obviously real user testing is better but it's sometimes it's just easier to prompt and get something and then you reflected it in the product with UI that feels familiar in a calm design style. So, I think that, you know, you're even seeing that with Airbnb's like recent redesign when they added like their new icons. Um, like it feels calming. It feels like fun. It like people don't want apps that are black and white anymore, you know, soulless like brands anymore. They want fun brands. They want calm and and then intelligence. Finally, intelligence, artificial intelligence is the foundation of the experience of the app. So every like all these uh you know entry points uh as they say in product design um are designed to get people to interface with the intelligent to get more smart to help you out more so that it's becomes a a partner you know not a piece of software.

Exactly right. Um so something that we have um as well is we have a vector database that actually stores memory from the user. So we have like actually thousands of signals that we save down, right? from like you know do you have a sibling, right? um what are your values or um you know how did you feel when you spent on you know fitness maybe if it's a you know higher spend category because you're going to Equinox whatever in San Francisco but if you care a lot about it and that's a priority we we remember that and we're not going to be like oh you know reduce spend here we're more like maybe there's other areas you can find efficiencies, right? so um It's I I think that like people's expectations of products as you said is going to change, right? with all of this available today where they expect things to be emotionally resonant and like taking into account not just function but feeling, right? like I want to use something that makes me feel a certain way because you know that affects how often you know I engage with it. And then to your second point, right? I think we want to start using things that feel as humanlike as possible, right? And you know, we're getting to a stage where it's like you're talking to someone who is almost like a a smarter best friend or maybe a best friend that like has spent 6 to 12 more months in this area and is like able to like actually give you good advice and suggestions to like help you get better as well. And and none of this was possible I guess in like this kind of scale and speed years ago, right? Uh right now it's like you can whip up I mean you can probably whip this up even faster than we did, right? I think you probably whip this up in like maybe a few weeks, maybe a week depending on like how complex your app is and like it's it's truly amazing. It's just such an exciting time to be building mobile apps right now. And uh before we get to your next slide, you vi you when you say how did you build this?

Yeah. Yeah. You what what tools did you use? What stack? Yeah. Yeah. So for like the front end and the back end

It's been in Typescript and then for the mobile app, uh, because we wanted to be like more platform agnostic across, you know, iOS as well as Android, uh, use React Native and using Expo for that. Um, so yeah, pretty typical. I think we uh, you know, don't do anything too like fancy outside of the the React stack.

Um, as for the LLM side, we actually string together a combination of uh, Gemini as well as Claude. uh, depending on like the I guess the function, but a lot of the initial insights, like these kind of uh uh I would say like more detailed insights, we use 37 on Claude because the reasoning is freaking fantastic. Um, and then elsewhere where it's like uh speed is required more, we rely more on on Gemini, so a lot of the coordination on this chat, this is Gemini plus Gemini tool calling essentially; that's that's going on. Cool. All right, let's let's continue. Yeah.

Um, so yeah, that's uh the gist of like our journey on the product side. Um, we've been out in the market for just a bit over a month now. Um, and uh, it's been super exciting. Um, we built them public on Twitter, which I can, you know, go into detail in the next next part, but uh, been really like uh grateful and overwhelmed by the response, and we've had like thousands of people download the app. Um, our most active users are telling us they use it every single day. Um, you know, they're checking their balances, they're doing like these AI check-ins to like understand their finances better. Um, it's still early days, but it's like so we're already seeing how um people are using our app uh nearly seven times more frequently than like your desktop dashboard app. And they're talking more about like the the habit formation that really is, you know, coming into play when you have this kind of experience.

How do you go from zero to thousands of customers using your product with basically no audience by building in public? Can you in this section, can you teach us how?

Yeah, absolutely. I mean, it's it always sounds kind of scary and impossible, right? Um, there's but really it's not to this audience. This audience, not to this audience. This audience has is bought into the concept that building a public works, but how how do you do it? Amazing. Okay, I will I will go into how I do it. Um, I I'll actually just jump into uh like what I what what it kind of looks like timeline-wise um for me uh really just in the last few months. This little jump right here is I think I responded to some guy's tweet about college uh not getting into certain colleges. So I remembered this one. By and large, you can see like there wasn't a lot here. And this is right here was actually when I made my first build in public post for Peak. And I know cuz it was actually on my birthday. And for a long time, you might not really see much, right? And I think that's where like you you kind of have to feel okay with um like posting and not getting a lot of engagement and impressions at the start because that is normal, right? Um, but you just have to find one opportunity where somebody catches wind of your post or your reply and it really just compounds from there, like visually like you can kind of see it. Um, yeah, it started to happen around here. Amazing. And can you maybe go into like some of the post? Oh, there. Yeah.

Yeah. So, um, this is the very first post, uh, I made uh and as I as I was evaluating it, I was like I think I got better at like figuring out how to post about things as well like but you know it was one of those things where I was like let me just get the first one out there. So this was done in March. Um, as you can see, it's uh it was actually 4500 initially in the the views, but I think since my account has grown, it I think a few people have come back and looked at my first post, maybe. Um, but yeah, I nothing crazy. I basically said uh day one building the next big personal finance app. I was like, okay, maybe I'll make it like a journal thing. So, like it keeps me accountable that like, okay, I'd do day two, day three, day four. Um, and then I say like, yeah, one of the biggest reasons why I built Peak was because running my own company made it harder for me to manage my money as disciplined as before. Couple with, you know, ADHD felt all over the place. So, and I actually uh have this like URL spend. That was actually a landing page that I created for this like, you know, very very early uh version of the app uh which we we uh you know, re-redirected and took down. But uh yeah, basically it was very very simple, right? This this challenge. Can I break this down? So, uh you know, I'm on X a lot. I get hundreds of millions of impressions every single year. And this is an incredible tweet. Um, and I want to break down what why it works and some tweaks I would make. Yeah. Go ahead. So, if you're building in public, um the the best thing you can do is get people to follow along in in your journey. So, the the smartest thing about this is you you're like day one building the next big personal finance app. So, it's like, okay, I might follow Sherry cuz like I want to see where this goes. People don't like not knowing the ending of the story. It's just human beings. So, it's like okay, I guess I'll follow you. Finance app. We all know that none of us are happy, ecstatic, or I shouldn't say none of us. I would say most of us, yeah, lots of us are unhappy with um at least 50% of people are unhappy with the financial tools they use. A lot of them are using like antiquated bank old school 200-y old bank banking apps that look like they're designed from the '9s. So building the next big personal finance app is aspirational. So you have a journey meets aspirational and then you're just you say honestly which makes me feel like you're a real human being, you know like we're texting. Writing like you're texting is super impactful on building in public an X. The only thing that I would Oh, and you say that like what you know, you know, run because running my own company made it harder to manage my money. Like you're you're a bit vulnerable. Being vulnerable also really works in in building in public. The only thing that you're missing here and this is like a personal thing but um if you would have said like I suffer from ADHD and like that would be for the people who have ADHD extremely relatable. Um, also having a link in here suppresses the the algorithm, but other than that this is like a nine on 10 tweet is really good.

Can we move on to the next slide?

Yeah, for sure. I know this gets like talked about so much of like you just got to show up. You just got to post. And I know like everyone's like but everyone says it. I've been doing it for a while. Like I I'm not going anywhere. It feels like I'm stuck. It's nonlinear. And I like could not believe when this day happened. And and actually I only found out about this because my uh I mean our app was melting essentially like I I was getting all these pings from Intercom like this this and this user like has filed a ticket or whatever and I like messaged my co-founder, right? And I was like um are there bots? Like what's going on here? It's like this like moment where you're like is this a bug? Are these bots like it could cross your mind, right? And then I was like, "No, I I think these are real users." Cuz we were like looking at like what people were doing. Like, no, they're actually trying to connect their bank accounts. Like, you wouldn't do that if you were a bot, right? So then I started to trace back and that's when I realized I think I was one of the first people that resp happened to respond to this one tweet that this guy Preston um posted. Um, he posted uh I want a finance app with three numbers. What I spent today, what I spent this week, and what I spent this month. And I don't think he knew it was going to go to like 6.5 million views. And certainly I didn't think like I I reply a lot on Twitter a lot. Actually, that's another piece of advice which is like when you're you know early on like reply a lot and then at some point your post will carry you more. But anyways, I was replying to like you know a lot of threads that I felt like were relevant, right? So, it it just so happened that my one response here got picked up. Um, and uh I added an app store link. Um, and uh I said something like you get a cute mascot that roasts you. Um, and then yeah, it it just kind of took off. Uh, and you never know when it comes. It this one came like a month into Build in Public. So, like I started what March 6th. Uh, so this was like a month and 3 days. And uh yeah, this is the first time like I hit some kind of inflection point building in public.

Yeah. You also what's good about this is you weren't like download my app. It's you know I think people don't like pure promotion. I feel like you humanized it when you said you also get a cute mo mascot that can roast you in an AI chat interface if that's your style too. So again, going back into going back to write like you text and write like you're speaking to another human being. I know that sounds obvious, but people they they get all corporate when they're trying to do this sort of stuff and that just doesn't work.

Yeah, I completely agree. Um, I feel like the so even the way that like um I write stylistically is um I I make it very conversational like a text message, right? It's like um you know I don't use capitalization. Uh I uh you know add little things like if that's your style too, which I don't know if I would have used back at Google. Um yeah, a little cheekiness. you know, you're like, it's kind of how you would talk with a friend, right? A little unfiltered. Um, but yeah, that's that's like actually my norm like how I talk here. Um, this tweet is just how I talk in like real life. There's not a big difference. Love it. Yeah. All right. What else?

Yeah. So, um, yeah. So, this like kind of set things off and then, um, I was going to show you like one specific tweet. I'm actually just going to jump to like my Twitter cuz I don't think I put the uh I'll share this tab. So, this tweet was basically the first day that I actually on my own post got more than like a thousand. And I mean, the algorithm does have an impact, right? In a way, it's like if you have like a moment like that and you kind of catch it and you post anything afterwards, it kind of like grows from there. So, I followed up, right? I was just basically like, "Hey, you know, there was a bunch of downloads I was not expecting posted by this person." Uh, and then we actually made some changes to the UI because it's not three numbers, it's four because it would have looked weird with three. I had a debate with my team on that. Um, but yeah, we we kind of shipped it overnight. We were like, you know what, there's probably something to this whole like simplicity thing where like people are obviously reacting to the the bloat, right? The Frankenstein dashboards where you're like, sure, like I can look at my cash flow in this fancy sinky chart with 12 colors, but like how do how much did I spend today? You know, like that's all I need to know. So yeah, we just like ship this right away. So um that's like I mean the pros and cons of build in public, I think it's a pro. like you get to move really fast. You like see feedback and then you build it. You get like you get reactions. Um obviously the the downside is like if you're you're you're kind of accountable to things, right? Like you're like, "Oh, I I see this. I want to work on it." Then people expect you to put something out. But I like it because it kind of keeps me, you know, on track and making sure that like I'm creating the best app for our users.

So there's a few there's a few ways to build an app that gets a lot of traction. One is trendjack. So you look at what trends are are going to be and then you like build for those trends. And I think that's like I think it's that's a huge way to do it. The other way which you know is something I I think about is tweet jacking. And obviously you think about that. So you what what you really did is you saw a tweet that got a lot of traction, millions of views, and then you essentially vibe coded a an, you know, that tweet into your app and then you latched on to all those people who wanted this product and then all of a sudden they're like, "Wow, this, you know, peak is for me. this is a product for me," and you can only do that today. Like you, you know, in a world where there's AI and you can ship things overnight, you can only do that today. So, I think this is brilliant. And then I love that you you included like the journey day, whatever you put up there. Um, yeah. So, day 22. 22. Yeah. Yeah. This is cool. I like this.

Yeah. So, I mean, this is more tech stuff, but I thought this could be maybe useful for people who are like, "Oh my gosh, I have to post every single day, like, how do I come up with ideas, like all that stuff, right?" Um, so I kind of came up with like things that I enjoy talking about almost every single day. And oftentimes these are things I'm already talking to friends about, right? So, it's not like I'm creating new behavior, it's just like it's another medium. So, um, you know, we already talked about the daily updates, right? Um, but one thing I'll add here is, um, I did start out with like, hey, you know, building the day one of building a next biggest personal finance app. I realized like, hey, that gets used if I use it over and over again. It's like, okay, like what's new, right? So, I actually started to um make thematic challenges during certain time periods. So, uh, I started like, uh, another one I called like the 3x retention in 30 days because our week three retention was not great at one point. Um, it was like 13%. Which you don't want as a consumer app. Not great. So then I was like, you know what, I don't mind. I'm just going to tell people our retention needs to improve, right? This is the number. Here's where I want to be. Here's like how long I want to take to work on this, and here are all my steps I'm doing it. Um, and uh, funny enough, we did two things that actually shot our week three retention up from that uh, from like 13%. essentially like 28% in like a short period. Um, we uh, improved onboarding completion, literally just asking people, hey, it's been 3 days, just connect your bank account and finish. And 50% of the people we just sent a reminder did it. And the second thing is we started to send uh these context specific notifications. So it's like uh like uh well we'll basically look at their behaviors right after a time period and we'll be like oh it's uh it's it's a weekend now. Um you know you might want to you know take a look at like you know your you know spend buckets right in in these particular areas like entertainment etc. Uh so it's like related to the time period and um that also brought a lot of people back into the app. So um we actually completed the challenge in four days which was kind of funny actually four or five days. Uh so I was like okay I mean I'm not done like working on retention by any means. It wasn't like okay we're done but it was kind of hilarious that sometimes like it's the little small things that actually can make a difference.

Um, totally. I mean, I think what what you're saying is if you're building in public, you might want to have one really big goal. For example, I want to make a million dollars ARR in 12 months and I'm building the next big tennis app. Yeah. And that's a really big goal and you can remind people of that goal, but you need to have sub goals. Mhm. Exactly. So if you want to bring people along for the journey, you need to have a set of sub goals probably every 30 days. Yeah. Where people are bringing you you're bringing people along and it's like okay today's cha uh you know this month's challenge is going to be retention and then okay day 31 happens. You're like okay new challenge happen. I'm trying to hire a designer and I you know and I'm going to go. So e every startup has a set of challenges and and focusing it around every two weeks or 30 days I think is really smart.

Making memes around personal finance talk more about that and talk more about your process for creating them.

Yeah. Um, so I would say like you know don't make memes for the sake of making memes although it's it's super fun but um you know one thing that um you know we all know is like just talking about finance itself without like thinking about the emotions and the context around it can be very boring right so making memes is kind of my way to figure out like what is it that like really like people feel like some kind of way about right, emotionally when it comes to money. And so I started to create memes actually to like, you know, get uh get that kind of I guess feedback from folks. And maybe I'll show some examples again. I'm just going to go to my ex because I think it's just easier to to share. Uh all right, I'm on my timeline. Okay. Uh so I started to post um like I I do stupid stuff here. I mean, this is pretty like just silly. This is a, you know, this is from the latest Avengers movie. So, it's a bit of the trendjacking, whatever. Like, I was like, maybe people, you know, get the reference. So, this is a a superhero or, you know, villain uh in the movie called The Void, right? So, basically uh you know, you go into this feeling of nothingness, right? Um, so I just like, you know, took a couple minutes was like maybe there's some joke that I can make about this and uh I was like, you know, like I feel like I am really bad at managing subscriptions cuz it's one of those things you just like pay for and forget and it just like goes into nothingness even though it's not really nothing cuz I'm paying for it. Um, and so I used like that kind of a you know like as a tagline, right? and then added this image which uh I just bas I think I prompted on this is GPT4 by the way. I prompted like New York skyline and I took a screenshot of like some you know picture of this super villain and I said put the peak face right here. Um and I use this to basically like kind of get more and more like uh arena brand awareness isn't it? Yeah in a way. And I I've been doing a few like a few of them like I did let me see this one was also yeah like wall of shame. I mean yeah I actually looked really smart. This is so smart. Like what you're doing is basically you you know some of the pain points of the audience you're going after, but you're packaging it in trends and memes so that the algorithms love you. and you're using like not enough people are are using like image gen and some of these tools to have an unfair advantage. Like if you're building in public, a meme a week a meme a week is going to be helpful. Like you will get thousands of impressions. Yeah. If they're good. If they're good. And they're they're not necessarily too hard to do. I even actually have this post here cuz someone was like, "How do you do it?" I'm like, "I think it's pretty simple. I actually recorded. Yeah. Can you uh make that full screen and start from the beginning? Yeah. Yeah. Uh maybe. Okay. So, let's see what your prompt is. Can you create Can you create Can you create a peak cartoon meme around the latest Thunderbolt Avengers movie, The Dark Void scene subscription

creep? Insert clip of lightning-lit abyss that place your 1399 free trial keeps teleporting to.

Okay. Yeah. By the way, um, this line I actually uh got from 03. I was actually using uh 03 to uh brainstorm like trends and then and then thunder uh I I saw the Thunderbolt movie and I was like, "Oh, that was one of the ones." So I I actually copied this from 03 into 40 and I just added this part at the top, right? Can you create the peak cartoon meme? And I added the logo. Um, this is the logo asset.

So what you're saying is you can even use these LLMs to give you ideas around what memes to create. You don't even need you don't even need the ideas themselves. So use use uh use chat GBT as a meme co-founder.

Yes. And start from there.

Absolutely. I mean the the one thing you still have to have is you have to have taste. I mean they'll give you a bunch of ideas. Like nine out of 10 are You have to decide like which one like, you know, catches people's attention is visually suitable and can relate to some message around your brand. Right.

I've actually never tried this. I've actually never tried to use uh I've never tried to use chat GBT as a way to come up with meme ideas. So, I'm going to try this after. I think this is cool.

Yeah, you totally should. Um I I I can uh see what my prompt was when I asked it for memes. I I think I said a few things um in there. So like pop culture like um cuz we're reaching an American audience. I was like sports, you know, sports stuff kind of works, celebrity gossip. So I gave a few categories as well. So it wasn't just like make me memes. It was like I I I added some details in it, but I I can send it along if you want to give a meme gener uh on GBT.

So, uh, I'll keep playing this, but you can see how the first one looked like not great. I mean, it was like the sentry guy wasn't there. Like, it's a dark I mean, I looked at this copy. I was, okay, like clearly it just copied word for word from the for language and like, yeah, I mean, it nobody would get what this meant. So, then I just like uh prompted it again. I was like, "Okay, use this picture of Sentry." Sorry. Uh, Sentry/Void. Uh, Sentry. I don't know if you did you see the movie, the latest Marvel one?

No, but I've seen like the trailer.

Yeah, it's it's actually pretty good. Um, I felt pretty disappointed with Marvel for a long time, but I actually think this was this was a solid movie. Um, I'll do like a very quick 10-second explanation of Sentry. Uh, basically when sentry is really uh sad and in a depressed place it turns into the void. So so devoid of stuff. So that's um the same thing. So I gave it I I basically ripped that photo. Sorry. O I ripped that photo from Google images I think. Um and then I came up with this which like obviously looks kind of this looks like some Statue of Liberty like Superman. Doesn't really look like the void. And then like I don't know what's going on here. Like peak with the whole I was like trying to like tell it to use a New York skyline and it like just took this text. I mean it just does stupid things but like I don't give up. I just like try to like see if I can be more specific and I was like actually use this image, right? And then um change the text. Uh, I think it like crapped out there. And then I said actually like add the peak logo onto the face and then the text should be this. So it's very specific and then you get that. So it's multiple iterations. Uh, like I think if you give up after the first try, you'll be like, "Yeah, it sucks for like image jet or making memes." I'm like, "No, you just got to like, you know, finesse it and you you get something out of it."

Awesome. Um, what else you got?

Yeah. Well, uh, let's go back here. Um, I think we did actually talk about this, but yeah, you can use like GPT and LLM to brainstorm. And in fact, um, not beyond just like meme ideas. Uh you can actually use I I use 03 for this mostly but you can actually feed it your Twitter handle, your socials and ask it can you like analyze like what you think would be three strong angles for me to talk about uniquely and like give me those categories. You could do that. Um, and you you'd be surprised like what it comes up with.

Yeah, you can do it and you probably should.

Yeah, absolutely. Yeah. Um, just have a few more things. I know like uh Yeah, there's a I put this in here because a lot of people like say that build in public has a bad rap because people are just talking like it's like they're just yapping or where's the product, right? And I think that's why the build part is really really important. like you got to show like you're interacting with people, you're taking feedback and all of that. So, I actually added this tweet in here because it's very recent. Um it's just from like a day ago, I think. Uh but I yeah, so this person Seb um had been using Peak for a month. Um I think discovered it from the tweet probably from Preston. Um and uh left feedback, right? It's like, "Hey, here's some things that like I want, you know, working better." The refresh, uh, the, you know, Gen Z lingo is something that some people like. It's some people don't. It's polarizing depending on who you ask. And then, um, this part is really where I kind of took out and said, "Okay, we're doing the one thing that we want to uniquely do well, which is story gen." And then, I mean, I always think when someone says borrow UX, like I I appreciate that. like it's a yeah it's something that I don't mind hearing. Um so I yeah I interact a lot with people on replies, DMs, probably have like at least 50 people like in some kind of direct line of communication that just you know sends me the stuff um while we're building Peak Public and that's it. That's pretty much it.

Yeah. All right. Yeah. Um, Sherry, thank you so much for your time and sharing your process and how you built this both from a product perspective, from an insights perspective, and from a marketing perspective. Um, I'm going to include the the links uh to where people could follow along your journey and Peak's journey uh in the show notes if people want to go and and and and get more of that. Um, do you have anything you want to leave, you know, the audience with before you head out?

Um, I think this is the golden age of building AI native mobile apps. So, um, I'm really excited to see what everybody starts building. Um, this is like a paradigm shift not just for personal finance, but it's for the whole suite of any kind of applications that we're going to use to like, you know, improve ourselves in some area, right? to like live a higher quality of life. So, I think this is like such a cool time to be building. Um, so yeah, just uh go out there and uh do some cool stuff.

Amen. Thanks, Sherry. I appreciate it.

Thank you. Bye.

Bye. Bye.