“If you haven't done great, well-thought-out customer research and can demonstrate that you will continue to do that, you cannot have my LPs' money.” — Michelle Gilmore Georgie Healy speaks with Michelle Gilmore, co-founder of Juno, an AI-driven interviewer built to deeply understand customers without bias. Michelle, a seasoned behavioural researcher, reveals the overlooked importance of thorough customer research, sharing why so many founders avoid it despite its immense value. She explains how Juno overcomes human biases in traditional surveys, enabling businesses of all sizes to access honest, real-time insights previously reserved for companies with large budgets.
Michelle outlines her rigorous approach to building an AI product, rooted in experimentation, continuous validation, and a willingness to venture into uncertainty. She also discusses the challenges and paradoxes of scaling an AI startup, including navigating abundance and staying mission-focused.
Throughout the conversation, Michelle offers candid reflections on her entrepreneurial journey, the reality of running a venture-backed AI company, and why success hinges on deeply understanding human behaviour. She also touches on Juno's thoughtful personality design, balancing neutrality with empathy to capture truthful customer experiences.
Chapters
1:10 Introducing Michelle Gilmore and the vision behind Juno ▶ 2:19 Why so many founders fear talking directly to customers ▶ 5:00 When should founders start customer research? (Hint: it's immediately) ▶ 7:24 What does a behavioural research expert actually do? ▶ 9:18 How AI helps overcome human bias in surveys and interviews ▶ 13:42 The hidden dangers of relying on biased data (Michelle shares a shocking story) ▶ 16:35 You've gathered customer data. What's next? ▶ 21:07 How Juno's personality was thoughtfully designed ▶ 25:59 Can AI models really detect human emotions during customer interviews? ▶ 34:05 The hardest part of building an AI product that no one talks about ▶ 38:23 Navigating the tough choices: are there companies Juno wouldn’t work with? ▶ 40:35 Michelle’s view on uncertainty and why founders should embrace it ▶ 47:01 What really keeps a venture-backed founder awake at night ▶ Transcript Search or select a timestamp to jump ▾ 0:02 Georgie Healy: Founders scale faster on Deel. Set up payroll for any country in minutes, hire anyone anywhere, get visas handled fast, and get back to building. Visit deel.com/dayone. That's d-e-e-l.com/dayone. What's the most challenging part about building an AI product that people might not be thinking about? 0:28 Michelle Gilmore: Because I have met too many founders or want-to-be founders that have done little to no customer research, and I find it baffling. It blows my mind. It is the single most important thing that you will ever do. 0:44 Georgie Healy: But aren't you, with all this data, aren't you like, "Oh, just let me build it"? 0:47 Michelle Gilmore: If you haven't done great, well-thought-out customer research and can demonstrate that you will continue to do that, you cannot have my LP's money. I don't understand how it is responsible to invest if that is not present. 1:02 Georgie Healy: What about human emotion? Can Juno identify frustration or even sadness or anger? Like, I can't imagine how you'd build that into a model. Hello and welcome to In the Blink of AI, where I talk to the brightest AI startups and innovators each week. I'm Georgie Healey, and this week I'm speaking to Michelle from Juno. In Michelle's words, understanding people isn't just about asking questions. It's about asking the right questions and deeply understanding the answers. And if you can't do that, you'll never truly know what they need, what they feel, or why they act the way they do. Juno leverages AI and Michelle's 12 years of expertise to help businesses, governments, and organizations listen to customers with depth and clarity. Guys, this was a pure joy. I genuinely had the most insightful and fun time speaking to Michelle. I not only learned a lot about, you know, an exceptional person building in the industry leveraging AI, I saw firsthand by my questioning, by my unconscious bias, how like that can creep into questions and why a product like Juno is so valuable. That wasn't planned, and I just can't wait for you to listen to the show. Let's do it. Hi, Michelle. Thank you for joining in the Blink of AI. We're so excited to have you. But first of all, let's start with a quick explainer on what Juno is. 2:42 Michelle Gilmore: Juno is an AI-led interviewer. So what that means is that Juno talks to your customers for you so that you don't have to. It gives you an unfiltered line into your customer's thoughts, what they really think in real time, and it does it better than any other product on the market because I did the job for 12 years. 3:06 Georgie Healy: I am going off script already. It's question 2, but I have to ask you. 3:11 Michelle Gilmore: Please do. 3:12 Georgie Healy: I was part of a VC-backed cohort of founders, and again and again, none of us wanted to speak to the customers. What's that about? Why do people not want to? 3:23 Michelle Gilmore: I have a lot to say on this. Please. We may need an episode just on this, and I'm, I think maybe we need a different podcast to talk about this because it might be a little bit off topic. 3:33 Georgie Healy: I'll start a new one just for this. 3:35 Michelle Gilmore: Okay. There is a little bit of fear that is fed by what if I've come all this way and I ask people and they hate it? Because I have met too many founders or want-to-be founders that have done little to no customer research, and I find it baffling. It blows my mind. It is the single most important thing that you will ever do. I talk to my customers all day, every day. It is the highlight of my day. And the reason for that is it makes me such a great product designer. It makes me such a good system designer. If I go and design some logic architecture after talking to a customer and they have demonstrated a mental model to me, I am a better designer for it. I don't— 4:24 Georgie Healy: Yeah. 4:24 Michelle Gilmore: Understand why you wouldn't want to mainline this gold. And I also understand deeply why humans don't like it and why most are bad at it. So they should just use Juno to do it because then they don't have to worry about it. 4:38 Georgie Healy: Well, exactly. I was one of those founders that was like, I feel sick about like talking to my customers. And I'm not surprised that VCs in particular were so excited to back you, Michelle, because they see their founders do that and they're like, oh my God, please, I'm begging you to talk to customers. 4:58 Michelle Gilmore: Do you know what I find very interesting is if I was a VC, when I start my fund, a requirement that is going to be in my diligence is if you haven't done great, well-thought-out customer research and can demonstrate that you will continue to do that, you cannot have my LP's money. I don't understand how it is responsible to invest if that is not present. 5:23 Georgie Healy: I would love to know what the rates of conversion are if you've already got this willing and active group of customers that are like, I am just waiting for the product to exist. Do you build a product and then reach out to customers, Michelle, if you are giving advice to founders? 5:40 Michelle Gilmore: No. 5:40 Georgie Healy: No. 5:40 Michelle Gilmore: It's, it's at every stage. So you can do conceptual validation. So before you have anything, you can test sentiment, or you can learn more about the behavior that you are trying to serve. You can do that just by talking to people. So you could say to Juno, I want to design a new way to, um, house and charge pods. 6:03 Georgie Healy: Yes. 6:04 Michelle Gilmore: That's a pretty expensive endeavor. I want to make sure that I very much understand the problem relating to the product right now. So I'm in a really hot room right now. I hate wearing headphones when I'm hot. Maybe that's just me, or maybe that's everyone. And there is a big market gap, which is headphones that cool you down when you are in this very suffocating sound booth. Is it a problem for other people, or is it just a problem for me? I don't know. Let's go and ask 100 people right now that fit my ideal customer profile, and then I will have a better understanding of my total available market. 6:40 Georgie Healy: Michelle, how many times have you heard people say, oh, I had that idea, I just didn't launch it as well? Like, like, imagine you had actually done the research and seen, oh my gosh, this is a real problem, not just something that runs around in my head, and then I never do anything about it. 6:57 Michelle Gilmore: I hear that a lot, I'm not gonna lie. 7:00 Georgie Healy: Okay, let's get, let's get to the interview. 7:03 Michelle Gilmore: Let's just tear the script up. Yeah, let's just go with it. 7:06 Georgie Healy: I've always wanted to do that, like in the movies where they do it, tear the script up, and they're like, literally, I'm gonna tell you what's wrong with this university. 7:15 Michelle Gilmore: This is the problem with this digital world. We no longer have these analog devices to express our emotions and punctuate these things. 7:23 Georgie Healy: Yes. 7:24 Michelle Gilmore: I don't have it printed out, so I can't rip it out, rip it up for you. 7:27 Georgie Healy: I know, I know. That was a miss on my behalf. I think I could have printed it out anyway, pretended that I was reading off a script. Look, I genuinely want to know, you've touched upon it before, having, uh, over a decade of experience in this space. You're a behavioral research expert. What does that mean? 7:46 Michelle Gilmore: If you were referring to me as that, you would mean, I imagine, That, uh, I have spent over a decade, uh, researching human behavior in the context of various humans, whether they were at their work, whether they were at home, whether they were, um, about to get on a plane or about to buy something in a supermarket. Uh, my company Neo, which was a, uh, channel agnostic design agency, so we designed product systems and services for various brands, mostly B2C brands. What we did better than anyone was deep behavioral research to inform the design of the thing that we were trying to achieve. And we got really good at that. So that's, I think, what, what that means is that I am very good at exploring human behavior and then turning that into practical insight. 8:39 Georgie Healy: Are you itching to just create a product again then at the moment? Like, I know your Juno is your baby and it's like an incredible, exciting product, but are you with this data? Are you like, oh, just let me build it? 8:51 Michelle Gilmore: Yes. And I believe that my investors and stakeholders want me to stay focused right now and that I'm not going to take that bait. And my focus is on making Juno better and this company succeeding. 9:05 Georgie Healy: You know, I back that 100%. And when it does succeed, I can't wait to see what you build as a product. 9:12 Michelle Gilmore: You'll be the first to know. 9:14 Georgie Healy: Yes, thank you. Yes. I'll send you my number. So, I read an article where you discussed researchers adding bias into survey questions. So obviously, you know, that's one way of getting the customer feedback is to send a survey. But if I'm writing it and I know what answer I want them to give to justify me building the product that we've talked about before that I've already sunk a lot of capital into, how does AI avoid this? How do you avoid this? 9:43 Michelle Gilmore: So AI generally, I don't think does avoid it, but Juno does. And Juno does because it has been made to, and made to by myself and my co-founder Josh. So people insert their own biases into the data. And by that I mean they may say, do you love this podcast? If I say that to you, I am giving you a lot of signals to tell you that I want you to tell me that you love that. Now we know that humans are more likely, particularly if there's eye contact, particularly if it's human to human, particularly if there is a little bit of, um, self-consciousness on the participant's behalf. So certain personalities are going to take the path of least resistance and they are going to say, yes, of course I love it. Now, if we ask a neutral question, which is, How do you feel about this podcast? What I am telling you is I am neutral, therefore it is a safe space and you can say whatever you want. Now, humans are really bad at that, generally speaking. Now, what we have done, and this is one of 51 behaviors that Juno demonstrates, right? We have ensured that Juno doesn't ask leading questions. 11:01 Georgie Healy: Mm. 11:01 Michelle Gilmore: Now, that doesn't mean Juno doesn't have bias. Nothing doesn't have bias. If it is, if it is an entity, then of course it has bias. The question is, what are we going to skew those biases towards? And Juno is a system that is evaluated every day by me, someone that has done this job for 12 years and is monitoring the behavior. So we are in a place to ensure that we are producing data that is as true as possible. And by truth, we mean it represents the opinion of the participant as they gave it. 11:38 Georgie Healy: And I'm dying to know, you gave that 51 types of questions. Don't give me any secret sauce, but like, does it depend on who you're asking the question to? 11:49 Michelle Gilmore: Are these clusters of different things where some have, you know, 10 over here, but 3 over there and Yeah, so our, um, a few of our competitors that are chasing us will absolutely be listening to this, so I'm not going to— 12:02 Georgie Healy: No, no, no. 12:02 Michelle Gilmore: I'm not going to give you the framework, but what I will do is tell you that we have developed a set of characteristics, traits, um, and a way to evaluate those minute by minute. Some of those evaluation mechanics are still done manually with me in relationship with that technology, and others we have now automated. But those ones that are automated are set up on a scoring system that I have initially informed. So Juno is built on a very, very guarded approach which privileges quality, and by quality we mean a high confidence in accuracy of the data. So when we say someone said X, we're very sure that they said X, and we're also very sure that we didn't lead them to say X. 12:51 Georgie Healy: I kind of wish I spoke to you before I sent out my first ever quiz about the podcast because I'm kicking myself now, Michelle. I wrote in it, literally verbatim, I said, would you recommend this show to your friends? That's leading! That's leading! 13:11 Michelle Gilmore: It is, it is. But hey, there is a whole industry on NPS and one of those questions is how likely are you to recommend this to a, to a friend. 13:23 Georgie Healy: Mm. 13:23 Michelle Gilmore: There's many board reports that still have NPS scores in them, right? And it is a, a very, very biased comparison tool. You, you are, you are there with the best of them. There are some people that are making incredibly important big decisions daily and are using very biased data to inform those decisions. So, um, I would let yourself off the hook there. 13:51 Georgie Healy: That's kind of you, but if I'm a large enterprise and I'm not using Juno and I've got, oh yeah, 99% of people said they would recommend a show to their friends, but actually they don't really mean that. What can happen? I guess you've kind of alluded. 14:05 Michelle Gilmore: Can I tell you a story? And I'm okay, I'm okay telling you the story because I, um, I saw it happen firsthand. So I was once in a large Australian telecommunications, uh, retail branch. And I saw a staff member chasing a customer out with a clipboard. And I said, "What are you doing?" And she said, "Oh, I need this customer to fill out this NPS survey because they had a really great experience and they're going to say really great things." And I said, "Why do you care?" And she said, "Because I get rewarded if they do." Oh no. So that whole system, not only the mechanism of NPS and the very notion of surveys in that context, but you are building a system around the method of surveys, that is inherently— I mean, that's not bias, that's just negligence, right? 14:57 Georgie Healy: Yeah. 14:57 Michelle Gilmore: And those outcomes are then going into steering committees, boards, and those boards are making decisions based on the fact that 72% of people last week said they were more likely to recommend Telstra than Optus. 15:11 Georgie Healy: Yeah. 15:12 Michelle Gilmore: How is this responsible decision-making? 15:14 Georgie Healy: Yeah. 15:14 Michelle Gilmore: Right? And the, the reality is it's not responsible decision making. And if you get any of those stakeholders alone, they know it's not. They just don't have another system to replace it, and they have to wind back so much. The great exciting thing I think is products like Juno are leveling the playing field because you no longer have to have big budgets to buy data that has high integrity. You can have very low budget, you can have very small amounts of time, you can have a very small team, and you can now listen to your users, audience, employees, customers, citizens in real time. And that's really exciting. It means that you don't have to have money to get good data anymore. And what's that going to do for the world? 16:01 Georgie Healy: Wow. And, and normally to get the kinds of quality insights that you can do with Juno, what would that take in the past? I'm guessing a huge team. 16:09 Michelle Gilmore: Yeah, a lot of money. A lot of time, huge, huge human-led effort. And hey, I was one of those practitioners for a really long time. 16:20 Georgie Healy: Mm-hmm. 16:21 Michelle Gilmore: And it was a lucrative and fruitful and incredibly inspiring endeavor for a time. But that time is gone, and I don't think that we can responsibly do that anymore. And we can't do it at the scale that we need to. We need to make sure that to do great listening, you don't have to have huge budgets. That shouldn't be an entry requirement, and it no longer is with Juno. 16:46 Georgie Healy: Hmm. 16:47 Michelle Gilmore: Okay. 16:48 Georgie Healy: Now we've got this unbiased data from all the customers and it's meaningful. Is that the hard part done? What, what, what do you do once you've done that? 16:58 Michelle Gilmore: Yeah, sure. So it, it depends on the customer and how they want to consume it. So the point of Juno is that Juno goes out and listens to the people that you want to hear from, and Juno then brings that insight back. You may want to listen to that, you may want to watch that, you may want a slide deck, a report, you may plug straight into the API and there may not be an interface at all. So the point is that we listen to people for you and you then hear those outcomes in whatever way suits you. 17:29 Georgie Healy: And you know, you talked about human intervention and things like that. Where do you intervene, Michelle? Where are you like, this is something— oh, you can't really tell me exactly, I'm sure. What can you answer in terms of human intervention? Big picture AI. 17:46 Michelle Gilmore: I think that Juno's still a very young company, even though it's built off deep domain expertise and the legacy of Neo, which was our design agency. And Josh and I have worked together for 12 years. Juno is, as an AI entity, in relationship with us as the humans and the founders and the CEO and the CTO. So, I see us as a team. Juno is a living, breathing entity that we are in relationship with, and we are augmenting our human capability with that technical capability to deliver Juno to the world. So it is a, it's a partnership, you know, it's a companionship. I almost feel like Juno is our third founder. I love it. My son thinks Juno's a person, by the way. So I think that tells you that is validation enough that I refer to Juno as a person because he is often asking me who Juno is and why Mummy spends so much of her time with Juno. 18:49 Georgie Healy: Listeners of the show, Michelle and I were talking about our children before this. I've got a 3 and 5-year-old, and Michelle's son is 4, just for context. He's not 14. Like, um, I love asking where people get their names for their models and things like that from. Where does Juno come from? Why did you pick the name? 19:09 Michelle Gilmore: Uh, we wanted Juno's name, which now feels like it was always inevitable that it had to be Juno, but there was a time when it didn't have a name and it needed to be universal. And by that I mean it wasn't, uh, culturally or, uh, language-specific or accent-specific. Specific? So could it be something that anyone can say? Anyone can say easily? Can it translate? Can it be interpreted? Does it have origins that we are okay with? So Juno has very loose ties in Greek mythology to a very strong, powerful female, which I've been asked a lot about. That certainly wasn't something that we looked for. 19:55 Georgie Healy: They're like, "I see what you did there," and you're like, "Mm, not actually, but okay." Yeah. Yeah. 20:01 Michelle Gilmore: I've been asked, I've been asked a lot about it as if it was intentional. Um, it's not, but I'm also okay with it. So, so once you, uh, get your wishlist, there's actually not that many options. Um, and then it became about what was taken, what wasn't taken. Um, and something that Josh and I resonated with just from a personality point of view. Did it feel like the entity that we wanted to build? Could it convey the characteristics that we felt were really important? And once Juno came into the world, it was almost inevitable that it had always been that. 20:35 Georgie Healy: I love the name. You know, it's, it's funny you mention that. It really shows your product background too. Like, if we're going to pick a name and it means death in a different language, you know, that's not going to be great for us and things like that. Yeah, it's incredible. 20:51 Michelle Gilmore: Yeah, it was a very, uh, logical and systematic view, and, and at the same time there's a whole lot of heart in the brand. There's a whole lot of, um, responsibility that Josh and I take very seriously. We're very, we're very proud of who Juno is out in the world and how Juno behaves out in the world, and we will continue every day to make sure that that continues. 21:15 Georgie Healy: And on that frame, like, Juno is kind of a person in my mind now too. How does Juno personality-wise— is it a very neutral party, a very, like, uh, a mixture of you and your founder together? Like what's the vibe like? If there's a celebrity, is it, is Juno like a celebrity I can think of? 21:35 Michelle Gilmore: So most large language models, if you've noticed, are quite agreeable. 21:40 Georgie Healy: Yeah. 21:41 Michelle Gilmore: And they are built at its core, one of the primary attributes is to be agreeable, if not praise the human that you are talking to. Juno does not exist to please people, certainly not our customers. Juno exists to get truth out of the people that it is talking to and to make sure that participants remain comfortable and safe. 22:09 Georgie Healy: Mm. 22:09 Michelle Gilmore: And so Juno is, as a data extractor, quite neutral, very, very task-focused. If you try to take Juno off track, Juno's gonna take you right back to the point. 22:22 Georgie Healy: Fantastic. 22:23 Michelle Gilmore: Juno's not gonna talk to you about the weather. Juno's not gonna make you feel good about yourself. Juno's gonna say, hey, we're talking about headphones, let's come back here. But if you say, I'm actually feeling a little uncomfortable with this questioning, then a little bit more personality's gonna step in. So we are thinking about these behaviors and how this entity behaves every single day, every minute of every day. And it is incredibly important. When trying to extract truth from another human in its purest form. And that means that you have to allow them to express themselves and to express themselves on their own terms, in their own language, ideally. 23:03 Georgie Healy: Thank you so much. That's, that's very helpful. And when, when someone's responding, how does Juno know when it's genuine feedback or kind of noise? Like how, how does it know? 23:15 Michelle Gilmore: Well, I mean, that's a very subjective notion, right? So if it's the person's opinion, then nothing is necessarily noise. Rather, what we need to make sure we do is over time build in ways to deliver data to our customers that we can be very confident in. And we have built a proprietary way at Juno to do that. That doesn't mean that the participant that is talking to Juno ever knows that that is the case, but there should be, and there is a way to say, relatively speaking, we were able to prove X, but we weren't able to prove Y. That's not about good or bad. That's about making sure that there is relativity in the data sample. 24:00 Georgie Healy: It makes me think of the times I've been in a room where there's training or a workshop and the facilitator will ask a question, and if that person answers is it not the way that the facilitator wanted or the wrong answer? The way they respond to that seems to dictate the way that participant will engage in the whole rest of the workshop. If they're like, Yes. That's not what I was looking for. Or even just, nope, not that. Anyone else? That person will pipe down instantly. Like they won't raise their hand again. I don't know if you notice that. 24:31 Michelle Gilmore: Yeah, absolutely. And that, that, I mean, that's just really crappy facilitation, right? We can, I think we can just that's just very poor, low-level showing up from someone in a position of authority. 24:44 Georgie Healy: Mm-hmm. 24:44 Michelle Gilmore: Actually, if someone has had the courage in a physical context or in a digital context to share, shutting them down is not only not productive, it's just mean. The great thing about Juno is you are having a one-to-one conversation with an AI agent, and it is private. And it is personal. And if you want it to be, it's anonymous. And we are seeing people be much more open, reflective, honest than I ever did in 12 years of, of doing contextual interviews all over the world. And I think that we are getting really interesting insight in topics that people usually feel very vulnerable in or very self-conscious in, whether that is, um— sex, money, children, relationships, being honest about what I actually eat versus what I want you to think I eat, all of these things that make us complex humans, we may just be able to start to get some real stories that matter out and from places that we can't necessarily step into as practitioners. So I think that's really, really exciting as well. scaling this capability to allow people that would never participate to participate in research is, is a beautiful thing. And I think we're going to create something really compelling. 26:09 Georgie Healy: What about human emotion? Can Juno identify frustration or even sadness or anger? Like, I can't imagine how you'd build that into a model. 26:22 Michelle Gilmore: I think if it, if it is relevant, Yes. So Juno's not a companion. Juno doesn't exist to do anything other than to listen well and to ask well and to understand reasoning and motivations. So if emotion is important and relevant, yes. When we did our, our study for The Australian, when they were doing their— Danielle Long used Juno to support her article when the ACCC were taking legal action against Woolworths and Coles. There was a lot of emotion in that study. 26:56 Georgie Healy: I can imagine. 26:57 Michelle Gilmore: There was a lot of Australians that were very angry. And when we spoke to 1,000 Australians as the US election was unfolding, again, we have thousands of hours of conversation of people hearing the election results and having various responses. And in that context, those emotions were incre— extremely important for us to gather and understand and quantify. In other cases, not so much. So it depends if it's relevant. And the great thing about Juno is Juno will determine whether that is important context based on the objective or not. If you're telling me that you're really sad because it's raining today and I am doing a study for a new set of head— headphones that take my temperature down when I'm sitting in a booth, I'm probably going to say, Michelle, we need to stop. Talking about the rain, or how is the rain connected to headphones for you? If it's not, let's get back to talking about sound and temperature and how we might change that. 28:01 Georgie Healy: I love this. And I kind of, uh, not that you asked, but I think I'd prefer it was a, a model that I was speaking to in that case, because then I'm like, oh yeah, good point, I am getting off topic, but it's a model. So it's like, you know, I didn't do it in front of the whole workshop. That's great. 28:17 Michelle Gilmore: Exactly. And, and as a human in people's homes or in people's work or As a human interviewer, and I'm— I was very good at it, I am very good at it, it's a tough thing to do to not acknowledge, listen, hold, and then somehow divert back to the reason why you're there. 28:35 Georgie Healy: Yes. 28:36 Michelle Gilmore: And it takes a lot of time and a lot of energy and big budgets, whereas Juno has a different role to play, and Juno is unapologetically focused on talking about the topic that we are there to talk about, unless you as the human can prove there's a link, which you may be able to do. You may say the rain is relevant because I'm gonna keep my headphones in and run outside. Juno will then say, ah, rain's interesting. Talk to me more about that. Just like I would, right? 29:07 Georgie Healy: Yeah. 29:07 Michelle Gilmore: But if I need to say, Georgie, for the love of God, please stop talking about the rain. It's very hard to do that as a human when you are in someone's home interviewing them. So we tend to be polite, particularly in Western dominant culture. We tend to be polite. We tend to privilege civility, and it hurts our data. 29:27 Georgie Healy: Yeah, I couldn't, I couldn't do it. And I'm interviewing you right now. If you said you were feeling uncomfortable, I'd be like, let's just talk about cats. No worries. Forget the AI. 29:38 Michelle Gilmore: I might throw something in and see, let's see. Let's see if I can, I can throw you. 29:43 Georgie Healy: I did realize as you were talking, oh my God, she's seeing everything I'm doing wrong right now. I've been doing it for 9 months. Look, I would love to ask you, because you're so clearly so attentive to the product you're building, one, because of your background, two, because of, you know, the human emotions and the kinds of data that you are capturing. Building this kind of product, I would love to know what mattered for you with the tech. Mm-hmm. What were you, what was on your wishlist or checklist before you built it? 30:17 Michelle Gilmore: Yeah, great question. Well asked as well. Very neutral. 30:21 Georgie Healy: See, now I'm back on board. 30:23 Michelle Gilmore: Technology for me is an enabler. So I never, I never start there. What I'm trying to do is think about the behavior or the outcome that I'm trying to get to, and then we as a team now at Juno, but me as an industrial designer and a system and service engineer always would step back and think, what is the desired outcome? And then how do we get there? Now that may be a particular material, a physical product, a digital product, a whole system of various touchpoints across different channels. It may mean AI, machine learning agents, not workflows, not who knows. I think that— It is about prioritization. It is about relativity. It is about trade-offs. There are always trade-offs, and we need to be mature, clear, and fast all at the same time. The great thing about having a co-founder that I've known for 12 years that I trust deeply is that we have that shorthand and we can move really fast. Our head of AI is very good at talking in trade-offs. And in a very positive way, but also in a rapid way. So if we are thinking about a pivot, he might say, well, we can do that, but it's gonna cost X, or there's Y risk, or these are the unknowns. Are we willing to wear those? And then we have a very mature conversation about whether we are or whether we are not, and we keep on moving. And the question is, how much do we want the outcome? And how do we do the least amount to test the next step? We are experimenting and testing all the time. Our, our whole world is experiments. We, we have a system of experiments that are running. There's any number of, at one time there might be 50 experiments running inside this organization. And at different levels, different fidelities, and every single one, if you pointed at it, I can tell you exactly what we're trying to prove or disprove. And then we deploy once we are sure enough. And sure enough means different things for each experiment. Based on a whole bunch of different attributes, but that is in our blood. It's, it's how we think. We can't not think like that. 32:38 Georgie Healy: Yeah, you're not going to get to stage 10 and then be like, ah, shit, we should have pivoted a long time ago. 32:44 Michelle Gilmore: No, it's, it's everything is an experiment. And this morning a customer asked me to do something that could have been ridiculous. And yet actually there was some really interesting fodder in there for us to experiment from. And so, we have turned that into an experiment that will yield incredible learnings across a whole bunch of other experiments. And having that frame of mind means you see things in experiments, and that, that is a superpower. 33:13 Georgie Healy: Can I nerd out for a moment? How many experiments is a good number? 33:17 Michelle Gilmore: I don't think there is a good number. I think that it is relative to the organization, the amount of humans, the systems, the infrastructure, the complexity. The results that you're trying to hit. We are trying to get more customers, keep the customers that we have happy, and make our product better every day. So that is the framework through which we make decisions. Does it make our product better? If so, how? Are we sure? No. Okay. Do we do an experiment? Yes. What's the least amount of time, money, resource we can allocate to that to know X? And is that proportional to X? They, uh, this is the way that we talk at Juno. 33:56 Georgie Healy: Not just let's always have 7 experiments, it doesn't matter if they're good, bad, and different. It's case by case. 34:03 Michelle Gilmore: Yeah, there is no formula that can save you here, I'm afraid. To all those listening at home— 34:09 Georgie Healy: Good and bad. Yep. To all those writing this down. 34:13 Michelle Gilmore: Unfortunately, you have to learn the capability. I'm sorry to say. 34:17 Georgie Healy: Yes. I've got 2 more questions before the rapid fire. Number 1 is, What's the most challenging part about building an AI product that people might not be thinking about? 34:28 Michelle Gilmore: For me, it's about thinking beyond the boundary of human capability and capacity and trusting that what's beyond that horizon could be a beautiful thing. We are not trying to make Juno replicate a human capability here. What we're doing is trying to push a technology to its very boundary to see what this— how big could this be? There's a component of this that is beyond our imagination. And I think that at times we can forget that. And I need to step back and remember that rather than just replicating what I did as a domain expert for 12 years. That is a huge, huge unvan— unfair advantage for us. It is also— Yeah. A curse if not managed well. And I think for the first few months there was a bit of, how did Michelle do it? And now I've started to think what might be beyond, what is beyond human capability? What is beyond human memory? What is beyond translation? What is beyond a one-to-one interview process? What are all of the things that were limiting us before that no longer have to limit us? Who could we hear from that we've never heard from? What organizations could we offer this to that have never done research before? All of those things I think are really exciting and scary depending on how much sleep I've had. 35:56 Georgie Healy: As you're saying that, I'm just like, I'm so glad you have a 4-year-old in your house because the questions they ask, the imagination they have, they don't have the, the 20, 30 years of filter of too hard, to like, they've just got first principles thinking. Like, it's so beautiful. 36:17 Michelle Gilmore: Isn't it such a beautiful thing? And let's hope that we can allow them to keep that. 36:23 Georgie Healy: Yeah. Mine's obsessed with bearded dragons and weird swords that he sees on Mona too and stuff like that. And he's trying to draw them and he's like, when can I have one? I'm like, hopefully never. But, but I like, I like the imagination that he's going to live on an island any day now. 36:42 Michelle Gilmore: I know it's a— it's— yes. Leonardo is quite excited by Maui's tattoos and we've been having— yeah, we've been doing a lot of drawing tattoos on ourselves. So yes, I'm very into this movie franchise. I think if they could just— let's just tone up the magic and tone down The— when Moana in Moana 2— spoiler alert, if you haven't seen Moana 2, you need to— 37:10 Georgie Healy: Spoiler. 37:10 Michelle Gilmore: You need to turn this off, like really turn this off right now. We're going to pause for a second. 37:15 Georgie Healy: Yeah, join us in 5 minutes. 37:16 Michelle Gilmore: Yeah, if your child is listening, do not blame me because I told you. So when Moana at the end of Moana 2 gets the tattoo, well, it's all over. 37:27 Georgie Healy: What do we do now? 37:28 Michelle Gilmore: Our kids are just going to be going and getting tattoos very soon. 37:31 Georgie Healy: Yeah, I, my husband and I don't have any tattoos. We're like the only millennials I think that don't. And I feel like, yeah, we're just gonna be seen as so uncool. 37:43 Michelle Gilmore: Like— There's only so many times I can draw that arm tattoo on my 4-year-old, by the way. I mean, I'm an, I'm an industrial designer. I'm, I'm definitely dusting off my, uh, drawing skills, but I feel like preschool is gonna start asking questions very soon. 37:57 Georgie Healy: And now all I wanted you to do is draw it on me, because I bet it looks awesome. I bet it looks sick. The number of times you've done it, I bet it's really good now. 38:04 Michelle Gilmore: I'll send you a photo of Leonardo's arm later. He'll love it. 38:08 Georgie Healy: Pumped. Okay. I have a quote from Alex Ryan from the Leukemia Foundation that said Juno helped make reliable and will help us make informed optimizations in real time. Hopefully I didn't butcher that. I did quote it from the website, but in a weird way. They think that Juno is just helping them really instantly make decisions through the, the data you're capturing, through the customer insights. That must feel really rewarding. But are there companies that you're like, we don't want to help you, or we can't partner with you, or, you know, because the Leukemia Foundation, that must be incredible to be supporting them. 38:45 Michelle Gilmore: Yes, I love Alex. Um, she's amazing, and she was a very early user of Juno, so shout out to her for trusting us very early on. Before there was a lot of proof points. Obviously Leukemia Foundation is an easy one, right? As you said. I think that as a consultant, it's easy to pick and choose. When you put a product out into the world, you have to let go of control to an extent. There will come a point where you can't filter all of your customers, if particularly in a self-serve world. Does that make me nervous? Yes, absolutely. It does, but I am trusting that process to unfold and I would like to think that on net balance Juno is going to be a very, very positive thing in the world. And there may be some use cases that I don't necessarily agree with as the founder, the co-founder and CEO of this company. And I am learning to be okay with that, although I'm not necessarily entirely comfortable right now, but I feel like I may get there. 39:49 Georgie Healy: As the interviewer, I'm like, let's make her comfortable again, quick. 39:53 Michelle Gilmore: A great friend and a mentor of mine told me really early on, actually, before Juno was funded, when I actually, I asked her this question and she said, sometimes when you're pioneering, you don't get to choose and this industry will be better if you're in it. And when I feel uncomfortable, I go back. To that. I'm paraphrasing her, it was much more poetic. But that's how I like to remember it anyway. And it gets me through those moments. And most of the use cases that come through Juno, I'm incredibly proud of. 40:28 Georgie Healy: You've been so generous to listen to, Michelle. Thank you. I've got 3 quick questions to finish off. 40:33 Michelle Gilmore: Please. 40:34 Georgie Healy: They're kind of the spicier ones. So I've repaid your kindness in heating up that sound booth a little bit. Towards the end, it's like a sauna, and then you can go ice bath it off. 40:46 Michelle Gilmore: Yep. Yeah. 40:47 Georgie Healy: Okay. 40:47 Michelle Gilmore: How— wow, how very founder of me. 40:51 Georgie Healy: Look, you're quoted as saying you're really uncomfortable with certainty. Is that why you're a successful entrepreneur, Michelle? 40:58 Michelle Gilmore: I don't think it hurts. I am constantly pursuing and pushing and wanting us to be in areas that we are not sure about, because I think that's where the magic comes from. So when I become certain, I push again. I don't think it hurts. I don't think it alone is the reason for success, but I think it's probably got an unfair advantage tag. 41:23 Georgie Healy: Absolutely. What keeps you up at night? 41:25 Michelle Gilmore: My 4-year-old. 41:27 Georgie Healy: Well, I know that's a fact. I know, I know that's true. 41:31 Michelle Gilmore: What keeps me up at night at the moment Relating to Juno, I think abundance is a big challenge. There is so much opportunity and the shot clock is on. It's very real when you are a venture-backed company. You need to succeed, you need to succeed very quickly, and we need to have incredible focus and have an abundance mindset at the same time. And those things are challenging to hold. There is a paradox there that you need to get very good at accepting, and that's hard. Um, sometimes I wish that there wasn't a need for sleep because there's just not enough time, and then I'm also very clear about my well-being and my responsibility to Juno. So I struggle to turn off and to let go, but I'm also well aware that that is critical to our success. 42:29 Georgie Healy: Last question. There are numerous podcasts and interviews written about you and your entrepreneurial endeavors. Congrats, by the way. 42:38 Michelle Gilmore: Thank you. 42:38 Georgie Healy: What's the biggest misconception about you, Michelle? 42:42 Michelle Gilmore: I'm gonna turn this one back on you and say, let's do a back and forth. So I don't know what the misconceptions are because I have very little time on my hands, but why don't you give me the perceptions and I'll tell you whether they're true or false. 42:56 Georgie Healy: Okay. I'll give you 3 vibes I get that come off the page, and then you tell me if they're accurate. How about that? 43:05 Michelle Gilmore: I love this. 43:07 Georgie Healy: Number 1 would be that you are a bullbuster. Like, you will just be the first to call people out on their shit. Sorry for swearing. 43:19 Michelle Gilmore: Who are you apologizing to? 43:20 Georgie Healy: I don't know. The, the show. We've said a lot worse. It's the children that are listening. 43:27 Michelle Gilmore: I wonder if I was a man, if that would be a perception of me. I'm not sure, but I am very happy to, um, call— to use your words, call someone out. Yes, absolutely. I am not shy with confrontation. Um, and I would even go so far as to say sometimes, um, I enjoy it. I enjoy debating. I enjoy disagreement. I feel like I am better with difference. You know, I don't wanna surround myself with people that agree with me all the time. I love having a respectful disagreement. I think that often in our culture, we see that as impolite. 44:15 Georgie Healy: Mm. 44:15 Michelle Gilmore: And I wish that we would embrace it more because 9 times out of 10, and don't get me wrong, there is the 1 out of 10, but 9 times out of 10, my spirit and my intent is to learn, and maybe that doesn't come across, but I am certainly not, not afraid of confrontation. I don't know that we busted that myth. I think I would agree with that one. 44:36 Georgie Healy: Yeah. And it probably says some weird stuff about me that that's the impression I get, because obviously speaking to you before the show, 90% of what we talked about were our children and, and, you know, soft, lovely things. And then it's like, but this is the impression I got from the page. And maybe that's a me thing. Maybe that's a me thing, not a you thing or an article thing. Something else I would pick up on is just the attention to detail in this interview and in the ones I've seen. I don't think you just give a cookie-cutter answer, and I don't get the impression you give the same answer every time. I feel like you really listen and you are really answering in that moment, if that makes sense. 45:18 Michelle Gilmore: That is absolutely what I'm trying to do. I hate pre-canned responses. I had to get really comfortable with interviews and I wasn't, so I prefer real connection. I prefer having a real conversation. I prefer saying what I think. I'm really bad at a poker face, so I think that if I was doing formal media training and my team expected me to say something in the same way every time, I would be banned from doing interviews and you wouldn't have seen me so much recently. So hey, I must be doing something right. But I, to use your phrase, cookie cutter, it makes me cringe. 46:00 Georgie Healy: Really? 46:01 Michelle Gilmore: It's just not, it's just not who I am. I believe that A, there's just not another way. For me to do this, but also I don't think success comes from following that approach. It sounds terribly boring. 46:14 Georgie Healy: Honestly, never know how you're going to answer a question, and I love it. I genuinely love it. It's probably why in every interview there's something new every time, because you're not just giving the same answer every time. 46:27 Michelle Gilmore: Yeah. And there's so much time between this interview and when I spoke to Mason for Wild Hearts as an example. There's a lifetime between those two interviews. And so I am a different person and I will continue to demonstrate and show that. And if I stop doing that, let me know. If I stop doing that, I think something has really happened and I've gone off path. Feel free to let me know. 46:54 Georgie Healy: I will, I will. The last thing I pick up on, and this is a true fact, I such a sheer amount of optimism, can-do attitude from everything. I don't imagine that you, uh, get kicked down and stay down for very long. 47:13 Michelle Gilmore: Oh, that's so nice to hear. I don't stay down for long. I do, I do absolutely fall, but I think that I come back reasonably quickly. I believe that falls are helpful. I think failing is useful. I get disillusioned, I get challenged, I get tired. I'm human. I get all of those things happen. Running an early-stage startup is really hard. It's really hard. And we don't talk about that enough. But I am incredibly optimistic about this company, about the future. I believe that we have all of the markers for incredible success, and I don't see why it can't be us. and until that changes, I will continue to show up and do this every day. That doesn't mean that I don't feel deeply within that. 48:02 Georgie Healy: I think 90% of our listeners are founders, so I really appreciate that answer. Uh, it's, it's hard to see, uh, you know, the small percentage of VC-backed and frankly female founders, and you got there on your own merit. But at the same time, it would sometimes be imagined you know, you just see the success, you just see all the interviews, you just see all the Blackboard portfolio company, blah, blah, blah. And I really appreciate your candor. 48:30 Michelle Gilmore: You're so welcome. I think that we, we need to talk about the hard things more and hold the paradox that exists in that something can be very, very hard and it can be amazing and it can be inspiring all at the same time. 48:46 Georgie Healy: I don't know if you've heard of the conspiracy of hawks and doves, but if everyone's a dove and everyone's sharing that experience, you're safe. But as soon as there's a hawk, then you don't want to be a dove anymore. And so if everyone's like, winning, hashtag, like VC-backed, like, then it's like, well, I can't possibly share that I had a down month or a down, you know, quarter or anything like that because I'll look like the loser. 49:14 Michelle Gilmore: I've never felt like I was part of that sort of the fat of the bell curve anyway. So it's, it's somewhere I'm actually more comfortable being a hawk or a dove in your story. I'm not quite sure. I think I'm the hawk. 49:26 Georgie Healy: You're in the group. 49:28 Michelle Gilmore: Yeah. Yeah. So I'm, I'm very comfortable to be outside the norm and the norm makes me a bit uncomfortable actually. And I don't think it's ever worked for me. And at some point I started embracing that because there was no other option. And we are all very complex and we are all going through ups and downs and no one talks about how hard it is. And we should. And I think it's hard. If it's hard for the right reasons, it's amazing, but it's still a challenge and it's a beautiful challenge. But with a problem like this, if it was easy, it'd be kind of scary, right? 50:05 Georgie Healy: If it was easy, I just don't think it would be rewarding. 50:10 Michelle Gilmore: No, it'd be terribly boring and I would have stopped by now. 50:12 Georgie Healy: Yeah. Michelle, this has made my whole week. Thank you so much. 50:17 Michelle Gilmore: You are so welcome. Thank you for having me. 50:19 Georgie Healy: I can't wait to share with the listeners. Before you go though, you've been so generous to us. Is there anything the listeners or I can do? What, what should we know? What can you shout out? 50:31 Michelle Gilmore: Uh, if you are a founder, if you have a product, system, or service, um, if you work in an organization that needs to know what people are thinking, they might be employees, customers, citizens, please use Juno. You can do 30 interviews for free in our unpaid product. It's nowhere near as good as our paid product, but get in there and use it and then hit us up for a paid version. There's a lot of really exciting things happening every day. The paid product is well worth it. I am at [email protected] . You can see Juno at heyjuno.co. So please let me know if I can do anything for anyone listening. 51:12 Georgie Healy: We'll have those links in the show notes, guys. Check out the website because the case studies are pretty insane too. So thank you so much, Michelle. 51:21 Michelle Gilmore: You're so welcome. 51:25 Georgie Healy: Thank you for listening to In the Blink of AI. You can check out the show notes for anything discussed in this week's episode, and we will be back next week. This podcast was produced by Day One with music by Dan Hansen and visual artwork by Sophie Tyrell. If you loved the episode, please tell your mates. And I love AI news. Please share your thoughts and suggestions to [email protected] . No transcript lines match that phrase. Try fewer words.
▶ 
▶ 
▶ 
▶ 
▶ 
▶ 
▶
▶
▶