Google AI PM Reveals the Tools 99% of Product Managers Don’t Use

1. 0:00 – 1:52

   Intro

   1. MNMarily Nika0:00

      Google launched AI prototyping on AI Studio. It's amazing. Under 10 minutes, you have a fully functional app.

   2. AGAakash Gupta0:07

      Google and Meta are leading the AI race, and she's the PM behind it. Marily Nika has been an AIPM for 11 years. In today's episode, she's showing you exactly how to use AI well as a PM. NotebookLM, Opal, AI Studio, you're not using these tools, but they can give you superpowers. Let's say they want to get your job. What's the 18-month roadmap to landing an AIPM job at a company like Google?

   3. MNMarily Nika0:30

      I tell PMs that they need to be like crabs. What does that mean? That means that you need to move adjacent to what you've been doing.

   4. AGAakash Gupta0:39

      When should you be using an AI prototype?

   5. MNMarily Nika0:41

      It used to be the case: idea, PRD, comments on the PRD. Whereas now, the workflow is you have an idea, you don't need to reach out to anyone yet. So what I do is I prototype and then bring my scientist and engineer.

   6. AGAakash Gupta0:53

      Should they be using GPTs like this?

   7. MNMarily Nika0:55

      If you want one takeaway from this podcast is don't be embarrassed to admit that you use AI.

   8. AGAakash Gupta1:01

      What are the top tools you use as an AIPM? Before we get into today's episode, if you can do me a quick favor and check if you have a following on Apple and Spotify podcasts and subscribed on YouTube, these are free actions you can take that really help the show grow. And if you become an annual subscriber to my newsletter, did you know that you get access to over $28,000 of premium products? That's right. Mobbin, Arise, Relay.app, Dovetail, Linear, Magic Patterns, DeepSky, Reforge Build, and Descript. They are all free for an entire year if you become an annual subscriber to my newsletter. So go take advantage at bundle.aakashg.com. And now into today's episode.

2. 1:52 – 6:41

   Why These 6 Tools Matter

   1. AGAakash Gupta1:52

      Marily Nika and I filmed my seventh most popular episode a year ago. Marily broke down everything on how to become an AIPM because she's done it. She's been there. She is an AIPM at Google. And today, she's here for a new episode to drop all of the AI tools she uses as an AIPM, how you can supercharge yourself as an AIPM with AI using the latest Google tools. For instance, are you using NotebookLM? Are you using Opal? Are you using Google AI Studio to their fullest power? Marily is, and today she's dropping all the expert tips and tricks. She's made the mistakes, so you don't have to.

   2. MNMarily Nika2:37

      [laughs]

   3. AGAakash Gupta2:37

      She's gonna drop the workflows that you can use to use these tools like an expert from day one. Marily, welcome to the podcast.

   4. MNMarily Nika2:45

      Hi. Thank you so much for having me again, Aakash.

   5. AGAakash Gupta2:48

      Yeah. It feels like every couple years we have to get back together for a collaboration. We actually collaborated even before the podcast on the newsletter a year before that, so we have been fast friends for a while. So nice to see you. Today, I wanna break down for everybody all these AI tools you're using.

   6. MNMarily Nika3:05

      Yes. Excellent. Let's do it.

   7. AGAakash Gupta3:08

      So walk us through this. What are the top tools you use as an AIPM?

   8. MNMarily Nika3:13

      All right. So there's something I call tool hopping, and here's the tools I will use several times every single day. All right. Number one, and this is in no particular order, um, Google launched AI prototyping on AI Studio, and it's amazing. It, it's unbelievable. It makes tweaking the idea that you have in your mind and the vision super, super quick. So iteration is wonderful on it. Number two, Google Labs launched this product. It's called Opal, and it's for AI mini apps. I'll show you a couple of examples on what I've been doing there. But instead of going on a playground to just try out the new different models that are out there, you can literally type in using English and see a mini app emerge. It's very fun. You- under 10 minutes, you have a fully functional little app that can use, like, NanoBanana and all the cool, um, models, and it's just super great. All right, next up, and, you know, there's different kind of hats I'm wearing when I'm using them. So I have my Google job, as you say, where, you know, I'm an AIPM and I wanna be more productive, so that's where AI prototyping and Opal comes in. Now, when I do my second job, which is, I mean, yeah, I founded the AI Product Academy and I teach on my bootcamp, that's when I use NotebookLM. And I have a super interesting demo to show you, which is specifically for my demo day. I wanna have a judge that's an AI judge, and I bring in NotebookLM. I load up all the presentations from my students, and then it assesses what people have created and chooses the winner, and the winner really gets swag and, and prizes just because AI chose it. Um, now back to my Google hat. Uh, when I wanna build a product, of course, first thing you'll need to do is talk to the users, talk to the people. I discovered this super cool feature on Perplexity that allows you to search only through social. It searches only through Reddit. And I get to have kind of the, the magic, the, the power of the people voice on my hands, and I mold it, and it helps me really become an amazing product leader 'cause I know exactly what's likely to find product market fit. Moreover, as PMs, we create artifacts, right? I create most of my artifacts on ChatGPT or using Gemini's Gem. Now, I have some custom GPTs that I've customized so, so, so much that I just cannot move to any other platform because it knows my voice and myself so well after, I wanna say years at this point. Well, yeah, after years of using them. And-Last but not least, this is probably the only meeting I don't have my Fireflies note taker. And, um, I use Fireflies to take notes, to be more productive, and make sure I will follow up on the things that are assigned to me.

   9. AGAakash Gupta6:15

      I love your tool stack. I love how it's not a trillion tools. It's very simple. You're at Google, so a lot of it is on the Google stack, which makes sense, right? A lot of PMs listening, they only have access to certain tools at their work. You have to make do with the work tools you have against these key use cases which she's outlined. And I think one of the ones that most people don't know about for Google is that AI prototyping use case. So can you show us that and how to do it well?

3. 6:41 – 11:05

   Google AI Studio for Prototyping

   1. MNMarily Nika6:41

      Yes. Sounds good. All right. The UI is familiar to users at this point. So as you can see, there is kind of a area where you describe your idea. You can select the model you wanna use. It defaults on 2.5 Pro, which is great, or Flash if you wanna go faster. You can upload, um, files, or you can actually do speech to text, which I've started to use more and more. Now, there's also a gallery here where you can see what other people are building, which is incredible if you wanna get inspired. I know we all have so many ideas, but then by the time you're ready to use the tool and build, you kind of forget a lot of the things. So it's, it's fun to, to see what other people are building. There is one thing I absolutely love here. I think they had, um, they were playing, um, ugh, let me see. Let me see. Sorry, let me pull it up and then I'm gonna continue. Um, doo, doo, doo, doo, doo, doo. I don't know if I can search for-

   2. AGAakash Gupta7:41

      Take your time.

   3. MNMarily Nika7:44

      Yes. Okay. I'm gonna continue now. Okay. So here's a super cool example. Someone created Dreamy 1995, which replicates Windows 95. And you can see you have your desktop, and you have kind of Chrome, but you also have Gem Sweeper. I don't know if you ever played this, [laughs] Aakash. I definitely played it all the time. But it's great that you can use your natural language to actually see and visualize something that brings nostalgia or something that has been sitting in the back of your head and you're now ready to build it. Now, one of the reasons I love using AI Studio is because of NanoBanana. NanoBanana is the best text-to-image model. It's just so, so good. So I just love, um, selecting NanoBanana, uh, here at the top so that I can give a heads-up and say, "Hey, I will be, um, using NanoBanana." All right. So should we come up with a random idea on what we can create?

   4. AGAakash Gupta8:49

      Um, I guess, like, everybody really wants to create at-scale ads these days, right? Because that's the name of the game with paid advertising is trying lots of different creative.

   5. MNMarily Nika9:00

      Oh, I love it. Yes. I love it. Um, so one of the ideas I had that I really wanna build and flesh out right now is to create, like, an easy way to generate content for LinkedIn, but not any content, content that could have yourself, things to brag about. And what I mean, I saw my friend the other day posted six photos of himself. One was him on his TED Talk. The other one was him doing some hike. The other one was him traveling in Greece. So it was a collage of all these kind of bucket list items. And I was thinking, "You know what? With NanoBanana, you can literally upload your photo and just have it generate a collage of yourself in all these cool places." So maybe we can try and see what it would be like to build an idea where the only input you have is yourself, like your photo, and then the output's gonna be a collage of you in all these super cool bucket list items so that it's kind of LinkedIn ready. I don't know. What do you think about that?

   6. AGAakash Gupta10:04

      Love it. Let's do it.

   7. MNMarily Nika10:05

      All right. So I need to write my prompt. Okay, so I wanna create a collage generator for LinkedIn posts. The input will be the photo from the user as well as, like, six bucket list items, like meeting, I don't know, Taylor Swift.

   8. AGAakash Gupta10:31

      [laughs]

   9. MNMarily Nika10:32

      TED Talk. Um, what else is super cool? Being in the most beautiful island and so on. So we need an app that helps us do this. All right. So that's what I did, and I'm just gonna build, and I wanna see what comes out of it. So as you can see, I already specified that this needs to be a NanoBanana powered app because I know that the input's gonna be a photo, and the output's gonna be a collage of photos, so you definitely need to leverage this.

4. 11:05 – 12:53

   Ads

   1. AGAakash Gupta11:05

      Today's episode is brought to you by Jira Product Discovery. If you're like most product managers, you're probably in Jira tracking tickets and managing the backlog. But what about everything that happens before delivery? Jira Product Discovery helps you move your discovery, prioritization, and even roadmapping work out of spreadsheets and into a purpose-built tool designed for product teams. Capture insights, prioritize what matters, and create roadmaps you can easily tailor for any audience. And because it's built to work with Jira, everything stays connected from idea to delivery. Used by product teams at Canva, Deliveroo, and even The Economist, check out why and try it for free today at atlassian.com/product-discovery. That's A-T-L-A-S-S-I-A-N.com/product-discovery. Jira Product Discovery, build the right thing.Today's episode is brought to you by Vanta. As a founder, you're moving fast toward product market fit, your next round, or your first big enterprise deal. But with AI accelerating how quickly startups build and ship, security expectations are higher earlier than ever. Getting security and compliance right can unlock growth or stall it if you wait too long. With deep integrations and automated workflows built for fast-moving teams, Vanta gets you audit ready fast and keeps you secure with continuous monitoring as your models, infra, and customers evolve. Fast-growing startups like LangChain, Writer, and Cursor trusted Vanta to build a scalable foundation from the start. So go to vanta.com/aakash, that's V-A-N-T-A.com/A-A-K-A-S-H to save $1,000 and join over 10,000 ambitious companies already scaling with Vanta.

5. 12:53 – 19:37

   AI Prototyping Deep Dive

   1. MNMarily Nika12:53

      All right.

   2. AGAakash Gupta12:53

      What are the other tips and tricks while using Google AI Studio for prototyping versus other prototyping tools?

   3. MNMarily Nika12:59

      Well, if you don't like the design or anything that it outputs, you just talk to it and you say, "Look, make the background darker with a better text contrast." I find that it really understands the tweaks and the visual tweaks better than other tools I've tried. And I mean, presentation is really, really important, and personalization. And I don't know if you're like me, but I love creating little apps for my kids. So, um, the fact that I can specify which font to use and make it fun is very, very important. All right, so we have the first version of this. You can see the very first, um, action item right here is to upload a file and then edit your photo. You can describe and say, "Well, here's how you can change this photo before we add the bucket list goals." Okay. Well, let's try it and see how it goes. I'm gonna upload a photo of myself. And I-

   4. AGAakash Gupta13:53

      And one thing that's different also is that most other AI prototyping apps are powered by Claude as the main model behind the scenes. This says Gemini is the main model. Claude almost always uses, like, gradients and stuff in its design. [laughs] You don't see gradient, you don't see purple here. That's just one obvious thing that hit me.

   5. MNMarily Nika14:10

      Yeah, that is a good point. All right, so I said add, add a retro filter, so that'll be interesting. Okay. Well, bucket list items like meet Taylor Swift, um, give a TED Talk, um, be at the most beautiful, beautiful island. Actually, work from-

   6. AGAakash Gupta14:31

      Can we try something hard like-

   7. MNMarily Nika14:33

      Yeah

   8. AGAakash Gupta14:34

      ... meet MrBeast or something? [laughs]

   9. MNMarily Nika14:37

      [laughs] Um, let's, let's do... Hmm. I don't know if that's hard. Like, should we come up with, like, some movie character?

   10. AGAakash Gupta14:46

       Sure.

   11. MNMarily Nika14:47

       Um-

   12. AGAakash Gupta14:48

       Meet Tom Cruise or-

   13. MNMarily Nika14:50

       Yes

   14. AGAakash Gupta14:50

       ... meet The Rock.

   15. MNMarily Nika14:51

       Tom Cruise. Yes. And then bucket list. Um, what goes on-

   16. AGAakash Gupta14:58

       Be in front of the Eiffel Tower or something.

   17. MNMarily Nika15:01

       Yes. Okay. Of the Eiffel Tower. All right, generate collage. Applying creative edits. But this is an amazing first iteration. Like, you, you don't need backs and forth, backs and forth. Now I'm trying it at once to see what comes out of it. Great. To see what comes out of it. Okay, so it needed a couple of iterations. Um-

   18. AGAakash Gupta15:29

       It's actually pretty fast, though.

   19. MNMarily Nika15:31

       It's pretty fast, I will give it that, yes. Oh, but it keeps, "Failed to call Gemini API." [laughs] I have exceeded my quota. [laughs] Um, I wonder if I can quickly purchase more. Well, you can see how much I've been using it, so let me just quickly-

   20. AGAakash Gupta15:48

       [laughs]

   21. MNMarily Nika15:48

       ... increase Gemini quota. [laughs]

   22. AGAakash Gupta15:52

       Always happens during the live demo-

   23. MNMarily Nika15:54

       Oh

   24. AGAakash Gupta15:54

       ... that you exceed your quota.

   25. MNMarily Nika15:56

       This is insane. Okay. To adjust, request quota adjustment. Oh, bummer. Um, system limits. I wonder if we should use a different account, but we'd have to redo the whole thing again. Um, you know, maybe, maybe for the purposes of what we're doing, maybe stopping here is good enough, where we saw the first iteration, we talk about what will be next, and then we move to the second tool, Opal, to do the comparison. I don't know what you think.

   26. AGAakash Gupta16:28

       Yeah, that's fine with me. I want to ask you one more question on AI prototyping before we move on, though.

   27. MNMarily Nika16:32

       All right, let's do it.

   28. AGAakash Gupta16:33

       When should you be using an AI prototype? Does it start in planning? Does it start in discovery? Does it replace a PRD? Across the product development life cycle, how are you using an AI prototype?

   29. MNMarily Nika16:45

       [lip smack] I love it. I love this question, and I was just talking about it with someone. So it used to be the case, idea, PRD, comments on the PRD, product reviews, alignment, right? Whereas now, the workflow is you have an idea. You don't need to reach out to anyone yet. You don't need to write a PRD. You get in there, you talk to AI Studio, you have your idea visualized, and then you pull people in. So what I do is I prototype and then bring my s- my scientist and engineer in, and I say, "Hey, here's what I'm thinking. Here's the functionality, here's the end behavior. What are your thoughts?" So we debate on the actual prototype, and that saves us so much time. That makes the work so much more fun. And it used to take convincing to get engineers on board, for me at least, but now they're more easily persuaded because I can easily communicate my vision. So I use it at that time. Now, if it's something super complex, documentation still really matters, especially if it's a highly cross-functional product, especially if more, you know, people from all over the world are gonna have to work on this, different time zones. So you need it for these kind of async reasons, but it's not as crucial as it used to be. So PRD is needed for super complex stuff.It's kind of more deprecated for startups, I would say. And the gap between what you see in a startup and mid to big tech is kind of widening now. Um, so the PM is really different in the startup versus mid, um, to big tech. But yeah, that's my new workflow, and I love it 'cause it's creative, it's fun, and it's not predictable. For example, look what happened here where, you know... Is that Sailor Moon [laughs] for the second one?

   30. AGAakash Gupta18:40

       [laughs] So one thing that everybody asks me when they're using air prototyping tools is, "How do I get it to use my design system?" How do you get Google AI Studio to match what their current site looks like?
6. 19:37 – 27:03

   Opal vs AI Studio

   1. AGAakash Gupta19:37

      apps. Can we look at-

   2. MNMarily Nika19:38

      Yes

   3. AGAakash Gupta19:38

      ... Opal next?

   4. MNMarily Nika19:40

      Yes. All right, so Opal is an experiment from Google Labs, and this is what the UI looks like. Again, there is this place where we add our natural language, or it has an interface where you can actually drag and drop. I don't know if you've used Zapier. It used to have-

   5. AGAakash Gupta19:57

      Of course

   6. MNMarily Nika19:57

      ... kind of the concept of Zaps, so this kind of is like Zaps. Um, I just love typing. I think it's the easiest thing. So I'm gonna create the exact same use case. So I will say, "Given a photo of myself," I will upload, "I want you to generate a collage of four photos and different bucket lists scenarios. Prompt the user for their photo and for four bucket list scenarios, or suggest bucket list items for them." And that's it. Just hit this, and you wait, and you'll see the w- automated workflow getting populized. You'll see the automated workflow getting populated, which you can edit, or on the right side, you can just use the app. All right, so let's see what happened here. So you see I said that the input needed to be the user photo and bucket list scenarios, and then the next thing in the workflow would be, okay, well, here is your prompt, Opal. The prompt got generated, like the whole thing. So for my one sentence, it got expanded, and you can see that massive thing, which is, okay, well, you're an imaginative, highly skilled image prompt generator, and so on, right? Here's your input. Next thing. All right, you take this as an input. Image generation has limited free quota. Okay, thank you. I hope it's gonna work. And then the output's gonna be a responsive HTML webpage that deploys a collage of four images. You can also choose, um, like a theme. For example, sci-fi, claymation, cats, and all these things. But yeah, let's, let's just see what happened in the first iteration. It called it Bucket List Vision. It created this, which actually is pretty on point, right? Because it's four images that I asked for. So let's start. Upload a photo of yourself. So I'm gonna upload the photo of myself. Enter. Opt for bucket scenario. Meet Taylor Swift, give a TED Talk, uh, go to Eiffel Tower, and surprise me. [laughs] All right, let's see what happens. So now it's actually expanding from my bucket list prompts to a fully fleshed scenario images. So let's see what comes out of it. I will say it looks like it takes just a little bit longer than AI Studio. Also, the output is gonna be an Opal app or an Opal. It's not gonna be code you can actually take and change and deploy, which is fine. Like, for a simple use case, I think it's just super easy to, to use Opal. Now let's see what comes out of it.

   7. AGAakash Gupta22:59

      So when should PMs be building workflows in Opal? What are the other use cases?

   8. MNMarily Nika23:06

      I mean, if you wanna create an empathy map, the easiest thing you can do is go on Opal and say, "Hey, just an empathy map." You, you have it ready. You just type in the use case and who the user segment is, and boom, ready to go. Or when you really wanna use Nana Banana in smart ways. All right. I mean, look at this. Backstage with Taylor Swift. A TED Talk. It's interesting because the Taylor Swift image also has TED, which, you know, is not right. Oh, they all have TED. Well, suggest and edit. Okay. All images were the same. Have distinct, separate images. Distinct, separate images. Let's see.

   9. AGAakash Gupta23:53

      I think that some other interesting use cases I'm curious if Opal would do is, like, drafting emails or analyzing your calendar for meetings to remove or things like that. Can Opal be used for use cases that other AI agents can be used for?

   10. MNMarily Nika24:10

       Yes, although it's still experimental. Like, I would not 100% rely on it just yet.And things do change, like new versions keep coming out. So I wouldn't add it as part of your core workflow unless, you know, you've really tested it out.

   11. AGAakash Gupta24:25

       Got it.

   12. MNMarily Nika24:26

       All right. This still seems a little bit wonky. Um, so with that said, I've already built this myself, so let's open the version I've built. It did take a few iterations, but it eventually got there. So let me open that version. All right. All right, so in this latest iteration I have, you upload user photo, and then you separate the four inputs to be separate so that the system really understand this is a separate image it needs to output. All right.

   13. AGAakash Gupta25:03

       Yeah.

   14. MNMarily Nika25:04

       So let's, let's try this again. You upload your photo. That's the most, the easiest part. All right. And then I trained this one to actually give you examples like handshake, handshake with Taylor Swift to be the first one. The second one, sitting courtside at the Warriors game. All right.

   15. AGAakash Gupta25:26

       I'm glad you like the Greek Islands as much as me with Santorini there.

   16. MNMarily Nika25:30

       Yes. I mean, I'm Greek, so it's, I'm kind of biased, but yes, it's, uh... And I'm going to Greece in two weeks, and I'm counting down. I think it's gonna be super fun. Uh, [chuckles] top of the Eiffel Tower. All right, so now it's generating the image instructions. Hopefully, we're gonna get the, the output we're looking for. But imagine something like this, you can just take the image and post it easily on LinkedIn and, I mean, obviously it would be fake things to brag about, but it's totally something that you can easily do for, you know, more realistic things that you've already done and say, "Hey, I didn't take a photo when I did this, but here's the memory I have," right?

   17. AGAakash Gupta26:14

       I've often thought about, like, making those, like, realistic life shorts that you've probably seen where people have like, "Oh, this is when I graduated. This is when I got my first job." I don't have video of any of those things, so there's a use case.

   18. MNMarily Nika26:28

       All right. Oh my God, look at this. Look at this. Look at this. Look at this. So, well, [chuckles] I'm not really sitting at the Warriors. Um, Taylor Swift is more, like, in my office rather than me, um, being on stage with her, but this is me in Santorini, and this is me in the Eiffel Tower, and there is an extra one that's bonus, which is probably me in Greece snorkeling. [laughs] But this is great, you know. You can download the collage. Um, but again, fun. With a couple of tweaks, it's gonna get there. And yeah, I encourage everyone to try this out.

7. 27:03 – 31:50

   Notebook LM for Domain Expertise

   1. AGAakash Gupta27:03

      So workflows, really good for automating work that you have. The next one you have is your learning companion, NotebookLM, and I've been hearing a lot of hype about this. How do you use NotebookLM like a pro?

   2. MNMarily Nika27:14

      Here is my use case for NotebookLM. So NotebookLM, I don't know if people are familiar with it, but it's kind of your personalized research assistant. So whether you're a student or someone that wants to upskill, you load up only the content you care about in all these different formats, PDFs, handwritten notes, YouTube videos, websites, a Google Drive that if you put more files in there, it's gonna automatically sync, and you can learn things about a specific topic. Now, what I decided to try using this for is my bootcamp. At the end of my bootcamp, I have 150 people, and they all have formed teams, and they pitch their product. Now, I record every pitch. I have audio clips of all of them. And what I do is, and I'll show you an example right here, I upload the audio clips of each team, and this is my demo day on June 13. And then I go to Audio Overview, and I say, "Okay, do a deep dive. You are judges of Marily Nika's AI PM Bootcamp, and here are all the pitches. You need to select the top two based on creativity, impact, and simply better storytelling." Okay. That's it.

   3. AGAakash Gupta28:45

      So is the advantage here compared to using Gemini that it can handle more context?

   4. MNMarily Nika28:51

      It's that it only uses the knowledge you provide.

   5. AGAakash Gupta28:57

      So it's so context.

   6. MNMarily Nika28:59

      So you can... Context. And you can, of course, create a custom gem and do this, but I don't want to create a custom gem every single time I want to use NotebookLM, right? So this is like a custom gem that you can use and reuse and reuse, and the best part is these audio and video overviews, and of course the mind maps and the flashcards. But what I'm doing now is generating audio. It takes a few minutes to get generated and, you know, imagine my big bootcamp and the end of it, and everyone is pitching, and they're kind of waiting, and then we play it, and they hear who the winner is. Now, what's fascinating is that, you know, I have actual judges, like I have some VCs to be judging. Um, it's 100% correlated to what the VCs will pick and what AI picks. So sometimes I tell AI, "Hey, here's what the VCs picked. Choose something else." Um, so it really works, and it's such a cool, fun use case because it speaks back. I tried something like that with my kids when, you know, the AI spoke to them about how well they did in their drawings and their homework, and they, they lost their mind, and now they want to do it every day. Um, but yeah, let's see what comes out of it. [chuckles] All right, it takes a while.

   7. AGAakash Gupta30:12

      So if you secretly want to motivate your kids to do something, you're telling me use NotebookLM.

   8. MNMarily Nika30:19

      But it's such a fascinating tool. I just wish I had this when I was growing up and going through my PhD. I remember the first day of my PhD, I walk into my supervisor's office, he gives me five books like this and says, "Take this, come back next week, and we talk about what you read." And I said, "Wait, what? Where do I start? What's, what's going on?" So, uh, this would have really changed my, my life. But, um, on the other side, I'm glad I got trained the way I did. You know, these are skills you, you got that really helped my entire career. But yeah, things are changing in how we learn and how we do things.

   9. AGAakash Gupta30:59

      So NotebookLM could handle five whole books?

   10. MNMarily Nika31:02

       NotebookLM can handle so many things, and I think the biggest thing for me, here's the other use case I have. I had interviewed for this other company, I'm not gonna name which one it is, and they sent me a four-hour YouTube video that was the investor relations presentations by the leads. It was four hours. My interview was the next day. And what I did is I uploaded the, the job description on NotebookLM, and I added the YouTube link of the investor relation meeting, and I said, "What do I, applying for that job, need to know from that video?" And it gave me 15 points. I memorized them all, and then I went to the interview the next day, and I was

8. 31:50 – 34:39

   Ads

   1. MNMarily Nika31:50

      like, "Oh, yeah, just like what was announced in, in this video," blah, blah, blah. [laughs] And I just did so well, um, mainly because of this, right? You, you would not be able to do anything like this before.

   2. AGAakash Gupta32:01

      I hope you're enjoying this episode with Marily. If you are, be sure to check out her course, the AIPM Bootcamp. This episode is brought to you by Maven, where that bootcamp is held, and you can get a special code by using my link in the description to get a discount. So be sure to use my link in the description. Check out Marily's AIPM Bootcamp if you want to learn from her 11 years of AIPM experience at Google and at Meta. And now back to today's episode. Before we dive deeper, let's talk about something every PM faces: getting alignment on product decisions. You know that feeling when you're trying to explain a user flow to engineering or justify a design choice to leadership, and you're just describing it with your hands? That's where Mobbin comes in. Mobbin is the world's largest library of real-world mobile and web app designs from industry-leading apps like Airbnb, Uber, and Pinterest. Instead of spending hours taking screenshots or hunting for inspiration, you can instantly find exactly how successful products handle onboarding, paywalls, checkout flows, whatever you're facing. Over 1.7 million product builders use Mobbin to benchmark against best-in-class products and show their teams proven solutions. Whether you need to convince stakeholders there's a better way to handle user activation or research how top apps approach feature discovery, Mobbin gives you the visual proof to back up your product decisions. Check out mobbin.com/aakash. That's M-O-B-B-I-N.com/A-A-K-A-S-H and get 20% off your first year. Today's episode is brought to you by the AI PM Certification on Maven. Run by Miqdad Jaffer, who is a product leader at OpenAI, this is not your typical course. It's eight weeks of live cohort-based learning with a leader at one of the top companies in tech. As you know by now, the future of PM is AI, and this certificate will give you the learnings plus the hardware to show you are ready for an AI PM role. I myself took the course and recommend it. Put on by the amazing team at Product Faculty, including Mo Ali and Paul McHern, it's worth it. Former students come from companies like OpenAI, Shopify, Stripe, Google, and Meta. The best part? Your company can probably cover the cost. So if you want to get $550 off, use my code, AAKASH550C7. That's A-A-K-A-S-H-5-5-0-C-7, and head to maven.com/product-faculty. That's M-A-V-E-N.com/P-R-O-D-U-C-T-F-A-C-U-L-T-Y. Yeah, 'cause I think Gemini has like a million token context window, so this is able to handle much bigger context.

9. 34:39 – 37:23

   Notebook LM in Action

   1. MNMarily Nika34:39

      All right, so it's ready, and it has an interactive mode. Imagine calling in on a radio show and you kind of get to speak to the hosts. So now we're gonna speak to the hosts that are presenting the award for my bootcamp, okay? Let's see. All right. Cut to the chase. Who are the winners?

   2. SPSpeaker35:00

      Welcome back to "The Deep Dive." Okay, so you sent over this, uh, really fascinating stack of sources from the AI Product Management Bootcamp Demo Day, and our mission, well, it's pretty clear, right? We need to act like VCs or product leads here.

   3. SPSpeaker35:12

      Exactly. Sift through this, uh, really diverse... Oh, hey, I think our listeners got something to say. Ha, I love the energy.

   4. MNMarily Nika35:20

      But this is fascinating. We can talk to these hosts and ask them questions. Imagine what this can do to our kids. When our kids grow up, this is how they'll be used to learning. It's just unbelievable. And, you know, for my purposes, it makes an amazing kind of ending to, to my bootcamp. Um, but yeah, that's how I use it, and it is very different, and people love it.

   5. AGAakash Gupta35:45

      So what are the use cases for a PM of NotebookLM?

   6. MNMarily Nika35:49

      A PM of NotebookLM is if you wanna build a product for a domain you've never touched before. Let's say you're getting into a new zero-to-one healthcare product, right? You upload books around medicine, videos. There are so many free resources on YouTube, but it's just so overwhelming to figure out what to study, what you need to know, what to read, what to watch, so you load everything up. Sometimes I don't really filter what I load up for a use case, but then I tell it, "Here's what I want you to derive from this content. Also, if this content is not relevant to me passing this interview or to me building a product, just ignore it, and I'll find more content." So I do it a lot for new domains where I need to build products for. I do it a lotAfter hundreds of pages worth of, um, UXR, where, you know, interviews of users, and I say, "Okay, so what do people think, like really think about my product? What should I really improve?" Or can you read between the lines of what people are saying in these user interviews? And as you know, a lot of user interviews are two-hour clips per user, and imagine that after 200 users, right? So this helps you distill content, and it's such a superpower that PMs need.

   7. AGAakash Gupta37:11

      I've been preaching NotebookLM for quite a while, so the sole context and the big context window are just very hard to beat, very powerful for PMs, like you said, for user research, for new domains.

10. 37:23 – 40:00

    Perplexity for Reddit Research

    1. AGAakash Gupta37:23

       The next tool you talked about is Perplexity. How are you using Perplexity?

    2. MNMarily Nika37:28

       Right. So this is Perplexity. The reason I love using this is because it has this little filter. So I turn off search across the web, and I turn on search on discussions and opinions. And I say, "Would young professionals be interested in a ring that tracks their steps?" And you immediately see what people are talking about on Reddit. And look, at least 20 sources, and you have the link so that you can go there and see what people have to say, and discussions, and you can get there. So you do this, and you immediately know what people are talking about, and you can make decisions as a PM. Now, I take things one step further, and I can say something like, "Okay, so give me a list of features based on what you read that I need to incorporate in my MVP to maximize the chances of me finding product market fit." Now you'll know what the entirety of Reddit is interested in, and again, it's reviewing the sources. I guess this is what would come after... On Notebook as well, if you had uploaded all the knowledge of Reddit on Notebook, right? But we don't have that. Perplexity gets that. There you go. So you have your must-have core features, accurate step tracking, all-day comfort, battery life, and so on. So it's just phenomenal, right? Look at this.

    3. AGAakash Gupta39:08

       Wow.

    4. MNMarily Nika39:09

       Look at this.

    5. AGAakash Gupta39:10

       Pretty differentiated use case for Perplexity there.

    6. MNMarily Nika39:13

       Yeah. That's it.

    7. AGAakash Gupta39:15

       So you can get in deep to understand what users are saying on sites like Reddit. Does... Are there any tips and tricks or advanced ways to use Perplexity?

    8. MNMarily Nika39:26

       Tips and tricks. Um, that's literally the only use case I use it for. I will say the next thing I wanna do is download the new browser and kind of see how that changes my entire workflow. But, um, yeah, like that's, that button here is the next thing I'm gonna try, and then we can meet for the third podcast and see. [chuckles]

    9. AGAakash Gupta39:50

       Awesome. So your next tool here is the one everybody's probably heard of, ChatGPT, but you have a very specific use case for ChatGPT.

11. 40:00 – 55:16

    ChatGPT PRD Generator

    1. AGAakash Gupta40:00

       How are you using ChatGPT?

    2. MNMarily Nika40:01

       All right. I use ChatGPT because I have PRD generator that has my voice. It literally speaks like me, and I've trained it to either enhance my current product with AI, so like one-to-one products, or to brainstorm a new product, like zero-to-one products. And I've been using this so often. What's interesting is that I, I pa- I posted it on LinkedIn, and it has over 10,000 people that have been using it as well, which tells me something about this, right? Now, we were on Perplexity before, and we have a list of, um, things and features that our, you know, ring for fitness would include. So we copy this, and we can go here and say, "Okay, let's brainstorm a new AI product." Now, I've trained this to ask and really probe for questions. "Well, what space are you interested in? Who are the users? What goal do you wanna achieve?" Okay, so I will say, "This is for young professionals that want to get healthier. It's a ring wearable, and these are the features it should have." Oops. Da, da, da, da, da. Stop. Stop. Stop. All right. Let's do this again. "These are the features it should have." Okay. So now it is gonna... It confirms that it understood what I asked it to do, and it wants me to confirm what is the goal. So it makes you really think that you have a PM against you, 'cause like, "Wait, I'm not gonna build unless I know what your goal is." So I will say something like, um, well, "Increase ring sales," right? Let's say that's what I wanna do. So then it says, "Okay, so your recopy is target user, product, goal of adding AI," and so on. And then that's it. It's g- it's calling it Microcoach, has the use cases, AI health summary for sharing, AI habit designer. Let's go with the AI habit designer. So I trained it in a way that it will do some due diligence before it gets the PRD going. And as you see, here's the entire PRD. It has my voice. It has user personas, Alex, 29, Maya, 33, use cases, AI use case, tech stack, key features, market research, and all this good stuff. So, um-It also should generate a UI mock, um, so that you're ready to go. It's amazing

    3. AGAakash Gupta42:45

       What's your workflow for using this? Like, how much would you edit this on top of this?

    4. MNMarily Nika42:50

       Usually I will... Well, that's the thing. If I do things for my day job, I'm not gonna add my confidential things here, right? I am gonna use, um, you know, 1P tools. But if it's something, you know, outside of that, the next thing I will do is I will copy that into a Google Doc and then take it from there, which is not optimal. I feel that we're still missing the step that's gonna take you to the last mile, so to go from that to a Google Doc. Or the other thing I will do, I'll copy that as is, and that's gonna be my input to an AI prototyping tool like AI Studio. So yeah, depends on the use case and what I wanna do, but that's why I say I do a lot of tool hopping, because, you know, these are different tools, and every single day they're all interconnected to some capacity.

    5. AGAakash Gupta43:38

       Some people are saying, like, you shouldn't use AI for PRDs because it leads to too much slop and a lot of overly use of language when making the documents longer, and it doesn't... A lot of the process of getting, um, a good PRD is writing it, because writing is thinking. What is the right amount of AI to use when writing a PRD?

    6. MNMarily Nika44:01

       Well, as I said, I've over-trained this to really have the most important things and to sound like me. I even uploaded a bunch of my PRDs that I could share externally so that it could, um, you know, generate the right amount. But even if you compare the worst generated PRD with nothing, the worst generated PRD is still gonna be an amazing starting point for you. I'm talking about basic things like you don't have to make, to copy a file so that you can, you know, write different title. Even having a title in place is a benefit. So I absolutely think that it does pay off getting the time to get the PRD generator right, because it is gonna save so much time for you down the road.

    7. AGAakash Gupta44:52

       Hmm. Okay. Awesome. What about product strategy documents, other artifacts that PMs need to be creating? Should they be using GPTs like this?

    8. MNMarily Nika45:02

       Um, let's see. I use custom gems or custom GPTs for all my artifacts. Now, there are some artifacts I don't really often do. For example, landing reports, right? When you launch something and you wanna see how things went. I don't think my landing reports are very consistent with each other, so for that I'm not gonna create a custom GPT. But I have a PRD generator. I have a PR FAQ generator, which is also very, very useful. Um, but I don't think... And for product strategy, I don't have, um, a, a consistent use case there, so I don't, I don't use it. But yeah, these two, PR FAQ and PRD, I always use the custom GPTs.

    9. AGAakash Gupta45:44

       What is the difference between a PR FAQ and a PRD?

    10. MNMarily Nika45:47

        The PR FAQ is a very interesting practice that I first heard from some Amazon colleagues, and they said it's essentially after you know what kind of what you wanna build, you imagine yourself at the end of the road where you, as if you were writing a press release. So they say, "What would the press release look like? What would the FAQ for the press release be like? What would the quotes praising this look like?" And that really helps you visualize and understand the end goal so that you can work backwards from that as you're building the product. And in the beginning, you know, it didn't really resonate with me 'cause I've never used it in the companies I worked for. But the more I started introducing it and training my mind around that, I realized, "Okay, this is, this is a good practice," and we always use it as part of my bootcamp. So, um, a PRD is about the requirements of what you will be building, and then the PR FAQ is just a press release. It's much, much, much shorter, and it tells you what the end goal is and what things are gonna look like at the end.

    11. AGAakash Gupta46:50

        Hmm. Makes sense. Do you see any mistakes PMs often use when using AI to write artifacts that they should be aware of?

    12. MNMarily Nika46:59

        Yes. Big one. If you want one takeaway from this podcast is don't be embarrassed to admit that you use AI. There's so many people that will share their PRDs with me, and they're gonna feel bad if I can tell AI was used. So the biggest mistake is hiding that you used AI to generate your PRD. Don't hide it. Thrive. If anything, tell people, "Hey, I used this PRD generator for this PRD. Here's the link if you wanna use it as well." So that's the number one thing. It's not embarrassing to use AI at work. We need to normalize this. It's the end of 2025.

    13. AGAakash Gupta47:36

        I guess the reason people are scared is they just feel like, "Isn't the PRD, like, what they hired me for?" Like, um, "Is the AI gonna replace me if AI can write a PRD?"

    14. MNMarily Nika47:48

        Who is gonna replace PMs is people using these tools when you're not. This is who is gonna replace PMs. And I see this resistance a lot on my students as well, but then the fact that they're in my bootcamp means that they're starting to worry and say, "Okay, it's time for me to start gaining different perspectives and switching the way I think." So we need to adopt this. Using PRDs at work is not cheating. And it used to be the case, I remember it very well. I think three years ago, I remember saying, "Ugh, the past three weeks, ugh, I was just creating the PRD." And now it is, it seems so funny that I spent three weeks on a PRD and backs and forths and comments. Now I will have Gemini take notes from a product review. I will share my PRD with it, and then at the end of the product review, it will have adjusted and changed my PRD and ask for permission and say, "Hey, how about this? How about this? How about this?" And I'm kind of done in, in two days, if you will. So it's just, um, the way we work is changing, and it's evolving, and it's, it's a wonderful time to be a PM.

    15. AGAakash Gupta48:55

        Speaking of taking notes, you use Fireflies as your note-taker. Why, and how can people use it well?

    16. MNMarily Nika49:02

        I use Fireflies for what I do outside Google, and the reason I do that is because whenever I have a meeting, anything that's on my personal calendar invite, I can't use one consistent app. But Fireflies joins everything. So Fireflies will join you in Zoom. It will join me on my Google Meets. It will be with me everywhere. It's very interesting. There was a note-taker, I think it was Fireflies, but I'm not sure, I'll look it up after the podcast, where they actually announced that, "Hey, the person that was joining your meetings before GenAI was an actual human pretending to be an AI," and they were manually-

    17. AGAakash Gupta49:45

        Yeah, it was Fireflies.

    18. MNMarily Nika49:46

        It was Fireflies. Okay, good. [laughs] And they were pretending to take notes and then send you the actual outcome. And I was just thinking about this. This is the ultimate MVP, right, of an AI tool. Obviously, this is very scary, and who knows who was joining, but it was just so interesting. Um, but yeah.

    19. AGAakash Gupta50:02

        Yeah, they basically set the lawsuits up by admitting that themselves. [laughs]

    20. MNMarily Nika50:06

        So how did we find this out? Did they even admit it? Like, how, what happened here?

    21. AGAakash Gupta50:09

        Yeah, the founder posted on LinkedIn.

    22. MNMarily Nika50:11

        Yeah. I mean, it's a great story, right? I think it reminds me of Airbnb and how they started when they... I think they were reaching out to Craigslist postings, and they were saying, "Hey, I'll come to your place, take better photos, and then we'll get a cut if your house ended up renting." Anyway, it's... I just love these bootstrapping early days of, um, super cool things.

    23. AGAakash Gupta50:31

        Here's the post from the Fireflies founder. [laughs]

    24. MNMarily Nika50:35

        Oh, yes. Oh my God, this is awesome. This is awesome. Yep.

    25. AGAakash Gupta50:40

        So why Fireflies? Why not another note-taker? What makes Fireflies particularly the one PMs should use?

    26. MNMarily Nika50:46

        Because we hassle a lot, and we use a lot of different meeting tools, and if you're Zoom or if you're on Google Meets or if you're anywhere, it's just gonna join in on its own, so you don't need to worry about that. My favorite note-taker is Gemini for Google Meet, but you can bring Gemini on Zoom, and I do end up having a lot of Zoom calls because of, you know, my, my academy work. So I think I use a combination of Gemini when I have Google Meet and then Firefly when I don't.

    27. AGAakash Gupta51:15

        Awesome. That is your AIPM tech stack. What is your opinion on trying new tools? Like, how often should PMs be trying new tools? If... What is on your personal tool list to try more of? You mentioned Perplexity Comet.

    28. MNMarily Nika51:29

        The next thing in my list when I find time is to try out the two new different browsers, so Atlas by OpenAI and then Comet by Perplexity. I think every single PM, no matter the seniority, no matter the company, no matter how busy you are, no matter how many deadlines you have, you need to block your calendar for one hour a week and call this experimentation time. You need to not be married to a tool in the way that we used to be with Excel and Microsoft Word because you're likely to change your workflow every single week or every single day sometimes. I always, whenever there is a tool I love, I say, "Okay, this is great," but I know in the back of my head that there may be something that I resonate with more the next day. Now, in terms of figuring out what to try next, best thing to do is follow our newsletters, um, follow the rundown, fo- follow, you know, the LinkedIn creators. If something worthwhile spending time and experimenting on comes out, you will see and hear it there first. But please open up time to do this. It is very important for you and your team and the future of your career, honestly.

    29. AGAakash Gupta52:45

        There was this viral post recently by Peter Yang where he interviewed the head of design at Cursor. They reached $29 billion with no full-time PMs. We just walked through your AI tool stack. A lot of PMs are kind of nervous about the future of PM with these PMs being so AI-powered. What's your take on this, and how should PMs really be thinking about the, the future of the AI-powered PM?

    30. MNMarily Nika53:09

        Our craft is here to stay. We're not going anywhere. We still need to know the what and the why and the how to measure success. The tool is not gonna tell you what success means for the mission of the company. The tool is not gonna understand the morale of the company. It's not gonna understand the whys behind strategies that may be very bold from your competitors, let's say. In some of the companies I was working for, I remember it like yesterday. I remember we had kind of a roadmap we were executing against as an org, and then this other company came out to launch this random thing no one ever expected, and then we just scratched everything we were doing, and we started going after this. AI could never strategize in that way that could be so bold and say, "Okay, let's go after this." Some things really need judgment. Some things really need not to be cautious, or sometimes you really need to be cautious. It really depends. Um, but yeah, it does not replace personal judgment, and us as PMs are the ones bringing the why and the PM craft that is here to stay, for sure.
12. 55:16 – 59:30

    The 18-Month AIPM Roadmap

    1. AGAakash Gupta55:16

       So that's the masterclass on AIPM tools. I have to talk a little bit about AIPM because you are widely acknowledged as one of the world's leading experts in AI product management. So if somebody wants to become an AI product manager, let's say they want to get your job, what's the 18-month roadmap to landing an AIPM job at a company like Google?

    2. MNMarily Nika55:41

       I tell PMs that they need to be like crabs. What does that mean? That means that you need to move adjacent to what you've been doing. Here is an example. I had someone in my bootcamp, and he was working for Hearing aids, and he was telling me he always wanted to be a PM, but how does he break into PM? And, "I can't find an angle. I don't know how I'm gonna do it," and all these good things. And I said, "Just be a crab and bring what you have." And he said, "Well, I don't have anything. I was in a completely different domain." And we talked about it, and I said, "You're so wrong. Let's check it out." And then we went to the Apple Careers website, and they literally were looking for a PM for AirPods for the hearing aid feature AirPods now have. And he was just shocked. He said, "I would have never looked there." So we need to be open-minded. We really need to bring in the previous experience we have 'cause that's gonna set us apart. Another example, someone reached out. I think Meta had some role in sports in product management and AI. Um, and he used to work for ESPN. He was in sport journalism. And he said, "Hey, how do I pass these PM interviews?" And, you know, we really talked about it, and I said, "Don't come in to that interview being afraid that you are not good enough of a PM. Come into that interview saying, "Hey, I have the expertise needed. The rest I'll be able to figure it out as long as I show potential, as long as I show that I've done, you know, my work in these frameworks." And if that is the, the kind of unique weak point I have, then so be it. So bring your previous experience. Now, that's for AI experiences PM. Remember last time we were talking about AI builder, AI experiences? For AI builder PMs, it's more like platform APIs, SDKs, like bug and stuff. Um, it helps if you're technical. It helps if you have a CS degree or, um, you know, if you're a software engineer of some capacity. If you really want to become a PM in 2026, I think the best thing you can do is, um, other than joining my bootcamp, uh, the best thing you can do is just become AI literate. Your newsletter has so much content about AI. My newsletter has so much content about AI. Just understand the unique intricacies that AI brings. Understand the dependab- how dependent we are on data, understand the probabilistic nature of AI. It's all these little things, um, that really pay off. And that honestly will increase people's confidence into even apply for this role. There's so many people that would not apply because they don't think they're qualified enough, but that AIPM role is still getting shaped. Some people hiring for AIPMs don't really know what they're hiring for yet. So you want to be a part of the conversation, please apply. Please get AI literate. And please, you have all the resources out there for free, and we're here to help.

    3. AGAakash Gupta58:55

       Should aspiring AIPMs learn how to code?

    4. MNMarily Nika58:58

       Aspiring AIPMs should learn what goes behind coding. They should understand what version control is. They should understand why there are different languages and what's different behind these different languages. They should understand what an ID is, what an API is. So how to productionize and why productionizing is so challenging. Um, but you don't need to learn to code per se anymore, which I never thought I would be saying, but...

    5. AGAakash Gupta59:28

       So everyone

13. 59:30 – 1:02:10

    AIPM Interview Red Flags

    1. AGAakash Gupta59:30

       gets the AIPM opportunity, I feel like, these days. But where I feel people really struggle is the interview. And you've probably been on both sides of the table here for these AIPM interviews for years now. What are the biggest red flags you see in people interviewing for AIPM jobs?

    2. MNMarily Nika59:48

       There are so many things that come to mind. Number one, you see ex-machine learning scientists coming in for the AIPM role, and they say, "Okay, well, I've done AI. Now I'm gonna do the product." But they still really need to nail the PM craft. Like it's still product we're talking about. It's still about the users. Use cases don't change, but the way we go about fulfilling these use cases actually changes with AI. So we really need to demonstrate that we have that PM craft. And a red flag I see is someone coming in saying, "Hey, I know AI well. Here's what it is. Product, let's all sit together, create roadmaps." But I say, "Wait, we really need to represent the users here. It's all about the why. It's all about the mission. Let's set that." So when you ask how would you build a smart shoe or something like this, you don't want people to dive right into the solution, right? And they may be amazing at using the different trade-offs and all these different algorithms, but they really need to hear the standard PM stuff like the why, the who, how to measure success. It's just such a big thing. Um, another red flag is... And I'm surprised that red flag still comes up. It's people still mix up product and program. And-I, I interviewed someone a few months ago, and it was for some company, and they sa- and I said, "Well, why do you wanna be a PM in AI? Like, what's, what is it about?" And they said, "Well, I execute well, I'm very organized, I move fast." And I said, "All these things are not product specific. These things can be TPM specifics or program management specifics." And they were very surprised by the answer, and they just could not come up with a good reason as to why they wanna do this. And I feel we, we've talked about this, you know, how the PM is all about the what, and then the TPM is all about the how. Now, I think there were some companies that call- called program management product management. I think Microsoft w- used to do this. I think it finally changed. Uh, but yeah, there's still, if you go outside our PM bubble, there's still a lot of education that we need to do for PM, let alone AIPM in the mix, so.

14. 1:02:10 – 1:05:34

    The Future of Product Management

    1. AGAakash Gupta1:02:10

       So there was this director [lip smacks] at booking.com, and he said he hates the term AIPM.

    2. MNMarily Nika1:02:16

       Hmm.

    3. AGAakash Gupta1:02:17

       And he said, you know, there's so many categories of AIPM under it, and that's why he hates the term. Is AIPM a good term?

    4. MNMarily Nika1:02:25

       Very good question. It's so interesting. I was doing this before it was cool. I think it was 2000, I don't even know, like '18 that I, I said AIPM, and people said, "Well, what does that mean?" Like, "AI, what?" And we started having this title when, um, people had no idea what it was, [lip smacks] and then all, all, everyone started adding AI and saying, "I'm an AIPM, and I am AIPM." Now we're at the point where we do have many kind of subcategories of AIPM, and all these subcategories in two, three years will blur, so we're gonna come back to kind of that main AIPM persona. So, um, long story short, it's like saying you're in product. I know that I said this before, all product managers will be AI product managers in the future. We're kind of experiencing exactly that. It's like saying you're in product. But why is it bad? I don't understand why, you know, it's, it's bad. Sure, it does compress it, but so does VP of product does. There's so many different iterations of product. But yeah, the AIPM role is gonna be blurred into just PM, 'cause PM is gonna have an AI element in it. But for now, I think it's good 'cause it shows that you have the ability to understand that, you know, the underlying AI tech and the literacy I was talking about, maybe that you're working with models. It really helps if you wanna change companies and someone is y- looking for someone that's more familiar with this. So I don't think it's lazy in any way yet. Uh, but in the future, we're all gonna be product managers working on AI.

    5. AGAakash Gupta1:04:06

       Why is that the case? Because some people push back on me. They say, "Hey, 90% of PMs, they're not AIPMs. They're not working on AI features. They're not working on AI companies or AI products. They're not gonna become AIPMs."

    6. MNMarily Nika1:04:17

       Because eventually, no matter what feature you're gonna be using, it's gonna have sm- some smart functionality. Sure, if you work for a furniture company, um, if there's no smart tool to help you generate, like, and you design or personalize it in a way, you should not be called AIPM. But have you watched the "My Big Fat Greek Wedding" movie?

    7. AGAakash Gupta1:04:38

       Of course.

    8. MNMarily Nika1:04:39

       [chuckles] Okay. You know how there is, uh, you know, the dad says, "Show me any word." [chuckles] Every single word has a Greek... All right. So it's the same with me. Show me any product, and I will find how it uses AI. Maybe it's too early to say that, but in 2026, I think I'll be able to say this. Whether it's news or whether you're on Netflix or, as I said... Oh, I did a private training for a big retail company, and the first thing I asked is, "Why do you need an AIPM training?" And they said, "Well, because we have cameras, and we can monitor, you know, people walking in and which areas in the store they go to so that we can actually have insights and maximize what's gonna be more likely to get sold." So show me any product, and I will tell you how AI can be added to this product.

    9. AGAakash Gupta1:05:28

       Hmm. Makes sense.

    10. MNMarily Nika1:05:29

        We need a logo of, like, a Greek flag now for, [laughs] for this.

    11. AGAakash Gupta1:05:33

        All righty.

15. 1:05:34 – 1:06:47

    Outro

    1. AGAakash Gupta1:05:34

       Marily, you broke it down. You broke down the tools that people should be using from AI prototyping to PRD writers. There are so many different use cases as an AIPM. As Marily said, you should be trying out new tools for an hour a week. That's the most powerful thing. Take one of the tools that you, we even demoed here. Maybe it's NotebookLM, maybe it's Opal. Try it out. That's how you're gonna improve as an AIPM. [lip smacks] Until then, see you next time.

    2. MNMarily Nika1:06:00

       Thank you all. Thank you so much.

    3. AGAakash Gupta1:06:02

       I hope you enjoyed that episode. If you could take a moment to double-check that you have followed on Apple and Spotify Podcasts, subscribed on YouTube, left a rating or review on Apple or Spotify, and commented on YouTube, all these things will help the algorithm distribute the show to more and more people. As we distribute the show to more people, we can grow the show, improve the quality of the content and the production to get you better insights to stay ahead in your career. Finally, do check out my bundle at bundle.aakashg.com to get access to nine AI products for an entire year for free. This includes Dovetail, Mobbin, Linear, Reforge Build, Descript, and many other amazing tools that will help you as an AI product manager or builder succeed. I'll see you in the next episode.

Episode duration: 1:06:57