Designing GenAI for the Emotional Side of Money: Johannes Seemann

We're calling it intention instead of goals because goals feel like a destination that you can fail. Either you know, achieve your goal or you don't, but intention sounds more like a journey. It's what I'm doing, value a line and does it guide me towards where I want to be. Welcome to IDEO's Creative Confidence Podcast. Conversations with today's most creative leaders, change makers and design thinkers. In every episode we explore how to lead with creativity, build resilient teams, and tackle tough challenges with imagination. Hi everyone and welcome to the Creative Confidence Podcast. I'm your host, Mina C. Thuramon. Today's episode is hosted by a special guest, Becca Carroll. She's IDEO's Chief Strategy Officer and a Business Designer who thinks in systems and asks questions that get beneath the surface. She's the perfect person to interview the founders of IDEO's inaugural Startups and Residence Program. IDEO invited two early stage founders to work inside our studio. And today she's sitting down with the first of them. Johannes Seymun is the founder of Sooner, a fintech startup built on a single specific bet. That most financial tools have been designed for a rational actor who doesn't exist. After two years of in-depth research at Wells Fargo, interviewing real people about their relationship with money, Johannes came away with a different premise. Money isn't primarily a math problem. It's an identity problem. And most people are navigating it completely alone. Sooner is his attempt to change that. And this conversation gets into what it actually takes to build with AI and is spaced this emotionally complex, including three design challenges Johannes is currently wrestling with. If you're enjoying the Creative Confidence podcast, please subscribe wherever you listen and stay tuned for the second episode in this series. And now here's IDEO's Becca Carroll in conversation with Johannes Seymun. Welcome to the Creative Confidence podcast. I'm Becca Carroll. I'm the Chief Strategy Officer at IDEO. Today I am sitting in as a guest host. And before I introduce my guest today, I should probably let you know why I'm here and not the regular host Mina. At IDEO, a lot of my work right now is working with organizations who are thinking about how do we build for an AI native world? What does AI native even mean? And that goes beyond what tools should we use and really goes towards what does it mean to do that thoughtfully? What does it look like when we are trying to design for AI nativity, a technological nativity? But we want to be human centered designers. We want to build things that are good for people. So what does that actually look like? And I'm really excited to introduce as a guest today, Johannes Seymun, who is someone who I think has been living that question from the inside and really the frontier. Johannes has really been thinking about what people's relationship is with money for years. And AI is really offering this opportunity to build in a way that I think is really interesting and I hope you all do as well. And so my hope for this conversation is that you leave with a new way of seeing financial tools, but also financial stress. I think sooner and Johannes are going to make a really compelling case that most of what we have built so far in finance has been designed for what you might call the homoeconomicus, the really rational actor who is really good at math. And that the real opportunity that we have ahead of us is designing for how people actually relate to money, which is often not rational, which is often emotional and messy, and often more and more in isolation in a way that feels really lonely. And I think once you hear that framing and how sooner is building, it changes how you think about the whole category. Beyond that, I hope you all take away something about how to integrate AI into human-centered design, a picture of what it actually might take to build a business here right now, and maybe some fuel for whatever you're working on or whatever you're curious about building. So welcome Johannes, I'm so glad you're here. I'm so glad to have the opportunity to speak and at the same time reflect on what I'm building with my team. So let's start at the very beginning, because you are a human-centered designer and you're also an ideal love, which love to see it. I know you have taught interaction design at universities like CCA, you were IDO's first data hire, and then you ended up actually at Wells Fargo doing something that actually seems quite connected to what you're building now. And so tell me a little bit about that chapter. What did you learn when you spent two years in a big bank talking to people about their money? Yeah, that was pretty amazing and I would say even priceless and quite formative. So to provide a bit of a context, what the mission was is to set up a function within their innovation group. And this function was called the teaching design, which was a combination of bringing together business design as well as design research. I worked there as the design director and trained up a team and did baseball projects and created concepts and prototypes. What was important was to do these deep dive interviews, ethnographic interviews, some were in person, some were remote, and they were really long. And over the course of those interviews, we got so much rich understanding of how people relate to money. And pattern emerged and those patterns then eventually informed that there is unmet needs. And this is what also then led me to start building sooner. What were some of the unmet needs that you kept seeing over and over? And can I just say what a gift that you are able to do these long ethnographic interviews with people over time? At a time that it sounds like relationships with money were really salient and present in top of mind for them. Absolutely. And I have to say that it is pretty much top of mind for most of us continuously. And one of those stories that you come across is there is a lot of blind spots that people have. They may have a system in place. It is frequently informed by what worked in the past, so they're just continuing it. And they're missing out on opportunities to really change because they don't really have anyone that they can consult with. That need for a financial advisor, they assume financial advisor is not really for them. They don't have enough money. So in the most critical moments where actually major decisions are being made, they have to do and find the right solution on their own. That is something that was across the board that they thought financial advisor is not for them. I hear in that a deep amount of loneliness, a feeling like this is a journey I have to strike out on on my own. The other thing I wonder if you heard, I've spent a lot of my career in financial services in varying intersections. And one of the things I think is really weird about the industry is it attracts people who really like financial services. And then it can sometimes put blinders on what you create because you start to do research. You start to assume that everyone wants to approach financial services in the way you who work in the industry and know a lot about financial services wants to approach it. And I'm curious if that sort of loneliness and that feeling of heading out on your own journey was something that felt like a little bit of a surprise even internally. It felt like something where actually your co-workers didn't experience that but you were hearing it over and over again from people. The loneliness for being alone with your money basically and having to make all decisions on your own is unbelievable in its expression across the board. I have to say that even as the teams when I ran it, we had a really hard time talking about our own financial situations. But then over time we basically warmed up towards it and we opened up more because you don't necessarily have to mention amounts. But you have to be humble enough to actually share, I don't know that or I made these mistakes. So when we did for example co-creation sessions with people to develop concepts on how the bank could help them in the future, the warm up questions that I kept asking was like what was the best situation and worst situation with money but then also like what is your favorite tool. And people were hanging on each other's lips to really be like oh tell me how did you do that? How did you navigate it? And there was a lot of empathy coming across as well for these darker moments with money and the lighter ones. You know they were really applauding each other in a way and that shows you that there is this unmet need. Do you have a companion in your life? Some have a, you know, a privilege to have a partner that they can discuss money with. But frequently we also did diets with couples which were super interesting. That is interesting. Who is responsible for what part of the money? Sometimes it was only one. There was the pattern of like keeping it separate or divided responsibilities where one did the budgeting and the other one made you know more household day-to-day decisions. Interesting. I imagine one of the things you saw in there and I might be interpolating from your description of loneliness. It's just a lack of confidence. I don't know if I'm doing the right thing or not but I'm trying. I'm curious that that was a pattern you saw. Exactly. And that lack of confidence at times also leads to people not doing anything. And that is really like where also the name sooner comes in by the way because it meant making right decision sooner. Okay. Because compounding doesn't wait. It's one of the taglines that we're using. of what that really says is like the earth you have the confidence to make the right decision. The more it may actually accumulate for you, either you pay off that faster or you keep more money in your pocket for longer and get interest for it, it is also critical for them to not start avoiding those decisions. If things are too complex, we lean towards, like, let's put them off. I'll get to it when I have enough cognitive resources to deal with it. I wonder if there's a point here about a lack of financial confidence, as you call out, is making you stuck from taking action, which to your point of sooner like sooner can solve that, but also the lack of financial confidence, like it can reinforce that feeling of loneliness. That you're not sure if you're doing the right thing, you're not taking action, you then feel even more alone, and you kind of get stuck in this cycle. That is correct, but also people have, then, a really underlying stress. Like, you can't get it out of your head. You can't get it out of your head. It's always present, and that adds to the overwhelm. I think a lot has to do with important money decisions. There are so many dependencies that you have to consider. How does it impact your household, as a whole? Can you afford it? How does this trade off with your saving? Can you afford the vacation that you're planning for? How does it impact your future? Having someone who can help you to have at least a picture of those, what are your different options, the scenarios playing with you through the different scenarios? That is kind of what's missing. What you're describing, that someone who rides alongside you and helps you think through the life you want to lead, and how what you're doing with your money today makes that more possible or less possible. That has historically been the role of financial advisors. Historically, also, that job was really tuned towards the wealthiest people in society. It's getting a little better. You can one off, you can actually get advice, but in principle, the problem or the challenges that even people who have enough to get a financial advisor think may not be worth it. The other part is, which is really complicated with long conversation with one woman about is finding the right one that you trust. This suddenly becomes a relationship building. It's like finding a therapist. Exactly. You may not like the first one, but you're kind of committed, and you also don't know what to expect from someone, from a financial advisor, what can they help you with? Can you ask them this or that? Questions? Not everyone is also extra-right enough to do that easily, to go through the process of interviewing a couple financial advisors, perhaps, emotional, but we're really believing it that sooner, and also, we're getting all the proof points, the relationship is quite intimate, right? It's already intimate with your money, and then you have to talk with someone about your money. So there is a big hurdle as well, and that carries all the way through. It's like, what are the entities that a person starts trusting with their money decisions? And unfortunately, we heard stories over and over again, they turn to the wrong people. And I have this beautiful story, and it just touches you when you talk with the people, because it's like so human on how they're in the end, navigated themselves towards actually making the wrong decision. And there was, I'm gonna change the name now, but it was Rose, and she was a nurse. I think she was about to have her house paid off, and squirreled away a lot of money in savings, and was like, I had a terrible, as a child, a child taught, we were always short on money, and it made me at times really also embarrassed, but also gave me angst, literally angst, to be having enough right, because it was kind of projected from the parents, and I swore myself, I will never have that in my life. Totally makes sense what she does. Yeah, and good for her, for squirreling away enough to pay off her house, that's an impressive feat. There's an impressive feat. At the same time, she did not invest in the stock market at all. And the reason was financial crisis, but in general, it seemed like too much of a wild ride, and she was absolutely risk-averse. So then our interview goes on, and suddenly comes up like, well, there was this person in the gym, he worked in financial services, he became friends, he recommended to me to make an investment, and I lost $25,000, and that investment was in the stock market in the stock. In an individual stock? That's correct. Oh gosh. So you're asking, going back to the trust theme, that was not the right entity to trust, but it's so human to trust that, right? So human, let's transition a little bit. So you're doing these interviews, you're hearing these needs. I imagine you're looking out at what exists, and you're like, this is a bunch of spreadsheet tools, right? This is a bunch of math equation, ladder it up. Maybe it has a nice UX, where it gives you a beautiful graph, but it wasn't solving these needs that you were seeing. So you leave Wells Fargo, and you start to play a little bit with some of the tools that are available, and you have this light bulb moment. Something clicks for you. So walk me through this. Walk me through this moment of realizing I can solve this problem. Yeah, I wouldn't quite describe it that way. Okay. So I basically came away with like a whole boatload of different unmet needs. And a lot of them were just really so pressing. So sooner in the first iteration was the thought about like, how do we help people deal better over there? That there was one thought, but the biggest challenge was like, who do you trust with your decision making? Ultimately, when AI came around, I was like, this is it. Why were you like, this is it? Because they were already early science, end of 2024, that people started using, starting to use chatbots for so much more, to advise in their personal lives. And I was like, wow, there is a trust factor there towards AI. What would it mean if you put this into a safe, offering or product or a service, where you say like, this is a specifically capable GPT to help you with your personal finances. Maybe more trustworthy than the guy at the gym. Yes. With the stock investment recommendation. Yes, there was the goal. No, but these are like the single story that encapsulates the complete challenge around trust. And the other part was like, I'm by the way, ever masters in economics. So I became a true human-centered designer at AITO. But the masters in economics told me a lot about the field. And it is kind of this spreadsheet type of field. The idea of the homo-economicals that a human in principle we're making rational decisions was so wrong for so long. And in Kanemun eventually, you get the Nobel Prize for that as well. Like he moved it into the behavioral economics. And you may have all heard of it. That was, you know, still assuming now that humans are not rational, but irrational in their decision making. Or I hate this because it assumes there's an objective right. I'm like, humans are often very rational. It's just that the numerical conversation doesn't cover all of the criteria that they are considering when making decisions. Exactly. And so I hate the feeling of like, well, you're rational or irrational. I'm like, actually, maybe you are just broader and more encompassing. Yes. Then you can distill down to a single number. Essentially, it's like, hey, we're not human centered. All we are seeing is that people act irrational and not to their best interest when it comes to being good with money. Let's just make a trivial example. You're overshopping. There's a deeper psychological need for that perhaps. You're fulfilling a need or you're maybe just enjoying it. So if this is your lifestyle that you would like to live, maybe you don't have enough money, but it's still true to how you want to live. Basically, what I learned through idea of the human centered design and then at Wells Fargo was this is about financial psychology. This is about coming from the human centered angle. I think about it as the microeconomic aspect when you think about the household or the individual rather than the macroeconomic one as well. This was the basis to turn the equation on its head. So we're not starting with this deductive approach of putting everything into an organized spreadsheet and then competing on data visualization but starting with the person, how they relate to their money, what their relationship is, then from there saying, "Hey, how can we support the decision-making aspect?" And then we help you budget. Ideally, what AI should do anyway, because we hear from so many people that especially also our user group is, "Oh yeah, I should really budget." Or, "I tried it out and I don't know if it's worth it." Because life is just so chaotic at times that how do you fit that in a spreadsheet? To me, it's incredible to think about instead of budget as input into finance, budget as output. It's much more about who you are and who you want to become. And as you're saying, your financial psychology. And then from that, using these incredible tools that are good at inference, that are good at analysis, and saying, "Great, we can help you to get there." Yeah. Okay, so you see in 2024, "Okay, all of a sudden I'm seeing people talk to their chatbots, just general-purpose chatbots about money." And they're having the kinds of conversations that I was seeing in my research. They wanted to be able to have. And then you start to think, maybe I could form a business around this? Or what was the process like? There was a bit of a challenge for me. I got laid off in 2024, so I had extra time on my hand. So I started experimenting at the same time. I did an MIT course on no code, AIML. Because you're not an engineer. So you were learning this in real time. No, but I'm a builder. That means like, you know, whatever. I'd go to a tour and get my hands on, I'll try out. And I think what drove me more was like, I had an idea. And the idea was basically, how do you create, let's just say, personality style test on what your relationship with money is. Because we realize as well in our research, it's like, there's not that much self-awareness. There was a lot of, like, in conversation, people realizing things around what their money patterns are, or behavioral patterns are. So I started trying this out, very rudimentary Python. And eventually, there were tools available where I could actually create a full codebase. And that codebase, you have a conversation with a person about their relationship with money. That transcript then gets run through that. And at CN comes out a profile that is organized in different segments. So this is what I did. And I was really exciting the process when you do that for the first time. And something not really beautiful comes out the other end, but you're like, wow, I was just able to do that over the course of a couple of weeks. Obviously, now tools have become even more powerful to do that. But it was fascinating to me that it could hand off what I'm okay with, like data analysis to the AI. Did you test that with people? That sort of first prototype version? Absolutely. First, I did long forum interviews. Right? I did the classic one hour and then I created these profiles, gave it back to people, be like, is this useful at all or not? And then I got positive signals from there. The other part of it was, I wanted to have this in AI-moderated interview. This was also something in my mind that makes it more natural, similar to a qualitative interview rather than creating a sophisticated questionnaire. Like you may have taken for personality test. So a friend of mine and a quick shout out here. Thank you for the support, Triak. He's, hey Marvin, is a platform that's mostly a research repository for qualitative researches. But in this case, they also had an AI moderation tool. And he allowed me for free to put my interview script into it because it's now basically, now I'm narrowing from a one hour long forum interview and I'm trying to scale it down into getting enough and a precise conversation that provides sufficient data to give you a rich profile. I hope that still makes sense. I'm going quite into detail. But the point being is, we're tested how it feels like for people to talk with an AI about money. And interesting, what came out is you can kind of shorten the time, what takes longer when you have an in-person interview to warm up and for people to feel comfortable, trust, to speak with you about money. With an AI, it was more like, yeah, let me just dive into it. Well, you know, it's a computer. It's not going to judge you, right? Exactly. So you can, you can just be honest. That's the keyword judge. Yeah. So you did this. You ran this test and you got some signal and you said, there's a there there. But what I find so interesting about this is AI didn't create this insight. The way that sooner is AI native is not that you went into chat chpt or clod or wherever and said, what's the business I should build in financial services? You knew this, right? You knew over maybe I should have done that. But you knew sort of you had found and been validating the steep human insight for years actually up to the point that you were ready to run this prototype. So you were really zeroing in over time and with expertise into desirability. And then as you started to think about AI, I imagine that started to unlock things for you around both feasibility like, oh, wow, we can actually deploy this to people, but also viability to the point of financial advisors have historically been relegated to a small segment of society, right? And it's not actually viable to have everybody sit down with the human for multiple hours at scale. And all of a sudden, it's, hey, we have this technology and it could solve this problem. It could be desirable feasible and viable today in a way that it wasn't even five years ago. Absolutely. It's a game changer. The keyword was scalable and I would add 21st, seven. Financial advisors don't respond well to if they get calls and at 12 o'clock at night. I think this was the starting point. And I should preface here as well like the psychographic test that I was just describing, a kin of a personality that is around your relationship with money is really the data engine or the starting point towards creating the actual AI partner for your personal finances. So how this works is like a twofold. There is the, let's just say the on-ramp, right? There's a warm up scene again for the user. It's kind of a warm up towards saying like, hey, this is something useful that came out of this conversation that I had. Oh, I would love to continue that conversation and then it's an onboarding towards the native app. This is one idea or that is part of the product. But the other one is also the underlying information that comes in. It's already a starting point for the AI partner to make smarter recommendations or decisions. Spaisy, you pick up a richer conversation right from the get go, trade also a profile for the user that then can be for any future decision, you know, the AI can fall back on. And that's the last part. And it's already really interesting what behavioral patterns we find. And that's kind of the proprietary aspect of the GPT will be able to tap into these behavioral patterns, then have that as a reference as well. So it's not going to just have a smart set of here, personal finance principle, here are some psychological principles. Let's use those and help the person with smarter decisions. I mean, that at some value already, but it goes beyond that. It's just like, okay, there is unique behavioral patterns. I see those signals that means I need to account for when I help the person work through their options towards making a smarter decision. I wonder if you might give me an example of those unique behavioral patterns and the kinds of decisions that you might make. Because I hear you saying like, it will not be a future in which chat, JBT, your name or tool of choice, you'll just go into it and ask these questions because it's missing as you talk about this deep data profile. But also, I think you're saying like, there are certain things that are somewhat unique that are going to push people towards different kinds of decisions and sooner can help them do that. So I wonder if you might bring that to life a little bit with an example. One thing we found is that identity matters so much that you can chat like on what's the rational decisions or when someone says like, I'm really frugal, for example, they say, I am have the identity to be a frugal person and they will tell anyone who wants to listen. And that's very common that identity really drives behavior. But what it is not is the person doesn't say like, I'm frugal because of this and that. So that is one example. The other one is your identity for example is already lifestyle driven, meaning you may be living a lifestyle that you can't afford because this is how you see yourself. It's as trivial as that. But that's something incredibly important when it comes to decision making. How do you satisfy this idea of what their lifestyle is account for that rather than optimizing just towards the most rational? Here we go again, decision. Yeah, it makes me think about idea we often talk about trying to understand not just what people say and not just what they do, but also what they're thinking and feeling. And that sooner in sort of starting with this sort of psychographic analysis is really cutting to the heart of that matter, pardon the pun, but even sooner. And I think that's incredible because ultimately as you're calling out, that is much more of what drives behavior, what you're thinking and feeling, then necessarily, oh, well, I told someone I was going to do X. So I'm going to do X. I mean, this is also honestly the funnest part for me. Yeah. Yeah, they're discovery out of the information that we're collecting and being like, wow, these are just principles that are so obvious at the same time they haven't been accounted for. And let me give you another example. It's like sometimes people say, like, hey, I'm really not following the trends. And again, they may be telling everyone, you know, they're not really there, maybe end time materialistic or at the point is you will never be able to detach yourself from your social context. You will live that social context and it will cause you, you know, even if you don't fall into general consumption patterns, it's still something to account for. I hear in that, and I know we've talked about this, that the fundamental bet underneath sooner, much more than what the interfaces or how the product is experienced is that people are going to manage their money much more according to who they are and their identity than the information that they have or the quote unquote rational set of steps they should take. Exactly. AI is everywhere. It's exciting, but let's be honest, it can feel overwhelming too. How do you actually use it to solve problems, not just add more complexity? IDO use AI and design thinking certificate is designed to help leaders like you make sense of AI, giving you practical hands-on tools to enhance decision-making, streamline innovation and lead with confidence. Know high, just real skills. Check it out at ideo.in/aiSert. That's ideo.in/aiCeRT. Now. back to the show. So talk to me a little bit about the product experience. I would love to understand for what you're building. If I land on sooner, what can I actually do today? And then in the future, what do you imagine is going to be the end-to-end product experience? So right now, you will be able to get your money type a profile and that goes beyond being a profile. So you go through the, you know, psychographic tests of how I described it already. But that profile goes further. The profile also talks with you about your intentions in life. Besides those intentions, there will be also the early tips and ideas on how you can get started on things. I should also point out the intentions are so important. And we heard that already as feedback, it helps people to clarify what those actually are. Because frequently people have a very loose understanding. They say like, OK, yeah, I want to have financial freedom. And I want to travel. And I want to have security. And I think it was also MIT study, like HLAP, that found out that 80% use 26 terms to describe what their future looks like, where they have more time and money. And helping them become more specific about it, for example. Really think about what is one intention or aspiration that you want to have fulfilled it within one year? Or further out, medium term, long term, is surprisingly difficult for people. So that's already one benefit in this one. And from that, what is being prototyped has to do with the actual AI partner that will be available 21st-Saman for you. And maybe it starts out with this conversation around, like, remember, this is one intention. Do you want to make a plan around it? We're calling it intention instead of goals. Because goals feel like a destination that you can fail. You either achieve your goal or you don't. But intention sounds more like a journey. It's what I'm doing, value a line. And does it guide me towards where I want to be? It's more towards experience than perhaps ownership of something. Yeah. So it's not as binary as a goal, where it's like, I want to take a vacation to a tropical island. It's like, either data or I didn't. It also allows you more to adapt towards. And adapt the plan itself as well. Because that's the other thing is the best-laid plans. As much as you want to plan with money, things can happen that totally alter the course for you, a job loss or a bull run in the market or whatever. In good and bad ways, there can be significant changes. And so if you have binary goals, I imagine, that can be really unfulfilling. Yeah. And it also invites you to fail. I can, for example, imagine you sit like unexpected expenses. So I had a vet bill last year of $2,000. I didn't account for that. So how do I now recalculate my runway? But I bet your intention was to take care of your pet. Even if you didn't account for it in a budget, that you wouldn't say, well, it's not in the budget. I'm not going to do it. Absolutely. So poor dog. I was just thinking about having the conversation with my dog. I'm sorry, dude. I'm sorry. I have a spreadsheet somewhere that says I it's not in the plan. Like I assume this is a great example of where setting an intention might be much broader than a singular financial goal. Yeah. And also making sure that you're actually putting your time and effort against the value and how you want to shape your life going forward. I want to come back to something that you said around, well, yeah, you want to feel confident acting. You want to be able to make a decision. Because I think one very salient topic right now that I think everybody is wrestling with is how much decision making? Do we actually hand over to AI? That AI can do quite a lot, but actually what should it do? Do you probably listen to Sam Altman's interview a few days ago where he was talking through the way young people are using AI as an operating system for their life? And he was saying, many people feel like they can't make a decision without going to the AI to say which one should I do. And that transparently, like that scared me. Because I was like, I don't know if we want to be deferring all of our decision making to what is ultimately a computer system that is built on previous data, trying to infer the next most likely outcome. I don't know if we want that. And so I'm curious a little bit about how do you land on for sooner? There's a version where it's kind of think like Robo advisor to the 10th power, right? It does everything for you. You never make a decision again. And then there's a version where you are signing off on every single thing that it ever does. How do you think about balancing that line of giving people the control that I think they need to have autonomy over their lives and not feel controlled by a machine? While not overwhelming them with the number of decisions that lay ahead of them. So first and that came from financial advisors, quick tour on control. There's actually really type phenotypes here as well. There's the one that want to control very tightly. You could also call it the spread cheaters. But they want to make their own decisions in the end. But they want someone on the side who may be giving them good tips. There's the ones that want to kind of semi-handed off, collaborate on. And then the ones that are the robot voices. They're just like, I have a life to live. Let's just automate everything. For the control aspect, it's something to design for. My early signals tell me that it's so powerful and also intimidating that it's important to help understand the person that they are in charge. So how do you do that? Let me give a very trivially example. We first set up our AI moderator to have a free-flowing conversation with turn-taking, similar to exactly what we're doing right now. And it turned out that the preference was to actually have a push to talk. And that's in the function of control. You're suddenly controlling the conversation. It also besides that, it gives you more time to think. You're suddenly doing it on your own time. This needs to carry through. How do you design for that? That you actually hand control back? I don't foresee the financial decision to be made by the AI on that larger level where it comes to complexities in life. I think there's needs to be options that the person, or that the user can choose from. Because the field is evolving so fast as well, it's an interesting place to watch and see. How much are we handing off eventually in control? And what part does it play in our lives? Talking about Sam Altman and the idea of operating system. It's also ethical question, right? Do we want that? It's not like that we haven't consulted people all the time for even minor questions, right? If you have a partner. What's different now is that you're using the AI to do that. Does it make us less decisive? I don't necessarily think so. The future is really for human-centered design. Like my ideal vision is, you take away this cognitive load around money and decision making with something that has emotional intelligence as well as the functional expertise, which in theory should free you up to have more meaningful experiences in life away from making AI your one of the primary experiences of your life. That goes to something I've been thinking about for a long time, which is maybe part of the reason we are seeing people try to defer decision-making to AI systems as that decisions have historically been really hard to make. Feeling equipped to make those decisions has felt like a very high lift, such that the first sort of human instinct is to be like, maybe I just don't have to make it. But actually, if you kind of take sooner to its conclusion and imagine it across other industries, actually maybe the opportunity for us is not how do I make it that the decision that was really hard to make is now automated, but actually that the decision that was really hard to make is now very easy to make, such that it can be representative of me. You also mentioned sort of you keep being a human-centered designer at the core of what you are building. I observe you to be really treating it as like the competitive advantage that sooner has is that it is a human-centered design product. How does IDEO shape how you think about that? I think IDEO instilled in me the approach that you start with the human first. Think about these three-- we frequently say, these are the three pillars that we have to bring together for innovation, which is like desirability first, what the human needs. You mention already feasibility and then viability. So if you look out there, there is the one approach foundational labs have taken around GPT. And that is like, here's the technology. Let's see what users do with it. And the ones in a while they're like, oh, look what people do with it. It finds application. However, IDEO has a very strong standard. It sets itself apart by saying, we start with a human need. But that requires training, that requires listening to actually be really good at, to distilling out, even in between what people set, we always refer to it as they explicit needs and then implicit needs. We just touched on two, for example, trust, as well as control. So how do you design an offering that has a GPT and has the user interface that delivers that? And there's only one way. And again, it's what I picked up at IDEO as well. You start building. You start prototyping. I mean, this is what I did. And then through those iterations, and they can go very quickly. You could continue as feedback from users. As a founder, it's a bit different. because you hold the vision. And there's a really strong hypothesis. You assume that you just know a tad bit more than you heard from users. A strong opinion loosely helped. Let's just put it that way. Once in a while and you're being corrected. So you need to be open to that. Or maybe your questions get narrow and narrow is maybe how I think about it, which is like you hold at your core this problem you're trying to solve. And you have a hypothesis that like this manifestation of the product solves that problem. Perfectly expressed. But you might find that there's a different manifestation of the solution. And that I actually think that gives you an advantage, a competitive advantage as a founder because you're not stuck on well in building this. And it actually doesn't matter if it solves anybody's problem because this is what I've decided to build. Yeah. You're really focused on this is the problem I'm solving. And what's the fastest path to get a solution into somebody's hands. Yeah. And now you basically have better tools to do that. You can just prototype so much faster. Let me ask you a question because I sometimes talk to not just founders, lots of people who say we are user centered. And they say we have actually a human centered design phase at the beginning of what we do. And then where I always leave is but then what happens after the phase. How do you continuously apply HCD practices in what you're building at sooner? Or is it sort of I did this at the beginning and now I've got it checkbox and done. Yeah. I mean that was partially also why I've went on client side or well as far as like how can you keep this alive over a longer course of time. And I think frequently there's still how do you let go of what you've been invested in already. I think this is the biggest enemy of human centered design that there is a bias and loss aversion around this is what we're good at. This is our product or our resources that we can rely on. An object in motion remains a sin in that. Yes. And certainly you may even realize how valuable this feedback is. But this is why there is the other lens. It's feasibility and I think a lot with the feasibility also has to do with the talent and the tenacity to keep the pure you know insights or needs that you discovered alive and continue prototyping. I cannot repeat that more often every prototype is hypothesis that you put out there in front of people. And so it's a not a one-off deal where you have like okay here's our sets of insights. Let's just build something and invest a lot into it. It can be very nimble. That's probably a muscle that an organization needs to build. And it needs to be also integrated right I should say like putting research and design together. So taking designers on the journey with the user is absolutely quintessential. Or now even you know your ML AI you know expert. Can come alongside I think this is incredible. Absolutely. I actually think one of the biggest exciting things about AI for me is the ways that it allows us to translate and blur between disciplines that had started to feel somewhat siloed from one another because I think you find different things when you bring different lenses into that research. Yeah it's an idea and intersections. Always. It's perfect. I want to get into a little bit with you on you as a founder are really at the edge of what it's actually like to build with AI. Compared to maybe when you were at a big company and you had to wait for you know someone in technology to let the technology settle negotiated enterprise agreement handed off to you and so there was sort of more time. You I imagine are sort of living at the edge of what's possible to build with these tools. And so I want to hear a little bit from you on like what does it actually take to build with AI today still a team. It's not a solo thing. It's not a solo thing even though I have to say these first parts that you spoke about earlier right where you start prototyping and you can really excited. You know at a moment it felt like I'm collaborating with a really intelligent person. This faded away by the way. I'm now more back towards a tool you're helping me and yes you're smart and the conversations are perhaps even eloquent but it's not it doesn't have that feeling anymore that you're really working with people. What I should call out is that the field is moving incredibly fast. So it's probably whatever I'm sharing is going to be outdated. We'll be outdated by the time the atmosphere. Yes I have accepted. I know it's it's humbling and it's also kind of brutal as a founder to have to keep up with that speed but getting back to working with AI and why it has limitations at least for what we're building with sooner. I think about it and we need three architectures in the end. One is obviously our AI architecture right like what is the brain that also has this proprietary information that we're talking about and you know now so user really well you have an conversational layer and then the accountant how we call it who does the actual calculations because GPT is still not 100% reliable there so user machine learning approach. Then the second part of the architecture is actually UI because there is a paradigm shift on when it comes to interaction design and designing for AI. Do you still continue into the future using your screens that much? So we need to be open to that as well. At the same time the UI change is already for me in the way that you may not want to go really deep into your data so can they I really surface that for you and then thinking forward when you think about this all the data visualized session and everything else AI should evolve to that point where it can actually create the interface for you that's is most suitable. I mean that's now further out mission but the third one is then the rack the retrieval augmented generation that's basically the library that makes sooner brain smarter than you know going to a regular GPT with your personal finance questions. All of this needs to be brought together and that is a mix that I don't think you can just you know give to open claw. So right now even the test we're on AWS right so we're scalable that still requires back end. We're using two different GPT's for different purposes which one is better the other moderate at GPT one that we use gets deprecated you have to switch it out for another one you have to test with the voices which one you use so I check in with myself sometimes I'm like making this more complicated that it needs to be you know there's the promise out there AI could do everything and I think for creating a bespoke system and offering that actually captures a very complex human challenge around us and money. I don't think we're at the point where we can have the GPT run and create the solution. Yeah that makes sense I see why it's so seductive to think it's possible. You want it to be you want it to be possible but I also see how like that standard might actually not give as much credit to the superpower that is if it can get us 90% of the way there and then we can focus our efforts on that last 10% yeah and we can focus our efforts on what what needs to be true about that last 10% to give the data that you have meaning to people yeah and that might be about reliability like you're talking about that might be about trustworthiness that might be about interface that's a whole host of questions that you're raising that actually gives and creates I imagine for your team more time to focus on those important questions when you can trust that the sort of first 90% is handled I imagine it is easier to start sooner today than it would have been 10 years ago. Yeah I don't think in its its current iteration it would have been possible doing it way back so I strongly feel it's right time right place for what we're building and and with the capabilities that exist yeah I want to talk a little bit because you are so on the frontier and that some of the specific challenges you all may be running into when it comes to building an AI product for something that is as you're calling out psychologically really really complex and I know when we were chatting you were naming challenges that I just love to hear you go a little bit deeper on things like normative contamination or context raw or false coherence and these sound like maybe buzzwords but they have real implications for how you think about building the products and I wonder if we might just take a beat to talk through some of those. Yeah and that's a space that I'm also proactively looking for collaborators this isn't incredibly complex and I don't know where the frontier labs are with it at some of those points but normative contamination let's start with that yeah can you E5 for me a little bit it's absolutely so it's basically they I hasn't bias and it makes an interpretation and how this can play out during prompt engineering it became very clear so you basically take a transcript and you run it through different prompts and see what comes out in the in the profile so here's the happen the person said in the end I want to have a house in a garden and travel more with my family on the other side came out oh you really modest rather than ask for a penthouse you just want to have a house in a garden to it made it made that assumption I didn't say that no but you know what penthouse means right penthouse is a signifier that everyone pretty much knows is someone who made it yeah and this is basically the data said it was trained up so it uses this norming approach to make value judgments that is something we cannot afford at all and the reason why is now this is the psychographic profile of this person who is now modest and it informs in the future help that they are a partner provides oh this person is modest this is just one example right there may be other things where this value changes and actually leads to worse capabilities to support people in their own decision-making in the context of the life that they want to live. So that's the normative part, and it's fascinating because it's normed against what we perceive as a society. And it needs to be normed towards you. But the user actually stands for-- - You are. - Exactly. - Okay, normative contamination. I get it, it seems like a really hard problem, but also very relevant. What about some of these other ones, like false coherence or context-straw? - Let me start with the example. In a transcript we heard, "Hey, I want to be my own man and make my decisions and someone just kicked off their life in New York." And at the same time, the parents still pay part of the rent. - Back those two are identity points. - That goes to the identity point. It's a very valid intention that I express, as well as it's not the reality. So to make this coherent, it may be just like flattening it towards, he wants to be in control and does his own thing. But it doesn't account for the dependencies towards the parents. - It loses the nuance. - Yeah, it needs to hold both of those. And people have a lot of contradiction. And between those, those tensions are actually the richest part of how they can help. But you say one thing, your intention is this, and the reality is that, "Shifties right now want to make coherence, basically, and we need to keep it the incoherence for longer or account for it and name it for what it is." That's also ties then into the context brought that the longer these conversations go with all of these complexities of a person. And then you have to consult it on having options for making smarter decisions. You know, that context may get lost. It's as simple as that. - But I do think there's something really interesting to me in this idea of the technology is not breaking when it flattens. It's architected in such a way that that is the logical outcome. But when you think about money, which as you're calling out, is both highly unique to the individual in their psychology, but also highly ambiguous, that these questions become even more front of mind. And it both pushes you to sort of sit at the edge of, "Okay, what is actually possible? What might I design?" But it also pushes for all of us a way of thinking about, like, well, how do we design these tools to hold this form of ambiguity? And acknowledge that humans are really messy and they're really complicated. And they have a lot of context. And they don't want to be flattened into an average. And they want things that will design towards their uniqueness, rather than compress it into the most obvious next answer. And I think that's a fascinating challenge to solve it. Feel really lucky to talk to you 'cause you're sitting at the edge of it. - But you were also so true in saying, like, "You want to be seen and hurt," right? As an individual. And that's so important. We're doing early tests of our data library that we're using the REC. And the question that we're putting in that my collaborator developed, ATEE is, we're taking a combination of like a very emotional aspect and a functional aspect. Which also is part of our overall design that we call introspective finance. Which is like the introspection part is the psychological aspect. And the financial part is obviously, it's the functional aspect of a decision. So she came up with ATEE of using this combination for the early tests. Now one simple one is, I have to buy a new car and it really makes me anxious. So now that GPT and we're testing it against the general models that are out there needs to figure out what is this contextual challenge now? Do you address the anxiety aspect first or what's the car buying? Discipline around what should you do, what helps with that and then bring the two together? So for example, if you have anxiety, you lean towards either avoiding decisions or making them very impulsively. So that's something that they have to account for. And call it out and provide the guidance around like how can we make this a process of car buying that's more suitable to you? And go as far as like bring someone along for the ride, check in with me, upload whatever documents you get, leasing documents. And again, it can start out as much as like, "Hey, do you want me to calculate what your average monthly cost was for a year beyond your car payments?" And then we can figure out what you can afford. All of these decision points, basically making more confident that you're equipped and reducing side. That's awesome. I hear in that really a shift from, I'm not confident, I'm avoiding, I'm putting my head in the sand because it feels overwhelming. And as you called out, feeling that constant anxiety or angst day to day, to I feel like I have agency over my life. I think that's extraordinary. Talk to me about if sooner, if you all succeed, and I hope you do to your highest ability, what does that look like? What comes next for you? What comes next if you all succeed? I think my hope is that it creates a safe space for people. And that it addresses this, I'm not alone anymore with my finances. And that it keeps more money in their pocket for longer or pays off that faster. Just this little bit, right? This whole idea of sooner, being that free, just, you know, let's just say a couple of months earlier. It's a couple of months earlier, right? And then also cognitive space. Freeing up to engage user gain time for other things. I think it's also something I would like, by the way. It is a bit of an aspect of like designing for yourself there. I have to admit. Well, you know what? Money is a universal technology. And therefore, probably a universal need. I want to end this section with something you've said when we've chatted, which I can't stop thinking about, which is that the need has been obvious for years. But we didn't have the tools required for us to meet that need and show up for people with a solution. And I think about this all the time when it comes to AI. That the scale of problems we have solved is like this part of the bell curve. But the tail is so long and the amount of value we can create for people is so extraordinary. And I hope that any of y'all listening to this, you are thinking about this because I think there are so many human insights that we haven't acted on yet. So many problems that we haven't yet solved because the tools weren't ready. And all of a sudden, the tools might be much closer to reality or much closer to ready. And while I still believe the insight comes first, I think we have this extraordinary opportunity to say, hey, we've been focused over here at just this side of the bell curve in these needs we thought we could solve. And all of a sudden, we can go after so many more of them in a way that it's actually transformative and wonderful for humans. That's my non-duemer view of how we're going to use this technology to serve humans. And I think you were a great example of that coming through. Yeah, thank you. I think it's empowering. And I think people should think about these largest systemic problems, be like, how does it fit in? How does it make things easier for me? And also experiment for themselves, right? Maybe create that calendar supporting prototype using AI. Start prototype. Yeah. Where you don't have to look at the interface anymore. But it just tells you what to do or not to do. What's more, less important. Yeah, would be great. Maybe I'm just writing out, designing something for myself. There I should try. There you go. All right. We get to transition to my favorite part of the pod, where we do lightning round. So I've asked you a lot of questions about your job and your startup and what you're building. And those are all, as they say, rational questions that one would ask you. But now I get to ask some questions about Johannes. Everybody gets the same questions. Are you ready? Yep. OK. What is a book, a TV show, movie, piece of media that you would recommend to listeners? I am really into gardening and plants and in general. So there are YouTube channel called Mossy Earth. And it's just one of those really figured once with donations that grew basically across different places in the world to help restore coral reef or invest in reforsting in Iceland, which I didn't know, used to have more forest. And the nice thing about it is they follow along and show what impact it had. And now they have already a couple of years. So it's really cool when they revisit sites. And you see the successes. I really enjoy that. That's awesome. I'm going to check it out. OK. Who is a creative thinker that you are up to? I have to say there was a bit of a creative awakening when I joined IDO. And there's one person, in particular, that was very influential. That's my close friend, Miguel, who was a designer? Rad. And because I came, and I was basically, I'm an economist. Let's see how this is going to work out as a data. What am I doing at this weird creative agency? And it was beautiful because people were like, oh, no, no, you're a designer now. And I'm like, I always wanted to be a designer now. It was kind of an informal school I went to. I felt like, right? There were so many people at ID, honestly. It's hard to call everyone out. But that was the most formative I have to say. I feel it very much to myself, but I'm biased. OK. What is a skill you're currently working on? And what is a skill you're most proud of? A skill I'm working on, I couldn't help, but say, keep up on the say I race. But the personal and the professionals, I think I'll take. Yeah, basically getting my old newsfeed, right, organizing my newsfeed about things. trying to sort things out that are really relevant for sooner versus not because the field is also very broad, right? That aside, dog training actually. Ooh, okay. Yeah. I think I have some skill already. Okay. So now sometimes I help people out. Okay. I hear there's both art and science too. Yes, for sure. For sure. Science, yeah, again, behavioral. Okay. I guess. Yeah. The other one, I think this unique skill I have, like I can hold really complex systems in my head from start to finish. I'm very confident about that. I can attest that that is true. And I have senior demonstrate that skill. Thank you. What about a mistake that you've learned from? Oh, over and over. I think team, I have a tendency to be very trusting, give a lot of space. And I think it frequently is not the best way of committing people that creates rupture. But the other one is like money mistakes. I have that, well, you know, I have this high risk attitude. Come from Germany and inheriting money, money aspects from your family is a big deal. Inherited operating system was safe a lot. But I have this other side where I'm blowing all of my saving in yet another startup. So here we are. Well, it takes a level of risk tolerance to be a founder. So I'm not surprised to place that in your money life too. Okay, last couple, what is a philosophy that you live by? Everything can have symmetry and elegance. That is basically the design mission, right? You walk through the world and you're like, Oh, you know, this could be more beautiful. This could be more in line. And yeah, don't give up as a designer, right? Keep at it. I think you're getting to the next question, which is what is the advice that you would give to your younger self? Besides compounding is actually true. I would say, I think it's okay not knowing exactly what you want to be. I think that's the most important one. And I think right now life really confirms that as well. Trust yourself in these kills and the interests that you have. Yeah, take it as a journey. I think is at least what was very fulfilling for me and enjoyable. Amazing. Johannes, this has been an incredible conversation. Before I let you go, I have to ask, where can people find sooner, where can people try the psychographic test? Where can people follow you? So yeah, sooner, we are like at my sooner.com. This is where you can take sooner money type test. If you want to sign up and follow us into the future, perhaps be a beta tester. Go to hello sooner.com. And then you can also follow us on Instagram at hello sooner. Well, thank you, Johannes, for joining us for an incredible conversation. Thank you, Klaikov, for having me. Thanks so much for joining us. If you're enjoying the Creative Confidence podcast, be sure to follow us on Apple Podcasts, Spotify, or wherever you listen. The best way you can support our show is by sharing it with others. So consider dropping a link in your team's Latin channel or sending it to a friend who could use a little creative inspiration. We're so grateful for your support. The Creative Confidence podcast is brought to you by IDEO, the global design and innovation firm, and the team at IDEOU, the learning arm of IDEO. IDEOU's online courses give leaders the skills, mindsets, and tools to create meaningful change. Our podcast team includes Rachel Youngblade, Elisipagano, Aaron Bogar, Ashley Dema, Joanna Ogdenbergs, and Naya Harbin. To get invitations to live podcast recordings and hear about special offers, sign up for emails at IDEOU.com. Thanks for listening and keep fueling your Creative Confidence.