“PMs who use AI will replace those who don’t”: Google’s AI product lead on the new PM toolkit

Marily Nika 0:00 when all these tools we use every day started coming up, I started figuring out, how can I be an AI-enhanced PM? So how can I be better at my job, have more impact, be more productive?

Claire Vo 0:10 So what you're gonna walk us through is, this has inspired some ideas of, "Wow, if I had a fridge, and a video, and data, what product could I build that would be useful to me as a consumer?" Where would you start as a PM who is working on a smart fridge?

Marily Nika 0:27 Within minutes, I can literally see what the entirety of the world is thinking about. We have opinions in our hands, and we can split and filter out these opinions based on people that want this versus people that don't want this, and we can actually read and have them debate with each other so that we know what it would take to find product-market fit.

Claire Vo 0:47 Okay, you did [chuckles] your market research in three minutes, and your PRD generation in 90 seconds.

Marily Nika 0:53 Now, my job is not done as a PM. The next step is to actually create a prototype so that the people that I will present this idea to will see my vision having flesh. [upbeat music]

Claire Vo 1:09 Welcome back to How I AI. I'm Claire Vo, product leader and AI obsessive, here on a mission to help you build better with these new tools. Today, we have Marily Nika, a AI Product Lead at Google, and AI educator, who's thinking a lot about how AI is changing the art of product management. She's gonna speed-run us through using some of her favorite tools to do market research, build out requirements documents, prototype complicated things, and a few tricks on how to use AI video gen to influence your stakeholders and sell your vision. Let's get to it. This episode is brought to you by WorkOS. AI has already changed how we work. Tools are helping teams write better code, analyze customer data, and even handle support tickets automatically. But there's a catch: these tools only work well when they have deep access to company systems. Your copilot needs to see your entire code base. Your chatbot needs to search across internal docs. And for enterprise buyers, that raises serious security concerns. That's why these apps face intense IT scrutiny from day one. To pass, they need secure authentication, access controls, audit logs, the whole suite of enterprise features. Building all that from scratch, it's a massive lift. That's where WorkOS comes in. WorkOS gives you drop-in APIs for enterprise features, so your app can become enterprise-ready and scale upmarket faster. Think of it like Stripe for enterprise features. OpenAI, Perplexity, and Cursor are already using WorkOS to move faster and meet enterprise demands. Join them and hundreds of other industry leaders at workos.com. Start

Claire Vo 2:54 building today. Well, welcome to How I AI. I am excited because we have a very fun, kind of unique product use case you've been thinking about, and using AI to explore the space of what it means to both be an AI PM and how you can PM AI products. So let's dive in, and talk to us about what's been on your mind in terms of products.

Marily Nika 3:20 Yeah, sounds good, and thank you so much for, for having me. Um, yeah, what I like to say is that I was in AI before it was cool, and I was a pure AI PM, meaning I was working with scientists on training models for speech recognition, and it was great. But since 2017, when all these, you know, tools we use every day started coming up, I started figuring out, you know, how can I be an AI-enhanced PM? So how can I be better at my job, have more impact, be more productive? And I'd be happy to walk you through a couple of my, you know, how I, how I AI. So I'm just so excited to, to be here.

Claire Vo 3:57 Great. And so what fun product are we gonna talk about as an example? I know that you have a lot... Uh, you know, we were just talking before the show, you have lots of jobs. I think this is a little bit of a hypothetical that we're gonna work through, but I think it does show off some of the techniques that you've learned in the course of, of the last couple years in your career. So what are we talking about here?

Marily Nika 4:20 All right. So when I was trying to come up with a cool use case, when, you know, I was planning the, the podcast, I thought about what happened last Sunday. And last Sunday, I was literally walking out from my house, um, with my kids, and I look at the fridge, and I post the, that image that the r- fridge had on LinkedIn, and as you can see, that image went viral. So it said, "These items are expiring soon," and it said my Coca-Cola apparently was 80 days old. First of all, I didn't know Coca-Cola expires. Second of all, I did not know my [chuckles] I did not know my fridge could do that, okay? I know there is a inwards, um, facing camera, but I saw this, and again, I put my PM hat on. What an interesting insight! And as you can see, it says, you know, you can dismiss it, you can snooze. It does say search for recipes, you know, which is not great. But when I posted this, there were 700 likes, h- 100 comments of PMs commenting what they could do by having this information, and 13 reposts. So I'd love to unpack this as a use case and see how we can leverage this suite of amazing AI tools we have and just, uh, you know, see, see in action what it'd be like to be a PM for a smart fridge.

Claire Vo 5:39 This is totally product thinking, because you have a device or something in your house, and you get surprised by a user experience. And I, too, did not realize Coca-Cola expired. I thought you could use it till the end of time. Um, and so what you're gonna walk us through is, this has inspired some ideas of, "Wow, if I had a fridge, and a video, and data-... what products could I build that would be useful to me as a consumer? And you're gonna use that as an example to walk us through some of the really useful, uh, AI-powered PM techniques you've found over the past

Claire Vo 6:15 couple, couple months. So where would you start as a PM who is working on a smart fridge, uh, for busy families who c- who clearly have 80-day-old items in their fridge? [laughing]

Marily Nika 6:28 [laughing] Well, I- we have this magic in our hands, and the very first thing we do as PMs, right, is talk to users and listen to what users want, user research. Now, traditionally, you know, you would either leverage the UXR department of your company, or you would hire an external research agency. But with AI, within minutes, I can literally see what the entirety of the world is thinking about. So I open up Perplexity, and something that not a lot of people know about is that there is this magic filter here. You can obviously search the web if you want, or academic papers, or, you know, SEC filings on finance, but you can only search through discussions and opinions. So this searches through Reddit. So here's what I'm gonna do. I'm, um-- Would families be interested in a smart fridge? So if I ask something like this, it's like I get the entirety of Reddit and people's opinions just mined at once, and I can see references, I can see all the threads, and I can jump into them to actually see the discussion. And I see, you know, use cases that are kind of positive family use cases, common concerns, adoption barriers, um, you know, some sentiment and all these things. Now, this alone is pure gold and a treasure. However, what I love doing is taking things one step further. So here, bear with me, here's what I'm gonna do. Now, create two agents, one that is pro smart-fridge and one that is against smart-fridge. Use everything you read, and have these two agents debate at least, I don't know, like 20 times about it, and give me the minimum set of features I would need in order to convince the against agent. Now, we're diving into Reddit, we're-- we have opinions in our hands, and we can split and filter out these opinions based on people that want this versus people that don't want this, and we can actually read and have them debate with each other so that we know what it would take to find product-market fit. And we're-- we did that in seconds, and it's just such an incredible, incredible tool to have. All right, so here we have the pro agent's arguments, and then we have the against agent's arguments, and you can see the discussion is based on actual, uh, references from, um, from the actual discussion. And then there's a twenty-round debate of pros and cons and all these things, and the best part is the ending, where Perplexity told me, "Okay, here's what you need to do in order to build a fridge, um, that is gonna resonate with families, which is your, your target group." I can't even begin to describe how much time this has saved and how much more likely I am to find product-market fit by leveraging, uh, this technology. So now, naturally, the next step, um... there's something I, I do called tool hopping. So we're gonna hop on to our next tool, but I'd love to see if there is any comments or any questions here.

Claire Vo 9:54 One comment I do have on this is, so often we hear about AI, that it really wants to tell you your ideas are great, and so I think if you just said, "Do families want a smart fridge?" without any additional prompting, it would probably bias towards telling you that families do want a smart fridge. Maybe with some caveats, but it tends to be, uh, pretty agreeable when it comes to, um, [lips smack] uh, interacting on ideas. And so I love this idea that I wanna make sure people don't miss, which is actually prompting the AI to create two different personas in its analysis. Create a persona that's very for your idea or implementation or design or code. Create an agent that's very against your idea. Again, it doesn't have to be an idea, it could be, "Are you against the implementation of this code?" Again, a really good use case. Have them do a hypothetical or artificial debate on the topic, and then deconstruct that debate in a way that allows you to identify what would be the convincing arguments, capabilities, or implementation that might work. And so I think this is just a really great, um, little hack on getting higher quality, um, more critical thinking out of these LLMs that might be biased a little bit to agree with your ideas and tell you they're good.

Marily Nika 11:17 Yeah, I completely agree.

Marily Nika 11:19 And the most important part is I told it what format I need this response in, and the format I need it in is features, 'cause I'm a PM, and because features is the input I need for the next tool I will demo. So now what I'm gonna do is I am gonna copy the, the minimum set of features that Perplexity thinks I need, and then I'm gonna go to ChatGPT in order to create, um, a PRD. Now, I have created an AI Product GPT, as I like to call it, that has a template of my favorite PRD, and essentially, I prompt it with specific features, and it outputs a PRD in the style that I speak in my own voice. So the next step is I come here, and it says, "Ask anything," and I say, "Generate a PRD about a smart fridge-"... that has these features. And essentially, I am pasting the output for, from Perplexity. And now there we go. Immediately, it just generates and say, "Okay, well, the PRD is an AI-enhanced smart fridge with local-first intelligence." This is what it includes. Here's the problem statement, what users, you know, what privacy concerns consumers, DIY open source enthusiasts, summary, and feature analysis of architecture, key features. Of course, it's, there's prioritization and all these good things at the same time. Now, this is not gonna be perfect immediately. However, it gives me an incredible headstart in order to just have my strategic thinking in the, in my head, and spend, spend way less time in making a copy of the PRD and changing the title. So now I'm an AI-enhanced, um, PM.

Claire Vo 13:15 The other thing I wanna call out for, uh, folks that are looking at this is, this is a GPT specifically to create PRDs around AI-powered products, um, I noticed, or at least that's what it looks like. And so I do think that, um, you can create sort of general purpose document generators, but the more you can do a combination of structure and content in terms of your prompting,

Claire Vo 13:40 I think the higher quality results you're going to get. And so, okay, you did [chuckles] your market research in three minutes, and your PRD generation in 90 seconds, but, you know, it's a long way to the factory floor and actually building this thing. And so, you know, once you have this, how do you like to think about designing great products with, with AI? Or what's the next step for you?

Marily Nika 14:06 Yes. It's really interesting how this flow has changed, right? It used to be, "Hey, I'm a PM. I come up with an idea. I'd create a PRD. I debate with a PRD and collaborate with my scientists and my engineers until we're done, and then we create a little MVP." Now, my job is not done as a PM. Now that I have my PRD, the next step is I need to build. So I'm putting my AI builder hat, as I like to say it, and the next step is to actually create a prototype, so that the people that I will present this idea to will see my vision having flesh. So the next step is to build, to prototype. There is obviously many tools, um, that we can use. One of my favorite tools is, of course, um, v0, Lovable. Google AI Studio has prototyping as well now. Um, I will use v0. It's one of the tools that I have had, you know, the best kind of outputs with. And what I did is, I copied the output from my custom GPT, and then I'm about to paste it here. So the prompt I will use here will be, "Create the UI of a smart fridge, given this PRD as an input." And I will just paste the whole thing. Because it's a very long document, it takes it in as a note, and then I just do this, and then we just wait.

Claire Vo 15:36 And do you feel like these PRDs... Obviously, I do, um, you know, we have an integration with, with v0 and all these five coding tools, because we really believe that a PRD is a higher quality input to this. You know, have you tried this, you know, prompting directly versus just pasting in the PRD, and feel like that extra step really does help with the prototyping quality?

Marily Nika 15:58 I do a lot of experimentation-

Claire Vo 16:01 Yeah

Marily Nika 16:01 ... and I experimented because I wanted to move fast at some point.

Claire Vo 16:06 Yep.

Marily Nika 16:06 And I realized that the number of tweaks I had to do later on really got me overwhelmed and tired, versus me starting, spending some more time in order to make sure the input is very, very good. Um, it really

Marily Nika 16:20 pays off.

Claire Vo 16:21 Yep. And, you know, how do you think about where this might fit in the product development workflow? It seems like for you, a prototype is really a communication tool. So it's a tool by which, you know, if somebody read the PRD, you're like: Yeah, yeah, it's a, like, a locally intelligent smart fridge. Sure, whatever. But once you can see, touch, feel it, it becomes much more compelling. And so, you know, I'm curious, do you think this is a influence tactic that PMs are, are underestimating? Is it really to design the right thing? Where do you think PMs can use this kind of prototype?

Marily Nika 17:00 The perfect place to use this is at the product review. So I find that a lot of product reviews happen more often now with AI, because things move fast, because some people are, are really not convinced with AI yet. So if you enter a product review that's just about, "Hey, here's what I wanna do-

Claire Vo 17:20 Mm-hmm.

Marily Nika 17:20 ... Are you in or are you out? Here's where you need to invest your resources in," and they can get a link, and on their laptop, actually touch it, feel it, see it, move around, you communicate your vision and your passion more effectively, versus, "Here's a PRD that I turned into a deck for you, so that I can walk you through it." It's just-

Claire Vo 17:41 Yep

Marily Nika 17:41 ... so much more. It adds credibility, it adds value. It's just so, so, so much more effective.

Claire Vo 17:47 Amazing. Okay, and now we have an open fridge dashboard. I have to call out for people who are not watching [chuckles] on YouTube and are listening, my favorite part of this UI is the door status widget that tells you if the door is open or closed. [chuckles]

Marily Nika 18:06 It's amazing. My kids go to the freezer to get ice cream, and they don't close it. So I-

Claire Vo 18:10 Mine too. Mine too, it beeps.

Marily Nika 18:14 If you-

Claire Vo 18:14 You know, you, you and I have the same life. They, they close it, like, but it's this much, this much open.

Marily Nika 18:20 ... Yeah. Yeah, yeah, exactly. And the light is on, and they're in such a rush to eat the ice cream that they don't care. Um, okay. So yeah, we can see here fridge temperatures, door status, um, power usage was a good one. You know, the, the temperature that you can change, recent activity. Oh, what time it opened and closed, that's also good to call them out and say, "Did you have an ice cream?" [chuckles] Um, temperature completed, safety monitoring. It's all these controls, and you can see that this tool took the PRD as an input, and it added features on top of it. It guessed what the UI needs to look like. Um, so imagine your leadership or even your cross-functional partners to be able to experience your product idea. Imagine just how much more effective the meeting would be.

Claire Vo 19:08 Yeah. One thing I have to, have to call out is because you said that privacy and local intelligence was so important in your PRD based on your research, and I think that's a lot of people saying, "I don't want [chuckles] my, my smart fridge monitoring me. I don't want a cloud-connected smart fridge. I don't want advertising. I don't want monitoring." If you go up to, I think, hardware, I'm pretty sure... or diagnostics, I'm pretty sure your smart fridge is shipping with a GPU, which I find very, [chuckles] very funny.

Marily Nika 19:41 [laughing] Yeah. Yeah.

Claire Vo 19:43 You know? Y- maybe you can also run your coding models locally through, through your fridge, you know?

Marily Nika 19:52 Exactly. I love it. Little storage as well. Yeah, this is, this is great. Um, but again, this is a wonderful hundred percent model processing. Look at this, it has a little badge that shows you so that you can feel safe. Yeah, this is great. I would buy this fridge. I think this is, this is really phenomenal.

Claire Vo 20:07 This is very much like a Silicon Valley fridge right here. It's in dark mode. It's, uh, it's got system diagnostics. As long as you can, like, root into your, your fridge, people are gonna be really happy. [chuckles] You've seen the doom and gloom headlines: "AI is coming for your job." But the reality is a little bit brighter. In Miro's latest survey, 76% of people say AI can boost their work. It's just that 54% still don't know when to use it. As a product leader and a solo founder, I live or die by how fast I can turn fuzzy ideas into crisp value propositions, roadmaps, and launch plans. That's why I love Miro's Innovation Workspace. It drops an AI copilot inside the canvas, so stickies, screenshots, and brainstorm bullets can become usable diagrams, product briefs, and even prototypes in minutes. Your team can dive in, riff, and iterate, and because the board feels like a digital playground, everyone has fun while you cut cycle time by a third. Miro lets humans and AI play to their strengths so that great ideas ship faster and happier. Help your teams get great done with Miro. Check out miro.com to find out how. That's M-I-R-O

Claire Vo 21:30 dot com.

Marily Nika 21:31 So we have this, and it's great. It's an amazing first iteration. Obviously, for a real thing, um, I would ask at least five to 10 follow-ups. Usually, I add some images as well. Maybe I add an animation, and at that point, things are great, but this is not enough. Um, I add one more step to my workflow, which is people really want to experience what this would look like from the user perspective. And nowadays, we have all these wonderful tools that go from text to video. And so the next step I would like to do is actually go to Flow. Flow is this, um, app by Google Labs, and then if you click on Create with Flow, you can type New Project. And then here, you can literally just go back to Perplexity if you want, or get your PRD and say, "Create a promotional clip for two- for a couple using a smart fridge that has these features." And I can paste, and then you can choose what, um, model you wanna have. So, for example, what, you know, aspect ratio is, whether it's gonna be fast, if it's gonna have audio. I think I'm gonna go with high quality-

Claire Vo 22:55 Yep

Marily Nika 22:55 ... just to see what can come out of it, and then hit this. And then we have to wait.

Claire Vo 22:59 Yeah, with movie magic, we can make people wait a little less, uh, as this generates. But I love Veo. It's such a great model.

Marily Nika 23:09 It's such an amazing model. I've been using Sora, too, at the same time. I have the, the app on my phone.

Claire Vo 23:15 Yeah.

Marily Nika 23:16 And obviously, you know, uh, that generates in seconds, which is amazing.

Claire Vo 23:20 Yeah.

Marily Nika 23:20 And you get the cameos and all that stuff, but they, they roll it out as part of a social network. I don't know if you've tried it.

Claire Vo 23:28 Yeah, I've not abused it.

Marily Nika 23:29 It's interesting. It's creepy at the same time-

Claire Vo 23:31 [chuckles]

Marily Nika 23:31 ... but mostly interesting. Um, [chuckles] but yeah-

Claire Vo 23:36 What if I-

Marily Nika 23:36 I had my AirPods, and I took a photo of it, and I told Sora, I said, "Hey, tag Marily using these and make the ad an advert." And it zooms in, and it opens it up on the sky, and has me saying, "And I go..." Amazing times. Amazing times. Absolutely.

Claire Vo 23:53 Yeah, my, my kids find it very funny when I make, um, Sora cameos of myself being like, a, like doing kung fu fighting or doing something really funny. So it's quite, it's quite entertaining. I'm... You know, I think the, the social network aspect of it is a distribution, um, method for the underlying API models. So I think it just shows sort of power users what's possible with it, and probably then people get excited about commercial applications. But-... I think it's pretty, pretty cool. Look at this, door ajar.

Marily Nika 24:32 Right.

Claire Vo 24:32 They're really worried about that door being open. [laughing]

Marily Nika 24:37 Also, the screen is inside the refrigerator in one of the [laughing] What is going on? Okay, um, yeah, let's play. I hope the video- I hope the audio is gonna go through. Okay, let's see. Number one.

Speaker 24:47 Get instant local notifications, like when the door is ajar, ensuring your food stays fresh and secure. Get inst- [beep]

Marily Nika 24:57 [laughing] Number one.

Claire Vo 24:59 Well, so number one, again, for people who are not on YouTube and enjoying this high-quality AI-generated video, there is a screen on the refrigerator. This l- very attractive, well-dressed couple opens up the refrigerator, and there's another screen inside. [laughing]

Marily Nika 25:19 Amazing. All right, number two, let's see.

Claire Vo 25:21 Okay. [upbeat music]

Marily Nika 25:30 [laughing] Oh, two screens-

Claire Vo 25:32 Oh!

Marily Nika 25:32 ... again. That's, uh, interesting. That's interesting. Um, there's no voice.

Claire Vo 25:37 Yeah, so this one-

Marily Nika 25:37 There's cool music on this one. [chuckles]

Claire Vo 25:39 Yeah, cool music. This one, there is a screen inside. Basically, the only thing you can see inside the refrigerator is a screen that says, "The door is open."

Marily Nika 25:48 [chuckles]

Claire Vo 25:48 And then when you close the door, the screen shows up again. [laughing] Let's keep... Fingers crossed for number three being the winner.

Marily Nika 25:57 All right, let's see.

Speaker 25:58 It says we have enough for the lasagna recipe right here.

Speaker 26:01 Perfect. And I love that it's all stored locally, no cloud needed.

Speaker 26:06 It says we have enough-

Marily Nika 26:08 Whoa! I think we have a winner, Claire.

Claire Vo 26:10 I, I think so, except if you watch it again, I think the man has disappeared-

Marily Nika 26:15 Wait, really?

Claire Vo 26:15 ... and gets replaced by a lady. Watch. Watch this.

Marily Nika 26:18 I like this.

Speaker 26:18 He says we have enough-

Claire Vo 26:18 She appears, then his girlfriend shows up

Speaker 26:19 ... for the lasagna recipe right here. Perfect. And I love that it's all stored locally, no cloud needed.

Marily Nika 26:25 [laughing] It did disappear. Well, we don't need him anymore. We're good. Um-

Claire Vo 26:32 You know what? But it is vision setting. [laughing]

Marily Nika 26:36 That's interesting. Um, you know what I may do? I may actually generate myself, um-

Claire Vo 26:41 A Sora

Marily Nika 26:42 ... opening in Sora. I think that would be a very interesting.

Claire Vo 26:44 I think we should do. Okay, so we're gonna contrast and compare. How would I improve this? I think, you know, my feedback on prompting here is it's really over-indexed on this door is open use case. So I would probably pull that from the prompt, and I would also maybe, as a good product manager, one of my favorite parts to put in a, in a PRD, this is in our default chat PRD, PRD is actually what's called a narrative, which is a story about an end user getting value from the product or being successful with the product. And I like to write those stories as part of PRDs because I think it, it forces you to take a very user-centric lens on a list of features. And so I might generate a narrative here and try to, like, pick a, pick a hero story from the narrative, as opposed to over-indexing so much on the features. Um, but the quality is amazing.

Marily Nika 27:41 The quality is amazing, and this took, what? 30 minutes. If we didn't comment-

Claire Vo 27:45 Yeah

Marily Nika 27:45 ... and laugh, it would have been, like, a 20-minute workflow. So it's just-

Claire Vo 27:49 Yeah, yeah.

Marily Nika 27:50 It's phenomenal.

Claire Vo 27:50 Exactly.

Marily Nika 27:51 Amazing.

Claire Vo 27:51 Okay, so what we're doing now, just for folks that, um, again, are just listening and not on video, we are pulling up Sora. You have created a cameo of yourself, which is basically like an AI avatar of your face. Um, warning to people who are using Soras to create cameos, they really force you to continue to dress how you're dressed in your cameo. So I was wearing, like, very printed shirt when I did my cameo, first cameo recording, and now all my clothes in every cameo are crazy. Um, but you're generating OpenAI Sora video. H- did you use the exact same prompt, or did you change it a little bit?

Marily Nika 28:34 Very similar prompt, but I decided to add someone else in my cameo-

Claire Vo 28:38 Oh

Marily Nika 28:38 ... and I looked at who enables, who can do that, and I'm adding Mark Cuban to our-

Claire Vo 28:43 Oh, perfect

Marily Nika 28:44 ... video. A smart-fridge to Cuban. [laughing]

Claire Vo 28:45 I think we're gonna get, I think we're gonna get a Cost Plus Drugs advertisement, because also, I don't know if you know this, uh, folks that are listening about Sora, when you create your cameo, you can ship it with custom instructions. So you can say, "Never generate this," or, "Always generate that." I say, "Always make me look good" [laughing] on my cameo. And I'm pretty sure Mark Cuban put in his, like, in every video, make sure to put a Cost Plus Drugs ad in it. So I think we're gonna get a smart, smart-fridge video with a embedded advertisement for-

Marily Nika 29:25 That-

Claire Vo 29:25 ... Mark Cuban's company.

Marily Nika 29:27 You know, I had added, um, make me have, uh, pet Pegasus with me at all times. Everything I do, there is, like, a pet of Pegasus with me.

Claire Vo 29:35 [laughing]

Marily Nika 29:35 Um, but yeah, let's see. Let's see how it goes. Um, all right, my cameo is ready. First time I'm watching it. It's me and Mark Cuban opening a smart fridge. Um, all right. [chuckles] Let's see, let's see how it goes. I don't know what we're gonna... we're about to see.

Speaker 29:50 So this is the smart fridge you told me about.

Speaker 29:52 Yeah. Watch what happens when I wake it up.

Speaker 29:53 It looks like a normal fridge.

Speaker 29:54 Until it starts thinking.

Speaker 29:55 Oh, it already knows what's inside.

Speaker 29:57 Mediterranean quinoa bowl, Greek salad, pasta primavera.

Marily Nika 29:59 I mean, first of all, Greek salad, that was interesting. I'm Greek. I think it's-

Claire Vo 30:03 I know, [chuckles] it must have known!

Marily Nika 30:05 Make it so. [laughing] But I mean, pretty impressive, right?

Claire Vo 30:10 Look-

Marily Nika 30:10 It came up with a little dialogue, and it's added the two people. I mean, this is magic. [chuckles]

Claire Vo 30:16 Look,

Claire Vo 30:17 PMs, what I want you- we've, we've had some giggles here, but let's just take a step back and talk about what we've done in basically 20 minutes. We have done full end-to-end-... user research using the wealth of information available in Reddit. We've come up with a highly defensible set of features, which we turned into a PRD. We turned that PRD into a prototype that then you could take into a product review and click through, and then just, you know, we- we're gonna work on our prompting a little bit and our cameos, but just imagine pairing that with some sort of video-generated visual or set of visuals that really just hammer home the value proposition of the thing that you're trying to build. And that package, which we just talked about, we just did, can actually be built in, like, 15 to 20 minutes [lips smack] without the chitchat, and it's just something that is- was never, ever, ever possible before. And so what I think you're showing is the way we do product has changed, the way we convince and influence our internal stakeholders has changed, the way we communicate both product value and functionality has changed, and if you use a combination of these tools, even if you hit some silly roadblocks, you've just, like, totally flipped product management and the process on its head and made it just much more compelling.

Marily Nika 31:41 It's just incredible how the role has changed of being a PM. You've gotta be AI-enhanced. And it's not like AI is taking over our role. If anything, PMs that use AI are the ones that are gonna take over the role of people who don't use AI. So it's just so important to explore and adapt and see which tools fit to your workflow, 'cause, you know, my choices might not be the same as other people's choices, right? Um, but I'm excited I get to, to share them with

Marily Nika 32:06 you.

Claire Vo 32:06 Great. So in addition to this, um, kinda new world AI PM stuff, you have one other kind of power use case of another tool that we, we've seen occasionally on the show, but not too frequently, and I think your use case is really interesting. So do you wanna spin that up?

Marily Nika 32:22 Yeah, sounds good. So outside of my day job as an AI PM, I'm an AI educator and the founder of the AI Product Academy, and I have a bootcamp. It's AI Product Management Bootcamp and Certification, and people get to join us for six weeks and build their own product end-to-end. They actually launch a product by having an en- an engineer attached to them. Now, we have a demo day, and that's the best part of the bootcamp. Everyone comes in, presents, they earn prizes, they pitch to VCs, and I wanted to add AI as a judge, and I thought about how I could do that long and hard, and I ended up using NotebookLM as a judge, and I wanna show you what this looks like 'cause it was super cool and super interesting. So on June 13, we had a demo day, and what I did is I recorded the audio of every single pitch, and I uploaded these separate files here. So this notebook actually has all these presentations, and then I went to the Audio overview, and I added specific instructions to focus on for these AI hosts. And I said, "Okay, this is the demo day of Marilee's AI Product Management Bootcamp, and I want you to select the top three apps based on these criteria. Number one, you know, how innovative it was; number two, impact; and number three, storytelling. And I want you to announce these three winners," and then I clicked Generate. And it's phenomenal. Now it's actually generating the audio overview, and you can imagine my bootcamp at the end is, like, 200 people really waiting to see what AI is going to say.

Claire Vo 34:21 Mm.

Marily Nika 34:21 And we actually did a test with... Uh, well, we have live judges and then real people and then the AI, and AI really chooses the best demos indeed. So it does listen to all, um, the audios, and it's just such an amazing use case. Everyone loves it.

Claire Vo 34:40 I love this so much, and it just opens my mind to a couple other use cases that, um, I've seen very commonly. So demo days, both for courses and for hack days or hack weeks, could be really useful. I have been in so many sales enablement or sales kickoff meetings where AEs do pitch competitions, and what a fun way to add sort of like a third-party analysis to pitch competitions and make those kind of fun. And so what I love about NotebookLM is you can do audio and multiple file in, and then audio and multiple file out, and I think that's really, really fun.

Marily Nika 35:20 It's great, and now you can actually do video overviews as well. I haven't tried that for my bootcamp, but that could be interesting as well. [chuckles] All right. So now there is a file that got generated on the right. We can either play and hear the whole podcast, presenting all the ideas and walking us through them, or we can do a super cool thing where you can have this kind of interactive mode, which is like an actual live radio podcast where you call in. So let's do that 'cause I really like this.

Speaker 35:49 Welcome back to the Deep Dive. Okay, so you sent over this, uh... Oh, our listener's got something to share. Let's hear it.

Marily Nika 35:55 "Hello, we're actually in a rush. Can you please announce the three winners right now?"

Speaker 36:01 Oh, absolutely. We hear you.

Speaker 36:03 We can definitely cut to the chase, but we do wanna give just a little context first.

Speaker 36:06 That's fair. We were just setting up the criteria for judging these fantastic projects.

Speaker 36:10 And we are totally ready to announce the three winners.

Speaker 36:13 The quality was so high across the board.

Speaker 36:15 These projects are tackling some huge problems, like health anxiety and life administration.

Speaker 36:20 We grouped the pain points into three big areas: health and wellness pain-

Speaker 36:24 ... the grind, which is time and life admin pain-

Speaker 36:26 And the third one is the knowledge and empowerment gap.

Speaker 36:29 Things like struggling with learning or bad leadership.

Speaker 36:31 We judged them on three things: in- Hey, yeah, what's up?

Marily Nika 36:35 ... Just tell me the winners. You're taking too long.

Claire Vo 36:39 Oh, we completely understand the need for speed. We got ahead of ourselves.

Speaker 36:42 You are absolutely right. We apologize for the delay in getting to the point.

Claire Vo 36:45 We've got the three winners locked in and ready to go.

Speaker 36:47 You wanted the winners, so here they are, no more preamble.

Claire Vo 36:49 The three top projects are Endopath, Study Buddy, and Front Desk.

Speaker 36:54 Those three really stood out, addressing severe-

Marily Nika 36:56 Woo!

Speaker 36:57 - quantified pain points.

Marily Nika 36:59 All right. So this is incredible, because it's not like this one agent we're used to talking to. It's, you're talking to them, and then the other person who's responding, it's just, you really feel they're real, and it's just magical. But yeah, that's my last use case on what I do on my bootcamp, and I think it's the most... my favorite thing to do on NotebookLM, by far.

Claire Vo 37:23 This was really good. I actually didn't know they had that feature, so I, again, can think of so many, like, training use cases, lots of things. Um, I love those little, little podcast hosts. They've got such a vibe, and this was such a great use case.

Claire Vo 37:38 Okay, I'm gonna ask you our one lightning round question that we ask everybody, and then I'm gonna get you out of here, 'cause you're a busy, busy lady. So we've seen you... You're a tool switcher, so I know that you switch tools, but I have a question. When one of those tools is, you know, let's say Veo. Veo's giving you refrigerators with screens inside, with screens outside, with screens on your face, screens on your freezer, um, screens on your screens. When it's really not getting you what you want, what is your prompting technique? What do you do?

Marily Nika 38:11 The best thing you can do is just kill that instance and start over, and use gen AI to help you write the best prompt. So I stop trying to guess it myself, and I use AI to help me with it, and the longer the prompt is, the better it's gonna be and the less likely to have more iterations. So yeah, use AI on AI.

Claire Vo 38:31 Okay, kill it, use AI, and make it detailed.

Marily Nika 38:35 That's it.

Claire Vo 38:36 I love it. Well, thank you so much for joining us. Where can we find you, and then how can we be helpful?

Marily Nika 38:41 Please find me on LinkedIn. I post content on AI product management and AI building, but you can be helpful by asking people to join my bootcamp. I have a cohort coming up in December, and as I said, people get three certifications, and they get to actually build a full product and start their own company, because I assign engineers to every team, and you can build the product end to end. So it's this new tool that's coming out that is not gonna need us to do tool hopping anymore. You do everything on this one tool, and it's amazing, and I can't wait to have people and see what comes out of it and what apps they create.

Claire Vo 39:21 Awesome, and I believe you can find that course... I think we're gonna get you listed on Lenny's List if you're not already, and it'll be at maven.com/lenny. So check out her course. It's gonna be amazing for any PM that wants to do this amazing work very, very fast. Thank you for joining us on How I AI.

Marily Nika 39:39 Thank you so much for having me. [upbeat music]

Claire Vo 39:42 Thanks so much for watching. If you enjoyed the show, please like and subscribe here on YouTube, or even better, leave us a comment with your thoughts. You can also find this podcast on Apple Podcasts, Spotify, or your favorite podcast app. Please consider leaving us a rating and review, which will help others find the show. You can see all our episodes and learn more about the show at howiai pod.com. See you next time!