Navigating Workplace Challenges with AI: The Tough Day Solution

30 Apr 2024 · 54 min

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Podcast Notes: Talking AI - Navigating Workplace Challenges with AI: The Tough Day Solution

Episode Overview In this episode of the Talking AI podcast, host Matt Paige engages with Katherine von Jan (KVJ), co-founder of Tough Day, an AI platform designed to help workers navigate everyday workplace challenges. The conversation centers around the transformative potential of generative AI in reshaping workplace dynamics and the innovative strategies behind Tough Day's development.

Key Highlights

Introduction to Tough Day

  • Purpose of Tough Day: A platform aimed at providing clarity and support for workers facing complex challenges in the modern workplace.
  • Founding Genesis: KVJ shares her journey from observing workplace issues to creating a solution that addresses employee needs, especially in a rapidly changing environment.

KVJ's Background

  • Former Chief Strategy Officer at Salesforce Innovations.
  • Strong advocate for worker rights and workplace equity, highlighted by her experiences during high school.
  • Draws from her extensive career in tech during the internet boom, mobile advancements, and early AI explorations.

The Role of AI in Modern Workplaces

  • Generative AI's Impact: Discusses how AI is redefining business models, team structures, and employee support mechanisms.
  • Human Connection: Emphasizes the need for AI to complement human interactions, providing objective advice rather than just emotional support.

Development of Tough Day

  • Initial Idea: The concept arose from KVJ's recognition of the struggles workers face, especially in today's hybrid work environment.
  • Early Strategies:
  • Utilized rigorous testing and validation of the idea through discussions with industry peers.
  • Engaged in "Wizard of Oz" testing to simulate AI interactions, which helped refine user experience and expectations.

Future of Work with AI

  • AI Archetypes: Tough Day is developing AI personas to better understand different user needs and challenges.
  • Employee Assistance: The platform aims to provide actionable insights rather than just emotional support, focusing on problem resolution.

Challenges and Opportunities

  • Market Dynamics: The discussion touches on the increasing complexity of building a unique product in a saturated market where many solutions are simply wrappers around existing AI technologies.
  • Building a Moat: KVJ highlights the importance of content partnerships and offering unique value to differentiate Tough Day in the marketplace.

Insights into AI Technology

  • Research Methodologies: KVJ explains innovative research techniques being employed to enhance AI training, including the use of synthetic humans to simulate worker personas.
  • Ethical Considerations: The team is committed to ethical AI development and ensuring that user interactions are meaningful and respectful.

The Bigger Picture

  • Tidal Wave of Change: KVJ compares the current generative AI landscape to past technological shifts, suggesting that the impact of AI will be profound and far-reaching in both professional and personal contexts.
  • Community Involvement: Encourages a diverse range of individuals to engage with AI developments, emphasizing the need for inclusive innovation.

Takeaways

  • Importance of Human Insight: Workers need a reliable source of objective guidance to navigate workplace complexities, which Tough Day aims to provide.
  • AI's Dual Role: Generative AI can serve as both a tool for efficiency and a bridge for human connection, enhancing workplace dynamics.
  • Strategic Partnerships: Building relationships with content creators and industry experts is crucial for developing a sustainable competitive advantage.

Key Links

  • [Tough Day](https://tough.day/)
  • [Connect with Katherine on LinkedIn](https://www.linkedin.com/in/kvonjan/)
  • [AI Opportunity Finder](https://hatchworks.com/ai-opportunity-finder/)

Conclusion This episode of Talking AI showcases the vision and innovative strategies behind Tough Day, highlighting how generative AI can reshape workplaces by providing clarity and support to employees facing challenges. KVJ's insights offer valuable lessons for anyone interested in the future of work and the role of AI in enhancing human experiences.

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00Season 3 of the Built Right Podcast is right around the corner, but we've got one big change coming your way. The Built Right Podcast is now the Talking AI Podcast, and we've got a lot to talk about in AI. In the Talking AI Podcast, we'll be having in-depth conversations with both AI experts and early adopters of AI. That way you can understand how the technology works and how early adopters are beginning to implement and, more importantly, get value from AI. Our guests range from AI research scientists to founders of AI products to industry leaders putting AI to work in their business. While you're waiting for season three, go ahead and subscribe on your favorite podcast platform so you don't miss an episode.

0:41And make sure to leave us a comment about the AI topics that you want to hear about. So get ready to talk some AI in the new Talking AI Podcast, coming your way August 6th.

0:53Welcome to Built Right, a podcast by Hatchworks where we help you learn how to build the right digital product the right way. In this season, we're going all in on generative AI with guests ranging from international AI speakers, founders of Gen AI products, experts in specific domains of Gen AI, and leaders across industries. We're here to help you figure out how to take advantage of this new emerging technology so you can win in the market. So whether you're an AI techie or just AI curious, we got you covered. Let's get into it. Welcome, Bill Wright listeners. If you're interested in hearing an AI founder story in the making from the previous chief strategy officer of Salesforce Innovations, this episode's for you.

1:35Catherine Von Jan, affectionately known as KVJ, is the founder of Tough Day. And we're going to get into more of what that is in a second. But she's an experienced innovator, thought leader on the future of work using technology, which obviously AI is going to be a big part of that. But she's looking to solve some of our biggest challenges at work today. And before Tough Day, KBJ led Salesforce Innovations as Chief Strategy Officer. She's an advocate for workers, teams, and trust, as well as a recognized leader in technology. And she's been shared by CNBC, Forbes, Fortune, MSNBC, Wall Street Journal, and Wired Magazine, which is awesome.

2:16And I'm just so excited for this discussion today. KBJ, welcome. Thank you. Yeah, so I'm really excited for this. But today, we're going to get into your story of creating an AI startup. And I call it an AI startup, but really, that's just like the enabling technology for what I believe is something very human that you have a vision of building. And then we're going to dig into how human connection and relationships with AI is evolving. Plus, you're a strategy person. I'm a strategy person, so I'm sure we'll get into some of the fun strategic stuff around starting a company and what you think about and how that's changing with Gen.AI.

2:55But first, I want to start with that founder story. And typically when we have a founder on, they're talking about their journey that already happened. What's so cool about this is you're in the midst of this thing right now. So we're getting it hot off the presses right as you're thinking about it. But start with like the why behind Tough Day, the vision there of what you're looking to build. Oh, gosh. Well, first, my story goes way back before building Tough Day and even before I started working. So I think I was always kind of destined to work on this issue. And then I could share more about why and how.

3:34But really what we're trying to do is support workers through their difficult challenges, navigating complexity and perplexing situations. We want to help give workers the clarity that they need, the expert guidance and objective guidance that they need to succeed and thrive in the workplace. And so much of that, it's evolving today, not to mention, you know, we've all kind of, a lot of us have moved to remote or some kind of mix of remote. And then, oh, by the way, you inject everything that's happened with Gen AI since 2022. So it's changing things like dynamically. Plus just, you know, being an employee nowadays anyways, it's just tough.

4:23There's so much stuff to weed through, right? it's hard i think i mean listen there are societal issues there's like yeah there's so much going on for workers today there's so much frustration and fear and grievance and confusion um i don't know that that that's all about the latest sort of exogenous forces and their impact i think that a lot of that has been um you know creeping up on us for many decades and i would say like my first experience kind of having that record screeching moment of like, hey, wait a minute, what is going on here was way back in high school, actually, when I guess up until that point, I had never felt like I wasn't accepted or capable or a leader in the classroom or amongst my peers valued.

5:17And I was leading, I mean, this is kind of an aside, but I was leading this Latin club as president. So my friend Julie and I, we got to high school and we just loved Latin. It's probably a little dorky, but we did. And we were like, we're going to make Latin cool. And we really did. We poured our hearts and passion into reviving this Latin club. And at some point we had hundreds of members. And there was a big induction that was about to happen. and we were inducting all the new members. And my Latin teacher told me that I could not perform the ceremony because I was a girl. Oh, you're kidding.

5:58No, and because originally, Julius Caesar would have done it and he had control and power. And so therefore, we need to find a man in the group, a young man in high school to lead this ceremony. and I was just like wait what this is crazy but I was like I don't know I mean that's just weird and I went home and I told my mom and my mom as a school teacher and fierce advocate for women like it's like oh hell no like that's not how this works and there's something called title nine and like let's go in and talk to the school about it and um and sure enough I got to be Julius Caesar and lead the ceremony and all of that.

6:44But I think even, you know, so there are things that people just do and they think it makes sense to them and they're in a powerful position to make decisions that, you know, unchallenged, that just frustrates people. It could have like totally diminished my educational experience, but I had an advocate. I had someone with knowledge and wisdom to say, hey, let's go have a conversation. quick break in the pod if you're listening to this podcast chances are you've been thinking about how to actually use ai inside your business and that's exactly why we built the ai opportunity finder it's a free tool that helps you uncover high impact tailored ai use cases based on your business your goals your pain points and your industry no fluff no generic use cases just real ideas that fit your business and the rank by roi potential it takes about three minutes to run and it's like having your own personal AI strategist for free.

7:42If you want to try it for free, check out the link in the show notes or go to hatchworks.com backslash AI dash opportunity dash finder. Let's try to let's let's talk this out. And we resolved it. And I had a great relationship with my Latin teacher after that. And it all worked out. But that advocacy and that clarity and that understanding of like, hey, this isn't quite right. And we can fix it. Like that's really what we're trying to do for workers every day. Everyone should have a place to go to get that kind of clarity and advocacy. I mean, not that we're not coming in and helping. We're not going to send in my mom or anyone else into the organization.

8:22But the reality is like workers just need the information and guidance about what to do. And then they can go back to their organization and have conversations that might be hard for them. And that's why we're here. Yeah, we need to replicate your mom as an AI to go in the system. But I always find that interesting when I talk to a leader or founder, there's always those little kernels from their life that manifest later on and drive this for you. I mean, most students probably just would have said, okay, and just kind of kept that in and not known where to go. Like, thankfully, you did have your mom you could go to that kind of led you through that, right?

9:05Right. Really interesting. Yeah, good. I would say like for most workers, if you kind of flash forward and there's a lot in between, but if you flash forward, there's there are a lot of workers when something happens, to your point, they don't know where to turn. They don't have a family member who knows anything about management or coaching or HR or employment law. By the way, like most things are not illegal. like like a lot of people might think like oh they're not allowed to do that yeah actually like a lot of things are allowed you're just kind of upset about the situation so let's understand what's really going on and how we can fix it and and if you only have inexperienced and like lovely human friends that you can turn to they're mostly going to support you that's what humans do for each other in terms of like, oh, I'm so sorry that's happening.

10:05And yeah, he or she must be horrible. And let me give you a hug and sort of huddle you and like talk it out with you. And that kind of perpetuates the problem where what workers really need is more clarity, more insight and kind of, I mean, I would call it tough love, like objective advice, like in order to solve the problem, not to recreationally vent about it and kind of have that cathartic moment of like release. It's more, it's more meaningful in that, like, hey, let's talk about how you can go back and have a great conversation and resolve the situation before it turns into something massively frustrating and brings your worker, coworkers down, distracts you from performance and all, you know, it becomes a mess.

10:54Yeah. It's, I think on the manager side too, like I know I've run into this where I don't always know where to point somebody either. Right. And, and having that kind of neutral third party in a sense, it will get into how this is actually the strategy behind how it works in a bit, but there's value in that to where it is somebody that is kind of, you know, outside of the, um, the nuance of the situation. right? Well, yeah, I think just the point about managers, like managers are employees too, and managers need support too. And often they can't find the information or don't have the source of truth handy and they can be ill-informed.

11:36So this is just as much for managers as it is for employees. So what was this inflection point for you? You're at Salesforce, this awesome position and role, what's that inflection point where, okay, I'm going to take this kernel of an idea, purpose, and vision that I have and turn it into something. What was that trigger for you? Was it inflection point of things going on in the AI space and now this is possible or was it just the right time and just kind of organically went from there? Yeah, I think there's a couple of things. So one, for better or worse, I've been around long enough in the tech industry, I've seen some things.

12:16So like, you know, and what I mean by that is like at the beginning of the internet boom, you know, I was at Lotus, you know, before we were acquired by IBM, but like sort of at the forefront of, of using the internet to do business. And, and then in mobile was kind of out in front pioneering new mobile solutions when that happened and then moving into social and peer-to-peer and all these different technological waves. When AI sort of started to enter the scene 10 years ago is when I got really curious about it. Wow, this can really enable so many human experiences, so much wisdom, knowledge sharing, expertise.

13:02We can really support people. The second thing was then using this next wave of technology around AI to go advance those things. So, you know, I left and started playing and putting little teams together and building little prototypes. And what happened in that process was I was getting tons of phone calls last summer from friends across the industry, people I had worked with, people I had mentored, people who mentored me and let, you know, people were distressed and struggling with layoffs, struggling with like, how is AI going to change things and worried about things. And I found myself kind of enjoying this process of talking people through like how they might think about it.

13:50And then I kind of realized like, duh, like this should be AI. It's not about me and my advice. Actually, it's about the wisdom of the best managers on the planet, the best mentors on the planet, the best HR professionals, legal minds, all of that wisdom brought to bear at scale to help workers in a safe way, in a safe place, to empower them to go solve their own problems. That was the aha. And then I can tell you about what happened after that. Yeah. So it's interesting too. You mentioned doing stuff 2016 and before, and a lot of people think this is like this new thing that just exploded on the scene, but it's been around for a while.

14:38I think 2017 was that big inflection point where the transformer architecture that really just accelerated this stuff took off, but it's, it's been around for a while. But I think the interesting point we chatted uh previous leading up to this you mentioned how you were doing several different things had some teams building pocs but then this like moment of clarity to create focus and how important that was in kind of getting things moving yeah i mean one of the things that happened as well is a friend turned me on to the book essentialism i don't know if you've heard it or if we talked yeah i've checked it out since we talked yeah Yeah.

15:19I mean, it was just so easy. Like, you know, probably there are other books written about this. It's just it was like there was a moment in this book that basically was about like you've got you've got your glass of water like you you can pour that into several glasses or, you know, like sort of how are you going to spend your energy on lots of little things or on one big thing? and then I was like yes like I have to do the thing that I care about the most that the world definitely needs and like I knew I had to go explore it and I knew I knew I had the passion for it and another friend said something to me like okay you were at Salesforce almost 10 years like is this problem something you want to spend the next 10 years thinking about it was like yeah So hell yes, absolutely.

16:11We have to make progress. We've been talking long enough. We have to make progress on this. so yeah and yeah i think what's cool too is it's the beauty of using this new technology and ai for solving big problems for doing good things because you know the way i think about it uh there's the debate of is it good or bad and i think it's nuanced like it's both right there's going to be good actors and there's going to be bad actors and with this big of a disruption it just means the scale is going to be that big on both sides of it so there's a lot of value and you know, people like you actually doing something valuable, important with the technology.

16:50But so you're, you're a strategy person. Take me through the beginning days. I love hearing how, uh, founders think about things strategically, uh, in terms of, you know, you've got it nailed in terms of, okay, I have my purpose. I got the problem I want to solve, but what are other pieces? And some of this is probably like, uh, happening right now in terms of how you're thinking about it strategically um go to market uh you know what customers you're talking to how you're refining the the early thing that will be the product take me through some of that yeah i mean uh one thing is just and i think all founders kind of have this is like um there's there's a point when you realize all your experience kind of was leading you here and you didn't know it all the way you know and you have some secrets um some wisdom or experience that is like hey the world should know this and I'm not going to write a book I don't want to write a book I'll want to build something and make an impact so um so I think that was part one part two or you know that that was important and necessary that then then the strategic part of like, well, how are we going to do it?

18:10Like, we need to start somewhere. So I wanted to first validate the idea. Like I talked to, like, I really talked to a lot of friends who knew me very well. I talked to people in the industry, working in tech, people who were CMOs and CROs and CHROs and VCs and lawyers. And, you know, it was kind of like circling lightly, like, here's what I'm thinking about. Let's catch up, have coffee. I'm just, I have something boiling here. And they're like, I just want to talk about it. And people were like, oh my gosh, you have to do this. So like that, that like, partly I was motivated. Partly that was also like a lot of validation just anecdotally.

19:03And then I was like, okay, I'm getting serious now. I'm going to need a team and I need to think through this with some folks. So I asked a friend who's an executive at Google if he would sponsor a dinner. I was like, I think we need to bring some legal minds, HR minds, great management minds, coaching minds, business minds together and I just want to host a dinner and get them talking about a couple of the questions I have and he said and it's such a gift he's you can't get anything done in a dinner what are you talking about and I was like well I just need I need a way to like bring people together and he was like how about I fund an off site like do you think you could get eight to ten people to go spend five days in the Caribbean somewhere and we just, you know, lock away and hack this.

20:05And I'm like, oh, like, I don't know. That sounds much better. Like, it sounds awesome. Thank you very much. Like, it's an incredible opportunity. And it turns out both old friends and new ones, people I had met like a month before were saying, I love this so much and I believe in you and I I want to come down and be part of this offsite. And that's very humbling and special. And it was just this enormous sense of love and appreciation for these people. And we could have walked away from it saying, like, this is a terrible idea and we should not do this. But, you know, with with three weeks notice, 10 people flew to the Dominican Republic and had five days kind of fun of sharing of dinners of, you know, just being together and, you know, one on one walks and meditation and yoga and and work.

21:06And essentially, we hacked the business plan, which I had been ruminating on. And I asked each person to do kind of a mini section on their area of expertise. So as an example, Jenny Fielding from Everywhere Ventures was there, and she did a session on how to think like a VC. and it was amazing and not everyone in the room had the same kind of expertise that other people had so everyone was learning and then we would do sort of workshops around each piece of expertise so in that case we hacked our investor deck um that's awesome yeah strategically it was not it was like bringing in the right partners i guess thought partners to help think through where we go and how we get it done.

21:53I have so many questions about the, uh, the workshop that would have been awesome to be a fly on the wall there, but like for, to take a step back for all the listeners right now, look what KVJ did, right? She went and talked to customers first and you always hear this, but so many folks go straight to, to building the thing. Uh, it's such a foundational piece, actually talking to folks, hearing their perspectives and the right folks, right the target folks you're kind of going after but you mentioned you were talking to some friends how do you avoid like the friend bias when you're pitching them a new idea and you know they want to tell you oh it's awesome that's great how do you avoid that I guess talking to folks outside of your friend group as well but I think I mean first of all I value candor and and I think all my friends kind of know like I'm going to tell them what I really think and I also know my friends tell me what they really think like I've had ideas where people are like you're crazy like that's never going to work.

22:51You know, so I think just seeking diverse opinions and having diverse friends give you a point of view on like what they're thinking. Like there was no way I could circulate with that, like, let's say 20, 30 people and have them come back. So like in circulating with all those different friends, like it could have been that one or two of them might have been placating me or coddling me or like that's sweet KBJ but I feel like they um fortunately I have friends that will tell me the truth those are good friends uh so you know you're starting this startup I'm curious with folks now with everything going on with Gen.ai and you have Sam Altman talking about oh there's going to be one person one billion dollar companies are you approaching anything differently as you're thinking about your team structure, how you're using technology in the building of it that may be different now than how you would have done it, say, five years ago?

23:57Is there anything different in your approach early on? Well, I had a startup before Salesforce, so it's an interesting thing to reflect on. I think it was doing things back then that actually now are acceptable and back then were not. As an example, using contractors all over the world. It used to be that VCs would say, what, you don't have a core team sitting together, living and working together, and now that's not a question anymore. In fact, they're like, oh, yeah, like you should just get the best talent wherever they are. So and and we can't afford to hire people full time. But fortunately, a lot of people are passionate about it and they want to work on it even as contractors.

24:47So that's one piece. I think, um, AI first mindset. So, you know, um, AI is not perfect, but you can kind of think of it as the first pancake in almost anything. Like what could I do? And it might be messy and it might not work exactly right, but you could probably expedite the process a little bit by, by thinking about using AI first. The first pancake. I like that. And I've not heard that analogy that's good yeah yeah so i think like using ai every day um well i mean we can get into like the research techniques that we're using because i think some of those are really new yeah i do want to get into that uh but just quick aside you just gave like the perfect pitch for for hatchworks right there so like a big part of our operations is down in latin america and it's so true like there is talent everywhere what we found interesting is like post-pandemic people were remote.

25:47The real important thing was, can you share the day's work? Can you be in the same time zone? But I just love how that mindset is progressing. It doesn't just have to be somebody co-located in the US. I mean, there's so much good talent everywhere. But yeah, let's get into that. You talked about some of the research techniques and how you're approaching this that may have been different, uh, free everything going on with generative AI. Yeah. So, I mean, and, and I should also say like, I am a co-founder. I have an amazing CTO who was at that Dominican Republic, uh, offsite and then, you know, fell in love with it too, and decided to join me on this journey, as well as, um, another one of our colleagues who joined.

26:34So there are three of us, three of us, as a core team. And I would say we're all inventing the research methodologies all the time. And kind of, you know, each of us has our own experience to inform that. I think, you know, we've all had some product management, some sales, some, some, you know, customer facing roles, strategy revolves, all of those things. So we're all kind of ambidextrous with each one of us having sort of a core, you know, like Alberto is deep tech, right? And I cannot code, at least not anymore. But so I think in terms of interesting research methodologies, I mean, first I should say, lots of the old methodologies still work.

27:28Like our employee need survey has been illuminating, And I could get into that. But, you know, so so don't overwhelm yourself with designing the biggest kind of research project. Like some of the old tools are still really, really good. So so we use those as well. I think the Wizard of Oz testing that we did as a methodology was phenomenally insightful. and just to explain what that is. And I think probably in different organizations do this and call it something else. But what we mean by Wizard of Oz testing is that users come in to engage with AI, but rather than having AI answer the questions, we've got our wizards behind the curtain answering as if they were the AI.

28:23So we would schedule user testers to come in. We used just a Google Meet. And we said, look, we're going to turn off the video and turn off the voice. And you're only going to be texting with the AI. And we told them it's not really an AI, but we're simulating that experience. And we were actually removing PII and using AI to see what would what would what would it say. But we also had a lawyer, a manager, a coach, an HR professional sitting around the table. So when a question would come in, kind of depending on the first five lines of the exchange, we would decide who's going to be the protagonist and play the AI and like actually be the one to type.

29:14And then the others are all going to discuss and give ideas like so we can move fast. So we were one of us was was representing the actual AI and saying this is what this is what quad pi, whatever any other AI would say, chat GPT. We used all of them about this topic. And then we also you then we also had just the humans answering. And the exchanges were amazing. I think a couple things that surprised us. One was everyone, 100 % of people were very quickly engaged sharing problems and sharing a lot of multidimensional problems. Like they kind of start simple and then get into a lot of complexity.

30:09They seem to do that much faster than you would do it with a human. Like if we're sitting here talking, there's going to be some time for us to get to know each other, for me to trust you and, you know, to know I can share things without feeling judged or feeling any shame about what I'm going to say. And I think a lot of that gets removed when a human is talking or chatting with AI. It's just like very matter of fact. It's interesting. Yeah. So it was like, and then we thought the sessions would last about 15 minutes. Like come in, ask your question or two and in you're out. And 100 % of the people went at least 30 minutes.

30:49Wow. That was astounding. And we had to like cut people off because we're like, we have another one scheduled. Um, and people were asking if they could schedule another user testing session so they could come back in. And so that was, they enjoyed it that much. Yeah, they enjoyed it. But even being unprompted, um, just spontaneously, they would say things in the, in the chat about, you know, this is really helpful. I'm going to try that. Like there was like 100 % of people said that they got value out of this and something that they can go try. So that was also great, great to understand. And then, of course, again, removing all PII from those transcripts, we can use those exchanges to fine tune our model.

31:42So the real human dialogues really help to train the AI to be curious, to understand. I mean, we're training. We know that you can't necessarily give someone the right answer without having context. So, you know, similar to if you went to a lawyer or coach or manager and you said, I have a problem, they'll probably ask you three to five questions about what's going on before they give you advice. so the same has got to be true of an ai that's going to support you in in these challenges yeah it's such a smart approach and the concept of wizard of oz uh testing it's been around for a while but there's the new element here just from hearing you talk about it you get to see how people are interacting with the simulated ai which gives you an idea of the experience how they would interact with it naturally with a human.

Read the full transcript

32:42But the genius part is you're using that to train the AI, train the LLM versus just saying, okay, we got off the shelf, chat GBT, go have fun. And that's genius. And it really is like, especially because this is all kind of new and emerging. So if you're starting an AI kind of focused company, like such a great approach there for listeners when they're thinking about it. And you mentioned you tried BOD and, you know, ChatGPT and some different ones. What was that process like of, like, deciding on an LLM? Did you look at open source options? What was that like? And I know you're not the technical founder like you mentioned.

33:25No, I think that there was, gosh, I'm even thinking back to, like, December. Like, everything changes so fast. But I remember when we first, you know, we're playing with Pi and it just felt so much friendlier and empathetic. And we were just literally asking questions of all the platforms to see like what resonates with us? What kind of brand experience do we want to have or user experience do we want to have? And, you know, we can learn from how they're delivering that. And then we could also learn where each of them fell short. I mean, obviously, we're not going to build our own LLM. We are going to fine-tune based on the area of expertise where we're focused.

34:18And, you know, there are a couple private ones that we're working with now. Yeah, that's a beauty. We're not looking to, I hate to say it this way, we're not looking to train ChatGPT or OpenAI's model to do what we're going to do, right? So we're building the layer on top. And you don't need to now. That's the beauty of this shift that's happened. It's truly kind of democratized the ability to use AI. And that was the big blocker because, I mean, you look at ChatGPT and these models, they literally are spending, you know, up to millions, billions of dollars to train the models. And they're running for literally days on end to train it.

35:05And now that you have these options out there, the open source options, they were actually doing a webinar today, which when this airs, it'll be in the past. But looking at how to leverage open source models via Hugging Face to just build a simple RAG retrieval augmentation generated chatbot. I'm sure you're familiar with that. Sounds familiar. Yeah, exactly. And that goes back to the context element, which is so critical. But I think the LLM component will become commoditized in a good way because it'll be available to everybody. And that's where proprietary data, what you're training it on, on top of it.

35:48And then the context element, that's where back to strategy, that's where the differentiation comes in. That's where you can actually build a moat. And I kind of want to shift there just back to the strategy side. I feel like the, the ability, ability to build a moat just got a lot harder. You're seeing a lot of products out there that are just kind of wrappers on chat GPT. chat gbt comes out with new functionality and it wipes out a whole slew of products so how do you think about moat building in this new age we're in yeah i think one part of it is about content so you know um content partnerships and um you know licensing is hard in this space it's like Nobody's figured that out exactly.

36:37It's not like we're Spotify and we can just, there's a model for breaking up the album into songs. In this case, it's like dissecting a book or articles or what have you into the soup. So it's harder to figure out how to do that. So we're thinking strategically about our partners in that space and doing POCs. I'm not sure. I can't really talk about who those partners are without their permission to do so. And it's a little early for that. But we're really excited about the organizations we're working with, leveraging their content because they believe in this idea as well. And we're kind of figuring out new ways to add value or exchange value with them.

37:35So it doesn't necessarily have to be cash for each time an article is referenced as a source in something. There's lots of other ways we can deliver value back to those publishers and authors and experts. So I think that's one part of the moat. I think there are some features that to your point, like probably will become part of the broader LLM space, but I think in the short term, they can help us win. And for this use case, So like there are things that can help us win with this use case. And once we're in, then like, you know, if there's another place for that, if we're already the trusted source, then we're the trusted source and it's harder to pull you out.

38:36um but i think go to market is the other piece is like the go-to-market partnerships and our strategy for how we deliver this um it's not i'll just say this it's not the usual suspects it's not the accentures and the sis and like a lot of tech companies will um you know we want to have relationships with those organizations but um we're not we're not relying on them as our go-to-market, so we're being smart about who makes the most sense in this context to partner with us and be the face of our organization. Yeah, it's really interesting. I think all new business models are going to emerge from this.

39:25You mentioned just the other day, I believe it was a partnership between Reddit and Google leveraging their data. And it's like, oh, well, here's this whole new revenue stream we have that we've just been building over time. So I think that's going to be a super interesting part in terms of how, how these business models get defined, how they evolve, because it's not to your point, it is kind of messy. Like you mentioned, like the soup metaphor, it's like baking a cake. Well, it's all in the cake. So which piece am I, you know, divvying up? I want to go back to one point you mentioned earlier, and it gets to kind of the human connection side with AI.

40:04And you mentioned how people were interacting with the AI, even if it was simulated, was completely different. And I talked to a friend the other day, they said on their commute, they're using the chat GPT function where you can talk to it. And they just talked to it on their commute. And I'm like, what are you talking to the AI for, you know, 30 minutes at a time? but it is a different interaction like how do you see that evolving uh over time with that difference of interacting with ai versus say another human and what happens there well i think it's important to recognize what the user in any of you know any of the use cases so like what the user wants to get out of it.

40:50Like, is it entertainment? Is it comfort? Yeah. Is it something actionable? So I think AI can do all of the above. It's just about design the experience for what your users really want and expect of you. So in our case, like we're not, We're toning down all the kind of cozy coddling kind of thing. Because we are hearing from our users that what they want is the thing that most of their, you know, friends and colleagues aren't giving them, which is a kind of a bit of tough love and say like, hey, you might look at it this other way. And it's funny because I think managers can often give their employees a bit of that.

41:46And the employee might think, oh, I'm not really sure what their motivations are. I'm not, you know, I don't know. Like, hmm, I better think about that. I'm not sure if that's right. and you're not going to open a book and read a book on management and figure that out right so like there's something in between that is like actually i want to go to a source of truth that can either validate with my what my manager said or my manager can go there and say like actually this is a much better way to answer this question than i would have it like i would have um but there seems to be a need for an outside source that is just totally objective and just this is how I would break down the problem like does it seem to be illegal unethical or just uncomfortable guess what like 99.9 % of the time it's just something uncomfortable and if you can acknowledge that's uncomfortable then we can talk about like how do you fix that and so um that's just it's just um i i can also imagine a super friendly cuddly teddy bear style ai that is really just there to like tell you how awesome you are like it's just yeah you want you know that's a different purpose which may be uh that may be a bad thing in the long run but i like the approach that you're you're taking where it is, you know, kind of the tough love, the getting to the real meat of the situation.

43:21But synthetic humans, big pivot right here. It sounds scary, but we talked a little bit about this and another just really genius approach to starting something today. What are synthetic humans? How are you using them? Yeah. So based on all the research that we've been doing, we've identified really 10 personas or archetypes i guess you know i like the term archetype um but we've put some persona around the archetypes so um we know what the top challenges are that workers are facing most most regularly we also know the kind of the one-off use cases or the the less common things but we've built these um we've trained 10 personas 10 ai to represent these different challenges or or represent workers who have these challenges and we've also designed them to represent um different geographies different underrepresented groups Just putting more context into who these individual personas are.

44:38And so imagine if you had a human come in and interact with the AI. In this case, we've got these AIs trained to represent an individual of a particular background with a particular lived experience with some particular problems at work that is asking us first like a whole bunch of questions. So our wizards can answer AI questions and get a lot more questions than what we would get from doing user testing. So complimentary. We also then moved on to have these personas asking AI directly a bunch of questions. Wow. Which is, it's gotten really interesting. I would say 50 % of the questions that these synthetic workers are asking are brilliant, are really insightful, really interesting.

45:38And 50 % are just like gobbledygook or not that just they're not interesting. They don't help us understand a problem that we want to solve. And so we're consolidating all the best questions from each of these personas back into the themes where we need to become experts and then we're able to use that to kind of direct what content we need to be able to fine-tune the AI to do to have those conversations and and we've actually like really had some breakthrough with that like I'll just give you an example we had a scenario about micromanagement and the questions weren't necessarily about micromanagement they were about behaviors that the AI could tell, relate to micromanagement and then could make recommendations for how to deal with a manager who is a micromanager and things you might ask about are ways to navigate that situation.

46:39And it was just like, wow, this is really cool. It's really working. I'm getting chills thinking about this in a good way, not because synthetic humans are a scary thing. But I go back to back in the day, previous roles, we would define personas. they would live in a PowerPoint deck and as far as they went but you're actually giving life to these personas and not only giving life to them but you're using them to further train and enhance in this case the LLM which is the underlying machine of what you're building it's crazy and so the synthetic human which is an AI is interacting with the LLM the tool that you're using so you have ai interacting with ai is there was there any just like weird like off the wall things you saw happen out of that or i mean sometimes just got really repetitive like it was it was it kind of got in like these repetition loops like they're not even having a conversation my manager's a micromanager oh it sounds like your manager's a micromanager that kind of thing would happen like um but no i mean we're we're really cautious We've had some outside guidance from AI researchers that have helped us to do this as well.

48:01So, I mean, we're trying to be very ethical about it. Yeah. But in terms of weirdness, like I said, 50 % of the questions and things that were coming out of that, we were just kind of like, let's scrap it. But the fact that 50 % was good, it's not about like, oh, now our persona is great. Like, actually, what we just want to do is get more research. And if this if this, we have to still evaluate whether or not each exchange with this synthetic persona is is worthy to learn from or not. And that's humans doing that. So. you have a whole new product idea right there you could do synthetic humans which i'm sure that maybe companies already doing this but you're building one yeah that i heard of doing this for market research where they've simulated coal yeah to your point about personas in a in a powerpoint isn't it interesting though like take them out of the powerpoint and have them interact with you Mm, that's crazy.

49:01So the last thing I want to hit on, you mentioned earlier being at Lotus, you kind of saw the birth of the internet and all these transitions with mobile and cloud and things going on. Where do you sit where we're at or where do you put where we're at right now with generative AI at what scale? Is it at the scale of the internet more or less? Bigger. Do you see like dot-com bubble hype? how would you equate it having worked through through that experience i mean i i feel like um there's there are a lot of there were a lot of technologies that i thought were hype like just personally like crypto i mean i know people are huge crypto fans i'm just like like um and nfts and stuff like that.

49:50We all have our opinion. But this feels, this is going to transform our lives in ways that are just really hard to even imagine because our work is going to change. Our relationships are going to change. I think we're going to have like bots doing all kinds of things for it. Like bots negotiating with bots to do a lot of the work. I think problems will fix themselves. There will be, you know, your car is going to know it needs something And just, you know, I think Tesla already does this, like just drive itself to the nearest gas station and get whatever you need if gas stations still exist. But I think that this is transformative.

50:34This is a tidal wave. And I think you've got people who are sitting on the beach looking at the weather thinking it's a sunny day and they see this wave out in the distance. It's like, it's pretty. And then you have some people running for the hills. They're like, oh, my gosh, like this is this is scary. And you have some people grabbing their surfboards and going out to like check it out. And those people on the surfboards are the ones who are going to define what this is. And I think that's a wake-up call for everybody. Like, who do we want doing this work and how and what solutions? And, you know, so I'm hoping that people really lean in.

51:15We need to democratize the innovation in this space and include all kinds of people in that really pioneering work that's going to inform how we live. Yeah, I love the kind of surfing battle wave analogy. And like at Hatchworks, we're doing exactly that. We do custom software development, but right now we're digging into every aspect of the software development lifecycle from strategy design, development, testing, deployment, all of that. and we're seeing where we can leverage generative AI in that. But there's tools, but then there's this foundational change of how you interact. We have architects now that are defining the skeleton of the architecture working with gen AI.

52:01Then you're kind of taking it iteratively as you go, getting the framing done. So you're not having to worry about the syntax and all these minutiae of things. You're focused on the broader thing and the more valuable work but it's been really interesting uh digging in that with with our engineers and architects yeah yeah so i appreciate you being on the the podcast so i i want to point folks to uh to tough day where should they go and you mentioned a survey as well is that something that's still running that people can go take where should we point people and how can they find you Yeah, tough.day, tough, like T-O-U-G-H dot day.

52:44And you'll find the survey. I'm assuming by the time this airs, it'll still be available for people to continue to share their point of view. And, you know, we'll have more and more available on that. If you're interested in user testing, we're always looking for user testers as well. There's a spot on the site where you can sign up for that. Nice. We got to get Hatchworks using it as an early user there. But thank you, KBJ. This has been one of my favorite nuanced discussions. So much great meat to this. This is one that I'll go back and listen to again. But thanks for being on. Thanks so much.

53:27Thanks for listening to Built Right. If you enjoy the show, give us a follow or subscribe on your favorite podcast platform. And don't forget to leave us a review. For more info on Built Right, visit us at hatchworkbuiltright.com. The single biggest mistake we see companies make with AI is they don't properly train their teams. We see it all the time. Companies roll out AI tools and expect people to just figure it out. But using AI effectively requires a totally different mindset and skillset. And that's exactly why we built training for every level of your org, from AI training for teams and executives to training engineering teams on our generative-driven development methodology.

54:07Or if you've already identified your AI use cases and want to just prioritize where to start, we offer an AI roadmap and ROI workshop to help you build a quick plan. It's all about going from we should use AI to actually driving real value with it. Head over to hatchworks.com to learn more.

From the publisher

Generative AI has set workplaces around the world on a path toward reshaping business models, team structures, and how they look for tech solutions. And it shows no sign of slowing down!

In this episode of the podcast, we hear about a new AI platform that offers a unique solution to modern workplace challenges.

Katherine von Jan, aka KVJ, is the Co-Founder of Tough Day, an AI platform that helps workers with everyday challenges. She joins the Built Right podcast to share her brilliant strategy for taking a unique idea and bringing it to market. We hear about how she first came up with the idea and put it through rigorous testing and research methodologies to test viability.

Katherine also shares her take on AI in the future of workplaces, the human connection in AI, and why Tough Day has been building AI archetypes to better understand different customer use cases.

Discover how KVJ's vision is reshaping the workplace with AI, offering a beacon of support and clarity for workers navigating today's challenges. Listen to our podcast for insights from leading figures in this evolving field. Subscribe, share your favorite clips, and join the conversation today!

Key moments: 

  • Why Katherine built Tough Day
  • The challenges of being an employee or manager in today’s world
  • Where Tough Day can help employees who don’t have anyone to turn to for advice
  • Katherine explains her early strategy that got Tough Day from an idea to a working solution
  • Why an AI-first mindset is needed today
  • How Katherine and her team put Tough Day to the test
  • Working with LLM’s that may change in time
  • Why Tough Day is building AI archetypes to understand use cases
  • Where generative AI is headed in the future

Key links: 


Mentioned in this episode:

AI Opportunity Finder

Feeling overwhelmed by all the AI noise out there? The AI Opportunity Finder from HatchWorks cuts through the hype and gives you a clear starting point. In less than 5 minutes, you’ll get tailored, high-impact AI use cases specific to your business—scored by ROI so you know exactly where to start. Whether you're looking to cut costs, automate tasks, or grow faster, this free tool gives you a personalized roadmap built for action. 👉 Try it now at https://hatchworks.com/ai-opportunity-finder/

Talking AI - Conversations with AI experts and early adopters

Welcome to the Talking AI podcast, where we dive deep into the world of artificial intelligence with host Matt Paige. Formerly known as the Built Right podcast, Talking AI brings you insightful conversations with AI experts, founders of AI products, and industry leaders who are leveraging AI in their businesses. Whether you're an AI expert or a beginner, our episodes will help you understand how AI technology works and how early adopters are deriving value from it. New episodes drop starting August 6th.

More from Talking AI

All 84 episodes
Navigating Workplace Challenges with AI: The Tough Day SolutionTalking AI · 54 min
Listen in VO