Skip to content
How I AIHow I AI

How to turn meeting notes into prototypes that your sales team can immediately demo to customers

Anjan Panneer Selvam is the Chief Product and Technology Officer at Acolyte Health, where he’s pioneering the use of AI across the entire product development lifecycle. In this episode, he demonstrates how AI tools can dramatically accelerate alignment between stakeholders, reduce development time from months to minutes, and enable teams to validate ideas with customers before committing engineering resources. *What you’ll learn:* 1. How to transform meeting transcripts into interactive prototypes in under 30 minutes using ChatGPT, Lovable, and other AI tools 2. A step-by-step workflow for creating market analyses and competitive research in minutes instead of days 3. How to build a “living product library” that allows sales and customer success teams to demo prototypes to customers before engineering begins 4. Techniques for using AI to break deadlocks with engineering by demonstrating what’s possible without requiring technical expertise 5. Why AI enables faster stakeholder alignment by converting abstract ideas into tangible, interactive experiences 6. How to use ChatPRD to validate product requirements and ensure you’ve considered all critical aspects before engaging engineering *Brought to you by:* Notion—The best AI tools for work: https://www.notion.com/howiai Lovable—Build apps by simply chatting with AI: https://lovable.dev/ *Where to find Anjan Panneer Selvam:* LinkedIn: https://www.linkedin.com/in/anjanps/ *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo *In this episode, we cover:* (00:00) Introduction to Anjan (02:36) How AI changes the relationship between product and engineering (04:08) Workflow for converting stakeholder ideas into prototypes (08:50) Using the Limitless pendant to capture meeting transcripts (12:45) Creating interactive prototypes with Lovable (15:57) Benefits of using prototypes instead of documentation (19:07) Conducting market research with Perplexity (21:45) Creating presentation decks with Gamma (23:08) AI doesn’t replace PMs; it elevates them (25:05) Using ChatPRD to validate product requirements (29:10) Building a living product library for sales and customer success (35:50) Breaking deadlocks with engineering using Rork for mobile prototypes (39:00) Takeaways for building with AI (42:34) Cultural implications of AI in product development (45:20) Strategies for when AI doesn’t give you what you want *Tools referenced:* • ChatGPT: https://chat.openai.com/ • Lovable: https://lovable.dev/ • Limitless: https://www.limitless.ai/ • Perplexity: https://www.perplexity.ai/ • Gamma: https://gamma.app/ • ChatPRD: https://www.chatprd.ai/ • Rork: https://rork.com/ • v0: https://v0.dev/ • Magic Patterns: https://www.magicpatterns.com/ *Other references:* • React Flow: https://reactflow.dev/ • Figma: https://www.figma.com/ • Acolyte Health: https://acolytehealth.com/ • Meta Ray-Ban glasses: https://www.ray-ban.com/usa/ray-ban-meta-ai-glasses _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Anjan Panneer SelvamguestClaire Vohost
Sep 1, 202548mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 2:36

    Introduction to Anjan

    1. AS

      now, that's a high-stakes meeting. They want something in their mind. Most product-led or more sales-led CEOs are very clear in terms of what they want.

    2. CV

      You wore this AI recording device in your meeting with your CEO, basically recorded this meeting. Then you're just taking the transcript of that meeting, and you are pasting it into ChatGPT to generate this prompt.

    3. AS

      At the end of this, you know, I have a fully functional now drag-and-drop canvas builder that could build a full entire user journey, and the goal here is not to say something is easy. It's more so to, are we all aligned and can be move fast, which I think is the next superpower as all these tools come out, and what could be possible?

    4. CV

      [upbeat music] 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, I have an episode that's both going to inspire product managers and CEOs and maybe strike some fear in the heart of product managers and engineers. I'm speaking to Anjan Panneer Selvam, CPTO of Acolyte Health, who is using AI in every stakeholder interaction, giving customers access and free rein to all the prototypes on their roadmap, and solving debates with engineering by building products. This is a fun episode. It tells us a lot about both tools and culture, and I think it's a must-listen for product engineering and startup executives out there. Let's get to it. This episode is brought to you by Notion. Notion is now your do-everything AI tool for work. With new AI meeting notes, enterprise search, and research mode, everyone on your team gets a note-taker, researcher, doc drafter, brainstormer. Your new AI team is here, right where your team already works. I've been a longtime Notion user, and have been using the new Notion AI features for the last few weeks. I can't imagine working without them. AI meeting notes are a game-changer. The summaries are accurate, and extracting action items is super useful. For stand-ups, team meetings, one-on-ones, customer interviews, and, yes, podcast prep, Notion's AI meeting notes are now an essential part of my team's workflow. The fastest-growing companies like OpenAI, Ramp, Vercel, and Cursor all use Notion to get more done. Try all of Notion's new AI features for free by signing up with your work email at notion.com/howiai.

  2. 2:36 – 4:08

    How AI changes the relationship between product and engineering

    1. CV

      Anjan, welcome to How I AI.

    2. AS

      Thank you so much, Claire. Very happy to be here. Uh, big fan of everything that you do, and, uh, also a big user of ChatPRD, so really thanks for putting it out there and helping a lot of us.

    3. CV

      Well, what I really like is you and I have something in common, which is we have this fairly rare but more common title, CPTO, thinking and caring for the product and engineering organizations. And before we dive into your use cases, I'm curious how that specific point of view has really changed how you think about AI or how you've adopted AI inside your startup.

    4. AS

      For a long time, I was jumping between CTO, CPO, and eventually I found this kind of, um, role, and I think, um, it's really big thanks to AI that the roles are now starting to... the lines are starting to blur for the roles, and it's now one role. Um, and, you know, early-stage startups specifically love the CPTO because they have one person to go ask everything to. There's, like, no two people in the room now. Um, and I think AI definitely has made it much more bearable and much more easier to manage the responsibilities that come with being, being both responsible for the product and the technology side of things. And most importantly, it, it cuts down the time required to-- that's spent in transitioning ideas, transitioning information, and now directly goes into how things can be

  3. 4:08 – 8:50

    Workflow for converting stakeholder ideas into prototypes

    1. AS

      made possible.

    2. CV

      What I really like about this role is that you get rid of this false conflict between product and engineering, but unfortunately, you do not get rid of, I wouldn't call it a conflict, but the debates and negotiations that happen between the product and engineering team and the CEO. And so I would love to dive into your first, uh, workflow, which is really about how to guide maybe an executive, whether they're your CEO or another leader, through an idea and really solicit out kind of use cases, user experience, all that stuff in a single meeting using AI. So let's dive into that.

    3. AS

      I work in B2B and enterprise applications mostly, so, um, uh, a very simple idea that came up, which I think, um, I've built this over the years, like the last seventeen years in my tech career, five different startups. I've built it at every startup-

    4. CV

      [chuckles]

    5. AS

      ... and the timeline always shortened. First, it took, like, six months, then it took a few months, then a few weeks, but now it's, like, in a matter of thirty minutes to two days, we're actually building it and also putting it into production, and that's how fast it's become. So there's this one idea that's, there's a problem that I love a lot. How can we build a user journey map within the application? So a lot of applications that I work on is where customers want to be able to tailor how the experience gets delivered to users. Normally, this-- Just imagine this: if we had no logic, we have to write down all the logic. A Figma designer has to go and design everything in Figma, then they have to think through all the possibilities. Next, engineering has to understand all the different permutations and combinations. Everything's written in document. Impossible to think of how much has to be done before we can even have something functional. So in this case, we are in a meeting, um, multiple stakeholders, um, and then the conversation comes up, "Can we build, like, a journey builder within our application?" So-... first step where I always start, um, I love starting with ChatGPT, because this kind of has become my brain dump of what I'm thinking through, and then it helps me consolidate. Um, so we'll start here, work through a very simple idea. So I want to be able to build a prototype here. Very simple explanations, right? Natural language that we're just conversing. There's no complex, like, here's, like, a structured PRD or a structured statement, just a pure brain dump of what everybody's talking.

    6. CV

      Yeah, one thing I'd like to call out here is, as much as I like to say, of course, that AI can do a lot of the product job, one thing that experienced product managers are very good at, that you've shown you're good at, is taking probably an idea from your CEO that's like, "I want a journey builder." Like, "I want people to be able to, like, build whatever journey they want," and you've actually given some pretty precise language to that. So there is still a job-

    7. AS

      Yes

    8. CV

      ... out there [chuckles] for, for us.

    9. AS

      [laughs]

    10. CV

      Despite perhaps my, my past, um, [lips smack] my past forecasts of the death of PM, I do think the ability to clearly articulate, either for a human partner or AI, something a little bit more specific, which I think you've done here, help, you know, a single-page canvas, user-facing tool, map out personalized workflows. Like, that's all very specific language that I think a great product manager can translate in very short order. So that's what you're putting into this prompt. And then what are you trying to get out of it?

    11. AS

      So out of this, I want actually a simple, structured, um, prompt that could go into another tool, like Lovable, v0 or Magic Patterns, any of those. Um, and, uh, the idea here is, I'm starting with something, and AI does a great job of understanding what else is needed. Because I'm saying I want it to be a Lovable prompt or a v0 prompt, AI understands what else might be needed. So as we go in, it starts to summarize everything that's needed, automatically picks, like, a pattern, like it's a left-side vertical toolbar that could have some... And the great thing it does is, AI helps normalize every good product manager's, uh, product sense or product feeling into very structured information. So that's what typically differentiates different product managers, is they have each a different way of explaining how they want it. AI doesn't judge you. It's like, "Okay, I get it, and I'm gonna help you." And I think this is where the biggest shift has been. Even product managers who initially want to think a lot, my push always to them is, "Just jump in and start typing. Don't think about what anybody's gonna think, what AI is gonna think, because it's very kind, very nice. Just jump in, start putting in." So one prompt later, not even too complex, we have a very nice prompt that I could then use, um, you know, um, into Lovable or v0. So from here, that's all straight it goes to. Um, directly we go into Lovable or v0, we put it in and start conversing... um, start building the prototype.

  4. 8:50 – 12:45

    Using the Limitless pendant to capture meeting transcripts

    1. AS

      One thing I wanna also men- will mention here is, it's not just limited to, um, something I type here. Um, I wanted to actually show in this use case, I was using the Limitless Pendant, um, something I was just experimenting with. This is actually live transcript that's coming from the conversation that we're having in the room, and I could go from this directly into ChatGPT as well, so helps normalize it. Or this could be a copy-

    2. CV

      Wait, I have to, I have to pause you. I have to pause you. Do you have the Limitless Pendant right now? Do you have it close?

    3. AS

      Yeah, yeah. Yeah, I have it.

    4. CV

      Wait, you gotta, you gotta show it, show it to us.

    5. AS

      Um, give me one second. Let me just-

    6. CV

      Yeah, yeah, yeah. Um, while you're, while you're looking, um, at it, or looking for it, I have to shout out my old pal Dan, um, who's the CEO and founder of Limitless, and will be thrilled to see [chuckles] -

    7. AS

      [chuckles] So this, this is my Lim-

    8. CV

      ... the pendant in real life.

    9. AS

      Yep, this is my Limitless-

    10. CV

      Okay

    11. AS

      ... Pendant. Um, I was actually one of the first few people to pre-order this. Um, I, I've worked with pendants my entire life, because I used to work in senior living, so we used to build-

    12. CV

      Yep

    13. AS

      ... like, the medical alert buttons. So when this form factor came out, I'm like, "I'm trying this." And this has been, like, such a game changer, especially, like, you know, you're sitting in a coffee shop, you're talking with, you know, stakeholders, partners, colleagues. Ideas flow, right? There should be no friction. "Oh, let me take a paper and write it down," or, "Let me remember it." So this just completely changed the game there. So this is, like, transcript, for example, from a meeting. We start here, and then directly go into AI. So straightforward and simple, so thanks for building this.

    14. CV

      [chuckles] Tha-

    15. AS

      [chuckles]

    16. CV

      ... thanks, thanks, Dan and other friends over at Limitless. I have to say, you did not tell me about the Limitless Pendant in our prep, and this has been-

    17. AS

      [chuckles]

    18. CV

      ... a How I AI first. So just to recap for folks that are maybe not watching or don't know what this pendant is, um, you wore this basically AI recording device in your meeting with your CEO, I'm sure, um, ascribing to all applicable disclosure laws about, [chuckles] about recording-

    19. AS

      Definitely.

    20. CV

      Of course.

    21. AS

      Yes, yes.

    22. CV

      Of course. Of course.

    23. AS

      Yes.

    24. CV

      And you basically recorded this meeting, and then that's streaming into the Limitless AI app. Then you're just taking the transcript of that meeting, and you are pasting it into ChatGPT to generate this prompt. We totally buried the lead on the, like, new How I AI workflow here. So again, it'll be interesting to see... I think we've seen a lot of, you know, desktop-driven meeting recording apps. Um, it'll be really interesting to see as these, like, wearables and devices come into the workplace-

    25. AS

      Right

    26. CV

      ... how that might shift. And I just have to ask you, you know, why, why the pendant? Why a hardware device or, or a wearable, versus, like, turning on something like Granola or another meeting recording kind of, um, product?

    27. AS

      I think, you know, sometimes you want it to be non-intrusive. I think that's, you know, um... I'm, I'm a big fan of also the Meta glasses. I'm waiting for better versions, lighter versions to come out. But this, this is such non- such a non-intrusive device, has amazing transcription, um, and the great thing is, it's not limited by language, too. So, you know, sometimes we have-... voice recorders, you have to open your phone, you have to set it on the table, or let me make sure I do it. So that's not always possible. When you're talking, you just wanna talk, you're thinking through, and you're naturally conversing. So it does a great, that, that's one, you know, great reason why I think I'm very bullish about hardware, um, that's going to carry AI, and, you know, it's gonna completely change what's possible.

    28. CV

      Okay, quick, quick request for folks watching or listening. In the comments, if you are also a hardware aficionado or have a use case around that, I would love to hear from you. Um, I, too, love the Meta, the Meta glasses. They are my daily drivers for, um, headphones and things like that, and I would love to hear more about how people are using, um, hardware

  5. 12:45 – 15:57

    Creating interactive prototypes with Lovable

    1. CV

      on How I AI, 'cause this has been a first. Okay.

    2. AS

      Yes.

    3. CV

      So side detour through, through recording, but you're basically taking these meeting transcripts, in the meeting you're saying, "Great, let's get it to a prototype," and then you're getting it into Lovable, for example.

    4. AS

      Mm-hmm. Yep.

    5. CV

      And this is an example of what you're getting out of it, which is this nice, beautiful workflow builder with a left-hand kind of card nav, which we had seen in the prompt, and some interactivity, um, I suppose.

    6. AS

      Yep, um, fully interactive. So here, now, I started with, um, you know, this prompt in, uh, ChatGPT. As I thought through this, and, again, this is one of the great things, right, when you're doing product and technology, you're constantly thinking of how this idea's going to evolve. So ChatGPT gave me a good starting point. I then went into Lovable, and as I was pasting in, I'm adding a few more points there. So, hey, I remembered I, you know, r- I ran into a library called React Flow that does great job at making, like, a canvas-based builder, so I put that in here. And I did not expect Lovable to know what React Flow is. There's always this like, "Oh, maybe it doesn't know," but AI always surprises me here. Somehow they [chuckles] it, it knew React Flow, and exactly that's what it went ahead to implement. And, you know, um, one of the things I always recommend is break down your concepts for AI, and you'll have a lot more success. So I broke it down, first starting with the basic idea, and slowly, as things progressed, kept building, building on it. Um, and at the end of this, you know, I have a fully functional now drag-and-drop canvas builder that could build a full entire user journey. And this, I, as I mentioned, I've, I've built this so many times in my career, and it has literally been, like, 17, 18-page documents that fully summarize what's possible. Pencil sketches, then, then UX designers go put it in Figma, all of that stuff, which, which is still okay to do now, but I think nobody wants to spend time reading 17 pages and then arriving at a conclusion. They would rather see, and this, this exactly does that, so very easy to do and, most importantly, talk about. So just imagine going to your engineering team. You're not giving them an 18-page document, but you're rather starting, "Let me show a demo of what I'm thinking." Now, we're moving from, "Oh, let's spend a week aligning," to, "Here's the alignment," in less than 30 minutes to an hour. And it's the same with the CEO, too. Now, that's a high-stakes meeting. They want something in their mind, and they're very, you know... uh, most product-led or more sales-led CEOs are very clear in terms of what they want. They might not explain it always, but now, here, they're visually seeing this, and they will say, "You know what? I want something else here that'll make it possible." So, for example, "I don't want too many variables," or, "I wanna be able to edit each part of whatever block that you have." And that's the great thing with these prototypes that we generate: everything is editable. It's no longer just static images that like, "Okay, let me put it on a board and explain it." It's fully explanatory, it's fully interactive, and then we go from this to first alignment with the stakeholders, which is, like, the leaders of the business. Everybody's on the page. Let's now move on to the next step. So...

    7. CV

      Yeah. I, I really love this. I wanna call

  6. 15:57 – 19:07

    Benefits of using prototypes instead of documentation

    1. CV

      out a couple things for folks that are listening or maybe not watching. One is this is really, it looks great, and Lovable's been a wonderful sponsor of the show, so I have to, have to shout them out. We do love them. And I do tend to pick Lovable [chuckles] when I want something to look nice.

    2. AS

      Nice.

    3. CV

      You know, like, the-

    4. AS

      Exactly.

    5. CV

      It, it tends to be out of the box. You know, it, it uses color intelligently. It has, it tends to have, you know, kind of beautiful, pretty modern user experience. So I do tend to pick this particular tool for something like this, especially where visual design is a differentiator. I think, you know, sometimes people ask me, "What, what prototyping tool should I pick?" And you're probably like me. It's like, "Well, whichever one I feel like is gonna do the job of the day." And, you know, when I need something that, uh, where UX is maybe a little bit of the differentiator or it's highly interactive, like data visualizations-

    6. AS

      Mm-hmm

    7. CV

      ... or a workflow builder, I've also found that Lovable does quite, quite a nice job. So that's one thing I wanted to call out. The second thing is, you know, I, I did a little, little tour in healthcare once, and these rules builders, flow builders are [chuckles] like, like everybody's built one or they've built 10.

    8. AS

      Yeah.

    9. CV

      And what's really hard for product managers in the past is you're exactly right. You have to sit there and write bullet points for every piece of logic, every customizable field. You have to go in here and say, "Okay, if I send an SMS, I have to check that the SMS has been, like, opted in. I can't send too many SMSs in a row."

    10. AS

      Right.

    11. CV

      Like, you have to do all this logic. It's very hard to document. It's very hard to test. And what I like about prototypes as sort of the source of truth for something that, that is this complex, is you can actually just click through and go, "Uh-oh, you're not supposed to put two SMS components in a row." That goes back into the prototype code. Or, you know, "You shouldn't do a delay on a delay." And so I think these are particularly effective for highly complex products-

    12. AS

      Mm-hmm

    13. CV

      ... because the alignment on the requirements and the testing becomes much more natural. It's very unnatural to read-... rule sets, um, it's much more natural to interact with something, something like this. So I think it's a really effective tool. And then the third thing I wanna call out, that in case people missed, because they're gonna be very mad at you, they're like, "Oh, no, you've given my CEO the idea, one, I can get a prototype in 30 minutes, two, [chuckles] that I can ship it in two days." But I actually think the thing that you called out was really important, which is those are high-stakes meetings, and they're high-stakes products when your CEO is asking for something. And the worst thing is you say, "Okay, I'm gonna come back here two or three weeks and show it to you."

    14. AS

      Right.

    15. CV

      And you show it to them, [chuckles] and it's totally wrong. They're like, "No, that is not-

    16. AS

      I didn't need it

    17. CV

      ... at all what I wanted."

    18. AS

      Right.

    19. CV

      "Not at all what I was thinking."

    20. AS

      Mm-hmm.

    21. CV

      Um, the team is angry, the CEO is frustrated, no one feels good, and so this is like a cheap path to failure, too, which is like, if you get this, and CEO is like... She's like, "No, not it," um, w- it, it costs you what? 15 minutes.

    22. AS

      Right.

    23. CV

      It's not a big, not a big deal.

    24. AS

      [chuckles]

    25. CV

      Move on with your life.

    26. AS

      Move on, yeah.

    27. CV

      So I think those three things are worth, worth noting in this particular workflow.

  7. 19:07 – 21:45

    Conducting market research with Perplexity

    1. CV

      But let's go to what you were talking about next, which is, great, we have, we have something like this. You know, CEO's gonna be like: "When can I have it?" And Sales is gonna be like: "When can I sell it?"

    2. AS

      Right.

    3. CV

      And a product manager is gonna ask their favorite question, which is, "Why?" Like, "Do people actually want this?" So-

    4. AS

      Exactly.

    5. CV

      Um, how do you do kind of testing? I'm c- I'm curious what, what the next part of your flow is.

    6. AS

      Right. So next part from here is three things. Um, I tried to put together some slides, um, thanks Gamma here, or Gamma. The same prompt that we were just using, um, I start here, um, with the Chat, um, same ChatGPT prompt, whatever came out. I go and iterate on this. So the next question, uh, and again, all of this can be done with the custom GPT tool that you can build for yourself, but I think, um, the new models have gotten good, where it understands what your persona is, where you're asking from, so we don't even have to go through that effort anymore. I just start, like, I wanna be able to do some market research. Can you now make this a little bit where I can put it into Perplexity or do some research, um, on it? So that's the next step that I start here, try to build a very simple, just, uh, statement of what we're trying to do, what is a little bit more about the actual use cases, what the customer types are. So on one side, I can use the same prompt, go into Perplexity, this particular use case that I just did. So here, one of the great things to do is deep research, um, which cuts down the time analysis time by, like, three m-... It brings it down to three minutes from whatever it used to take. Um, I think, I think the main advantage with a lot of this is it reduces the immediate emotional or cognitive burden on product managers or leaders, um, in general, to focus on what makes strategic value and forget about every other thing that's involved. So if we have to do an analysis, every product manager has their way of what all they wanna analyze, put into a competitive matrix or any of this. We get past all of that. That's what deep research, for example, in this case, does for us. We give it some information. Again, as I come in here, I put in some more information. I wanna build a market analysis on the following feature I'm working on, paste the rest of the prompt. It goes and builds everything that's needed as a starting point. So review this, so now your market analysis, to a certain extent, is done. This is, again, to build conviction within, first, the PM or the technology leader that I want- I know what I'm working on is of meaningful value, and I can monetize it.

    7. CV

      Mm.

    8. AS

      Very important in an early-stage start-up, even in late stage, but I think even more important in early stage, because if we're not seeing ROI immediately, that's more money we're burning, building products that don't generate immediate ROI. So

  8. 21:45 – 23:08

    Creating presentation decks with Gamma

    1. AS

      once that's done, um, export it as a PDF or goes into kind of Gamma from here.

    2. CV

      Mm-hmm.

    3. AS

      Build very basic slides. Um, makes it look pretty, so when we're talking to other PMs, other stakeholders, it's like work is already done. That immediately, you know, attracts them, makes them bring to the understanding that we have thought this through, which is true, which is all happening in our minds, but now AI is helping implement it. And then very basic slides to be able to talk through, um, talks about what are competitive advantages, strategic benefits, and then what the product is supposed to do.

    4. CV

      This episode [upbeat music] is brought to you by Lovable. If you've ever had an idea for an app but didn't know where to start, Lovable is for you. Lovable lets you build working apps and websites by simply chatting with AI. Then, you can customize it, add automations, and deploy it to a live domain. It's perfect for marketers spinning up tools, product managers prototyping new ideas, or founders launching their next business. Unlike no-code tools, Lovable isn't about static pages. It builds full apps with real functionality, and it's fast. What used to take weeks, months, or even years, you can now do over the weekend. So if you've been sitting on an idea, now's the time to bring it to life. Get started for free at lovable.dev. That's lovable.dev.

  9. 23:08 – 25:05

    AI doesn’t replace PMs; it elevates them

    1. CV

      One of the things that I find challenging working with AI, um, exciting and challenging, is AI has, tends to have an abundance mindset in that it tends to think all my ideas are great, and everything is opportunity. And so I'm curious if you've gotten a critical market analysis out of AI, or if it's always sort of seeking the data that tells you something is a good idea versus figuring out reasons why it might not work, or it's not the right investment. I'm curious what your experience has been there.

    2. AS

      Yeah. Uh, I, I mean, one of the things always, as you said, AI is bullish about everything that needs to happen. [chuckles] So I think some follow-ups are critical here. So in Perplexity, for example, when I'm doing it, I always have a follow-up at the end saying: "Can you play devil's advocate and tell me if it's really worth it?" Or, "Can you, um, do you see really value in any of this?" In other words, if you're going to ask AI once more to just make sure it is right, it is going to be accurate then to say-... compared to all of these other things that I generated, I think this is the reasons why it might be unique. And a pro/con analysis is, again, so easy, right? So you could just build it into your prompt or ask it. And I'll tell you, there are more features that AI has helped me say no to because it's all in front of me, and the data is readily available, which I think is, again, what is the most critical, right? You wanna s- save your time for the ones that are really valuable, really meaningful, but now you're no longer carrying the burden that, "Oh, my God, I spent three weeks working on this just to say no." It's done in an hour, and now you can feel more, you know, you can more freely say no to things and only focus on the things that are important. So I think it goes back to the earlier point you mentioned, too. Good product sense is still critical. Product managers are still critical. It doesn't replace PMs or what they do, it just makes them to focus on what the outcomes to be rather than the entire process. So that's, that's, that's where I have seen it be useful and

  10. 25:05 – 29:10

    Using ChatPRD to validate product requirements

    1. AS

      helpful.

    2. CV

      Okay, so in our stakeholder Olympics, you have hur- you know, crossed the hurdle of the CEO. We've got the CEO excited. You made a deck, so now all the product managers and executives [laughing] are-

    3. AS

      [laughing] All right.

    4. CV

      So you made a deck with numbers. That's the thing that makes product managers and executives happy.

    5. AS

      Yeah.

    6. CV

      What, what's your next stakeholder you go after?

    7. AS

      Um, more details, right? So there's always going to be a junior PM or the engineering leader who's going to ask, "Okay, tell me more. Um, I want to know what exactly you mean by this."

    8. CV

      Yeah.

    9. AS

      This is where ChatPRD is, like, extremely great for me because I use ChatPRD almost like a gate to, to the point that you mentioned, "Are you sure you wanna do this? Are you sure you have everything?" That's where I find the tool to be extremely helpful because while ChatGPT is there, it's generic. Unless you give it, like, a huge persona overload of what it needs to do, it's very generic. But ChatPRD, on the other side, is geared towards a requirements document or a product sense, um, that needs to be validated, and this is where I use the same prompt. But now, rather than ChatPRD going and saying, "Let me generate everything for you just because you asked me," it goes through a series of validations. "Have you thought through this? Have you thought, thought about what could be the disadvantages if you're building something? Do you have any specific in- inspirations?" This is where I find ChatPRD specifically helpful. Um, I have a template that I've built for myself, um, I call it, like, the CPTO stack, where there is a particular type of, uh, PRD that my team likes. So I sat with them, discussed with them, came to alignment, and then I went forward. Now, I validate all the questions here. This is kind of core of, or the last step before it's handed over to the engineering counterparts. Now they're gonna spend actual mental time thinking through the details. So all of it is done. It goes through a series of discussions, and we have now a fully functional PRD, um, easy to read, very small bullet points, no large paragraphs, no large JIRA tickets yet, and then it's now ready. And th- th- this is one that caught me by surprise, this future ideas and enhancements, right? So this is something I've been dumping into the chats as I go through ChatGPT, Perplexity, all of it. But ChatPRD did a good job of pulling this separately, what are future ideas and enhancements, what could be possible? Almost like a teaser to let your engineering team know, "Hey, I'm gonna come back with a V2 that's gonna be all this, too."

    10. CV

      [laughing] Yes.

    11. AS

      But it's, it's, it's a very subtle hint that gets their, you know, uh, creative juices flowing, and I think, uh, [chuckles] makes, makes the job easier.

    12. CV

      It's a, it's a very strange experience as a, as a founder or as a product person to have somebody explain your own product back to you. But I will say, for folks that have not experienced this product that I made, it really is just a replica of going through a product review with me, my poor, my poor product teams.

    13. AS

      [laughing]

    14. CV

      Which is, I've done so many, like, what we call, we call PRDR. We have a little song that goes like, "PR, PRDR."

    15. AS

      [chuckles]

    16. CV

      Um, it's very fun. But we have this product requirements desi- or a product requirements document review meeting, and I just found myself consistently asking, like, "Who else is doing this well? Are there other products that we, we like here? What's next? Like, is this v0? If this is v0, what's v1, what's v2?" And I do think this idea of using AI, no matter what your function is, as a gate to make sure that you've checked all the boxes of the next step without having [chuckles] to go to your boss or having-

    17. AS

      Right

    18. CV

      ... to go to, you know, your other stakeholders is, is a really effective way to sort of bring AI into the loop, um, in a way that ultimately saves everybody time.

    19. AS

      Right. Yep, definitely. And I think, you know, the one thing I would love, uh, which you're building already, is integrations. So from here it goes now to v0, which makes the job a lot more easier to keep iterating and refining. But I think as time progresses, more tools become integrated. It's going to be a big game changer for everything PMs and CTOs are doing, um, day to day.

    20. CV

      Yeah. I'm, I'm sorry to the engineers out there-

    21. AS

      [chuckles]

    22. CV

      ... but we will be building a Jira, a Jira integration to turn all of this into tickets.

  11. 29:10 – 35:50

    Building a living product library for sales and customer success

    1. CV

      Um, sad, sad to say. Okay, so you've gotten again through a couple key stakeholders: your CEO, your PM cohort, your AI CPTO here in ChatPRD. You know, let's get to customers. What do we do when we wanna actually get this in the hands of customers and validate it? What's your next AI-driven next step?

    2. AS

      Yeah, so customer success teams love this as well. So we built what is almost called, like, a living product library or demo library of pages especially, which are, like, close to the design language of our actual platform, and have these all as individual microsites that we deployed. So now, um, either sales teams or customer success teams, because it's intuitive, there's no, you know, um, there's no places where it's going to fail suddenly or I can't click because the Figma interaction was not done. No challenges of that nature. They have now the flexibility to be able to completely take this and show it to customers, get real-time feedback, so they're no longer worried about how to navigate. There is a fully interactive prototype. So in this situation, as soon as I built this, um, it was so funny, we had a stakeholder meeting in one meeting room. Thirty minutes later, we're with the customer actually demoing this because, you know, they wanted to see it. And-... what would have taken, again, like two, three months of discussions going back and forth, especially if they're a high-stakes customer, is now done almost immediately. And it does two things, right? Customers see it, they align and say, "You know what? This would be really helpful for me." They're no longer reviewing slides. They're no longer having to just listen to long talks about what a product could do. They're seeing what it's gonna be, and if they don't like something, they're right there telling us, "You know what? I think something could be different." And I think the outcome of all of this is, like, extremely fast alignment, which in my view, is kind of what is super critical now. As long as you're aligned with your stakeholders, with your customers, building anything else is easier, is more tangible, because there is no longer an uncertainty of, "Okay, let me put it into build. Will this work or not?" So that's what we do next, take it to customers, show them. Anybody on the team can do it, even our... I could trust my CEO to do it because, you know, it's a demo library. It's restricted to that. There's no page he has to go to where it's, like, suddenly-

    3. CV

      [laughs]

    4. AS

      ... going to give a 404. So all of it is very well-bounded, and this living demo library keeps growing. And, you know, um, it most importantly doesn't put any burden on the engineering team, which is what I love about being a CPTO. I can almost do my entire product function without worrying the engineering team, "Oh, please build this demo two days' time, one day time." [laughs]

    5. CV

      [laughs]

    6. AS

      "I just want this one button to work." It's all-

    7. CV

      Yep

    8. AS

      ... now gone.

    9. CV

      Well, and so I think this is worth repeating, because I think it's both very inspirational and also is going to give a whole bunch of teams a big heart attack. And so the thing that you've done here is you've taken not just this prototype, but basically your library of good ideas, fairly high-fidelity prototypes. You've put them in a kinda company-wide live demo library-

    10. AS

      Mm-hmm

    11. CV

      ... so anybody has access to these links, and then you're like, "Customer suc- success, go ahead. Like, sales, go ahead. CEOs, go ahead." And I know that you work for a relatively small company, and so that cultural change is a little easier for you, but I see so many product teams being like, "Oh, my God, don't t- don't tell customers we're working on this," or, "Don't show them the Figma. It's not gonna be like that." And I do think it's worth reconsidering how tightly we hold product ideas in the product organization, and I do think you're showing a different model, which is customers are understanding that things are in prototype phase, and they're-

    12. AS

      Right

    13. CV

      ... not live yet. Your team is intelligent and capable enough to manage customer expectations around things that may or may not come soon or later, and it's overall better for the product, it's overall better for the teams, and it's overall better for customers to just see what we're working on [chuckles] versus kinda keep it, keep it secret. And so I think the AI side of this is interesting because it becomes cheap to sort of show customer stuff, which I have also worked in the past where sales needs a demo of this or a prototype of that, and you're scrambling two or three engineers to just try to make something happen. That's not great. So now you've changed the cost of that, but then you've also changed the culture around just being more generous with showing your roadmap, showing your prototype, showing your things, things to customers. And so I think it's something that folks generally should consider is gonna become a shift in how product and sales teams work together.

    14. AS

      Mm-hmm. Yeah.

    15. CV

      And I think you're totally leading, leading the way here in terms of collaboration and openness.

    16. AS

      Yep, yep. Yeah, that's totally the product and sales collaboration, you know. Um, I had, uh, my, all my previous sales leaders always said, like, "If only you could show me what's possible, I can make this happen." So for the longest time, what was friction, it could easily now be converted to a partnership, and I think that's where AI is playing a big role. How can you reduce friction across different teams? How can you make the impossible possible? And, you know, um...

    17. CV

      Yeah, and one thing I wanna say, which is probably the fear in the back of product teams' minds or engineering teams' minds, is they say, "If I give this to sales and a customer likes it, I'm gonna be asked to build it." [chuckles] I'm like, "I'm gonna be stressed because all of a sudden we're gonna have demand." And I, as a product leader, I will tell you this, this is before AI, I have always said, "Wouldn't that be a great problem to have?" Like, wouldn't it be a great problem to have-

    18. AS

      Right

    19. CV

      ... that a customer wants to buy something so much for us that there's, like, demand on our time to build something very specific? I'm like, the most likely outcome of all of this is a customer looks at it and goes, "Eh, like, maybe." You know?

    20. AS

      Right.

    21. CV

      Like, they're like, "Eh, kind of."

    22. AS

      [chuckles]

    23. CV

      And I just, I think people are so afraid of preserving product energy and engineering energy, that they forget that those are like what I call rich people problems. Like, having too many customers ask you for too much product is, like-

    24. AS

      Right

    25. CV

      ... the best problem a product manager can have.

    26. AS

      Yep.

    27. CV

      And so I love this idea of just, like, getting in there and creating that problem for yourself, which is ultimately a good one.

    28. AS

      Right. I mean, I, like, it's almost like you would rather have this prototype and alignment, rather than randomly finding it on a quote that has already been-

    29. CV

      Yeah [chuckles]

    30. AS

      ... signed and executed. [chuckles] Right, so...

  12. 35:50 – 39:00

    Breaking deadlocks with engineering using Rork for mobile prototypes

    1. AS

      Yeah.

    2. CV

      We've all been there.

    3. AS

      Yeah.

    4. CV

      Okay, great. So you're, not only you as a product manager getting this in front of customers, but you are letting your team get this in front of customers. Um, so I think you've, you've managed stakeholders. It all sounds great. It sounds lovely to me, but we know there's still, there's still conflicts, there's still debates. So tell me a little bit more about how you're breaking deadlocks with, with AI.

    5. AS

      Yeah. So one, one other situation where, um, we, we, we ran into a particular challenge was an idea, um, again, that is where it required almost like a mobile app to be built.

    6. CV

      Oh.

    7. AS

      Um, and the whole idea there was-... we have this concept, could it work? And now, we have never done it for mobile. There is no mobile developer on the team, and now the team is stuck. Engineering is like, "Nope, I don't wanna even hear about this. Don't even tell me about this." [chuckles] And all we want to do is, okay, can we validate if it's possible, rather than, like, going hiring? That's the first step to do before we even start something. So what we did in this situation was, um, a particular use case again, um, to kind of also show that there is no limit to what could be possible, um, is where we went into building a mobile app. Um, this was an idea where we specifically wanted to be able to capture, and just a selfie of a person, and we are working with avatars, so we wanna be able to test Cappy, capture them in a happy way, in a sad way, in a normal way. And one of the big challenges we saw was enterprise customers or enterprise laptops always don't have the best webcam. So-

    8. CV

      Yeah

    9. AS

      ... now, how do we bring the solution to the customer with whatever they have, rather than shipping expensive kits? So we're like, "Okay, mobile phones are always great, the cameras are great. Now, how do we build something very quick?" So this is where, again, Rork comes into play. Um, I think, uh, they're an underdog in this game, or a lot of people are using it. Um, but when I found it, I was like, "Okay, this is like another frontier completely unlocked," because what was previously not possible was mobile apps with generative AI. So same prompt, right? Um, similar concept, started with ChatGPT. Um, a discussion in a coffee room where they're like, "Okay, what if we bring the mobile experience to the customers and have them just capture an avatar?" I'm like, "Sure," and then I go off. Ten minutes later, just one prompt here, we're able to now have a fully functional iOS app, um, or like an Expo app, that does only two things, right? This is not a production app again, but it does the main thing of building something that you could test on your phones now and shows what could be possible. And this is where a little bit of sensitiveness is critical, too, um, which requires understanding what engineering needs. We're not saying, "Engineering, here, I built it." I think that's not the right explanation to this or what we're doing here. It's more about, "It's possible. I think let's look at what is required to take this to production." So now this deadlock where there is a lack of knowledge of what could be possible is suddenly lifted to say, "Here's what could be possible. Could we explore more?"

  13. 39:00 – 42:34

    Takeaways for building with AI

    1. AS

      Right?

    2. CV

      Okay. [laughing]

    3. AS

      [chuckles]

    4. CV

      So I'm, I'm laughing because I... And I'm making this face. Usually, when I'm excited, I make this face, but I'm making, I'm making this face, which is I, I love this. This is like Claire Vo playbook. No, no lanes. Like, do whatever you want. "Sure, why don't we ship the CEO's idea?" [chuckles]

    5. AS

      [chuckles]

    6. CV

      Like, "Yes, let's show it to s- like, let's give it to sales." And everything you're recommending, I can see the flip side of, like, a product manager being like, "Why in the world would you give the CEO a prototype? Why in the world would you tell sales we can do that?" And then an engineer being like, "Great, this guy just, like, whipped up a mobile app and told me that he can code something that I can't code." And what I wanna call out here is, [lips smack] you know, your intentions, and, you know, some- someone like me that's really AI, trying to adopt AI, like, our intentions are good, which is there are sort of these, like, artificial rules and blocks that we in organizations have put into place. Like, you're supposed to argue with the CEO about the roadmap, you're supposed to never overpromise to sales, you're supposed to, like, trust that engineering will always tell you what's possible and what's not possible. And the reality is, like, we're all human, we're limited in cognitive capacity, we have good days, we have bad days, we have experience, we lack experience in places. And-

    7. AS

      [chuckles]

    8. CV

      ... there's just so much more access now to overcome those, like, misunderstandings, get more creative, be more inspired. And so, you know, what I would encourage folks to maybe take away from this, or what I'm taking away, is, one, there's a different way you can show up to your team, and it may, like, it may frazzle them a little bit, but I think on the net, it's pretty inspiring and very generative, and you can see how this is gonna deliver real vas- value for customers first. And then I would say on the other side of the table, let's be gentle, friends, which is like- [chuckles]

    9. AS

      Yes, yes

    10. CV

      ... I think people that are trying these new tools and trying these new technologies, you know, like you, maybe like me, are genuinely builders. Like, genuinely get excited about creating solutions, genuinely wanna solve problems, and so it's not about threatening people's expertise or, you know, trying to take their role or questioning them. It's really about, like, "I think this could be pretty cool to build for our customers. I think this would be great for our users, and I've found something that maybe can inspire us to do that." And so I think there's, like, a very important cultural takeaway from this episode that I don't want people to miss, in addition to, I gotta call out, um, I have not used Rork. So now you can AI prototype mobile apps. Did not know, now I know, um, to kind of figure out things that are possible with consumer devices. So I, I appreciate you bringing sort of both sides, the tool sides and the workflow sides, as well as the culture side to, to this conversation.

    11. AS

      Yeah, definitely. I think the, the one word I would use is, um, to echo what you're saying, is alignment. The goal here is not to say something is easy, it's more so to, are we all aligned and can we move fast? Um, which I think is a, is, is the next superpower as all these tools come out and what could be possible, right? And, uh, to, you know, all the big unicorns today that are coming out are small teams that are moving extremely fast, and I think we're all seeing living proof of that with everything that's going on. So yes, [chuckles]

    12. CV

      Yeah

    13. AS

      ... nothing to displace good old production-resilient code and processes, but this is all before that.

  14. 42:34 – 45:20

    Cultural implications of AI in product development

    1. CV

      ... Well, and I'll also say, I mean, annoying PMs have been doing this from the beginning of time, [chuckles] which is-

    2. AS

      [laughs]

    3. CV

      ... I have been, I have been guilty of being told no and saying, like, "I set up a, an, an Excel script that, like, kinda did the thing that I want," or, "I set up, like, a Zap that is basically the flow I want. Can we..." And at least now you get code. I mean, you know, you get code and something to click. So, you know-

    4. AS

      Right

    5. CV

      ... we will forever, perpetually, as product managers, find very annoying ways to ask engineering to build too-complicated things and scope creep. It will just happen till the end of time, and now we just have [chuckles] more tools.

    6. AS

      [laughs]

    7. CV

      Okay, so let's, let's stay on this topic. Let's get to lightning questions. We've been here a while. You know, zooming, zooming way out, I think the big takeaway from, for me, from this conversation is roles are changing, and the way companies work is changing. So I'm just curious, sort of what are your meta, you know, predictions on not how is this gonna change for an individual product manager, but really how are companies gonna start to operate differently when they have access to these tools?

    8. AS

      I, I think companies are going to figure out a process that helps them validate things faster and move faster very much early on. I think whatever used to be the bottleneck of, "Let's take time, let's wait for something to come back, let's wait for more analysis to be done," is going to shrink. Which I think in, in turn is really good, because you don't want confusion after you ship a product, which is what causes the most issues, right? A good PM is now stuck managing disconnects rather than what should have gone out to be successful after spending all the time. So I think companies are going to evolve to where they're spending more time with these tools up front, and then being very sure of what they want as they move into building. Um, alignment is going to be faster. I think these prototypes, from a perspective of aligning all the different departments, I think about also customer support, right? Enterprise applications, you wanna be able to support your customers. You have a customer support team. Rather than them reading booklets of information and preparing manual guides, they can now have a prototype. You could explain it to them with that. They are then training, like, hundred, hundreds of team members, like, "Oh, this is how you would support if something happens." This is, like, super valuable for large organizations. They don't have to wait until a product is done to be able to train. Now they can train on the fly, and you're no longer using obsolete technology, right? You're always staying cutting edge, which could be a big differentiator for a lot of large companies, too. So I think that's, that's where I see companies evolving, alignment fast, and then being able to ship, uh, with a sh- ship with, I would say, precision after that.

    9. CV

      Okay. I lo- I love it. It's like I'm smiling ear to ear. Let's move faster, let's build better things. Let's enable,

  15. 45:20 – 48:32

    Strategies for when AI doesn’t give you what you want

    1. CV

      you know, our peers in the company a little earlier, um, have them more prepared and give our customers a better experience. Okay, and then you've showed us nothing but perfect outputs, so I've seen a perfect Lovable prototype.

    2. AS

      [chuckles]

    3. CV

      I've seen a perfect Perplexity analysis, Gamma presentation, a chef's kiss, uh, ChatPRD analysis. I've seen all these beautiful things. Please tell me AI sometimes steers you wrong.

    4. AS

      Oh!

    5. CV

      [chuckles] And so wh- what do you do when AI is not giving you, giving you what you want?

    6. AS

      I think, one, take a break. Um, [chuckles] sometimes-

    7. CV

      [laughs]

    8. AS

      ... it's like, you just need to take a break and stop, you know, prompting until the end of credits. Um, so AI could be wrong. I mean, for, you know, the one demo that I showed, I have at least three demos that did not work, um, in terms of building when we're building something to perfection. That's where initially I was doing all of this inside v0 or Lovable, and then I realized, "Okay, I'm spending way too much time iterating inside the tool, which should not be." So I think naturally found ChatGPT to be the first starting point. So when AI doesn't work, you take a step back and use something else more broader. Two, be patient, because, um, AI can make mistakes. It's still a developing field, but the benefits, you know, far outweigh, um, the failures.

    9. CV

      Yeah, and we were, we were joking a little bit before the show, saying that how people answer that question is often a reflection of their parenting strategy.

    10. AS

      [laughs]

    11. CV

      And so I like the idea... Honestly, I need to hear this. Sometimes, when your, like, kid is not doing what you want, like-

    12. AS

      Right

    13. CV

      ... just walk away is usually- [chuckles]

    14. AS

      Yeah

    15. CV

      ... is usually the answer. You're probably not gonna win the argument with the, the five-year-old. Uh-

    16. AS

      Yep, yep

    17. CV

      ... so sometimes you just gotta go, "We're gonna take, we're gonna take a little break."

    18. AS

      Right.

    19. CV

      And so I'm gonna, I'm gonna take that into my, into my AI strategies as well. Well, this has been such a great episode. I've just really enjoyed the conversation top to bottom. Where can we find you, and how can we be helpful?

    20. AS

      Yeah. I'm on LinkedIn, um, Twitter as well, um, or X. I'm on both platforms. Um, LinkedIn is the best place to reach. Um, I'll leave my email, um, here, and, uh, happy to also talk, you know, um, if anybody needs insights on how to advocate for more AI in your product or technology workflows, [chuckles] happy to always support with that. Um, but, you know, I think, uh, we as a community, uh, need to step together to do this. Um, and not, not overselling your tool, but, like, the ChatPRD community or Lenny's community is where I've had very meaningful discussions. You know, I've had a lot of PMs talk about what could be done different, and I think, uh, yeah, that's where you can find me.

    21. CV

      Yeah.

    22. AS

      Mm-hmm.

    23. CV

      Okay, so find you on LinkedIn, and let's all chat about how we manage this transition together. Well, thank you so much. I really appreciate this. PMs, this is one that I hope you've listened to the very end. Thank you so much. I'll see you later.

    24. AS

      Thank you.

    25. CV

      [upbeat music] 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! [upbeat music]

Episode duration: 48:32

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode shqv90oAIkM

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome