What Does Product Leadership Look Like Inside an AI Company?

Watch the Episode

Transcript

Shobhit: Ready to move to the next level in your product career? I'm Shobhit from Intentional Product Manager. Join me as we discuss ways to help you stand out in your job search and your career so you can have more impact and make more money. Hey everyone, welcome back to the Intentional Product Manager podcast. I'm your host Shobhit Chugh, founder of Intentional Product Manager, where we help product managers step into true product leadership and build careers that they are proud of. Now today's guest is someone who I've admired for a long time. Karan Gupta is the CEO of Alice AI, a voice intelligent platform pushing the boundaries of secure enterprise ready AI. So he steered teams through IPOs, M&As, acquisitions and he's building in one of the most exciting frontiers in tech, the intersection of AI product and real world business value. Karan, welcome and I'm very thrilled to have you on the podcast.

Karan: Hey thanks for having me on. This is great.

Shobhit: Well, why don't we start with some quick fire ice breakers to get to know you beyond just what it says on your LinkedIn bio. Tell me one book that you think every product leader should read.

Karan: I was on a call yesterday. I recommended at least four marketing books to that person. There's a few really good ones, but I will name one. There's a lot of stuff by Seth Godin, stuff by Cialdini — it teaches you a lot about human emotions, human behavior, teaches you about marketing. So I think not just product managers but most leaders should read those books. One book which I do read every probably every couple of years is "What They Don't Teach You at Harvard Business School." It's by Mark McCormack and it's a really great book. It's about old school business. The principles still apply. There's a lot of strategy and there's like I don't know something like 150 case studies shoved into one tiny book. You know the difference is all those case studies are small, just like three paragraphs gets the point across.

Shobhit: So I know you just came back from a trip from Singapore and you do a lot of work across the world. Tell me one habit that helps you lead across cultures and different continents.

Karan: This trip was great by the way. It's always like a pleasure to work with people who have different backgrounds, different cultures, and ultimately you realize people are just people. People have the same concerns, they have the same responsibilities, for the most part — grown-ups, I think. So I've been able to hire people all the way across India, even Singapore, in Europe, Ukraine, South America, Canada, of course many parts of the US. One thing I realized is not everyone kind of follows the same schedule — people have different holidays, different festivals, different things going on in their families, their personal life. You have to give room to people to live their life and find a way to balance work and life together, so they have their own harmony, their own rhythm. But that being said, we put some systems in place. Sometimes those systems are very formal — for example, goal setting sheets, spreadsheets. I know companies use OKRs and we've typically used the simplest systems possible, which is like talk to everyone in the same language about what's needed to be achieved and by when, all the way from the senior-most role to the junior-most role. And then you treat people like they're grown-ups. We've given a set of boundaries, a set of guardrails, and now there's a lot of freedom, a lot of autonomy. We expect you to deliver your best — they're not going to micromanage you. So people then kind of have to find a balance between all the stuff going on in their particular lifestyle and their particular culture and what's expected by the company. And so you end up having the same yardstick for everybody no matter where you are, and the same terminology across cultures — people are talking about deliverables in the same way, timelines in the same way. If people are in a tough situation, they're able to say "okay, this is on my goals, so I have to deliver this." So that's just one example of a system, but I think those are the kinds of things that have helped us for sure.

It's something so top of mind — everybody talks about global things, but it's great to see the learnings from a real world experience that you're bringing in. That being said, I got to mention something: I love being with people in the same room. There's that level of energy — you cannot get that working completely remote all the time. So one way, when people are all remote and spread out, is to have at least one, if not two or three, occasions in the year where everybody can meet and get together, or at least partial groups, because those connections you make over a one-week period really give you the energy to go through many months of being in a remote location. So if you're going to be remote, not spend on an office and be all distributed, I think you have to park aside some funds for these get-togethers.

Shobhit: Yeah, one of my managers highly advised that early on when you start working with a team, make sure you meet them in person — that would be totally beneficial. Let's talk about one more topic. What is one way you personally use AI every single day?

Karan: What do I not use it for? I use it for a lot of things. Sometimes it's like my chief of staff, all the way to sometimes I'm literally just asking for recipes for pastas. I kind of went down that hole of using it for recipes because trying to use a recipe website these days is a total nightmare — the number of ads just means you can't really use it. Besides that, for work, a lot of things — just the other day there was a very long news article which I used, you know, I use Claude most of the time, to just digest and get to something I can absorb in a few bullet points. I rarely use it to generate content for me, because most of the time it generates so much verbose content that I have to strip it down anyway. But I might give it something and say "clean it up, shrink it, make it better," versus just giving it a straight-shot "give me XYZ document" — I don't really do that. And then coding — I mean it's a game changer when it comes to coding. My company uses AI for coding. Everybody has an AI software engineer.

Shobhit: It's sort of interesting that you say you're not using it for the use case that most people love to use it for, and many of the opposite ones — synthesizing things and using it that way.

Karan: I think what I realized is I try to write as little as possible, just use a few bullet points to get the point across, and AI tends to write a lot — even if you change the filters and stuff, it tends to just be overly formal.

Shobhit: Let's now transition. Let's start to talk a bit about your career, because it's really fascinating — all the way from Sony to Alice AI. You've also been part of two IPO journeys, you founded your own company. Tell me what first drew you into product and engineering leadership in the first place.

Karan: I remember the technology bug. I was still pretty young, maybe 8th grade or something like that, and I was going towards bio, on my way towards medicine — definitely very attractive to me. And then I got introduced to computers for the first time and I was like, oh my god, this is amazing, because the instant gratification of creating something and just seeing it live, you don't get in a lot of careers, right — something complex. Then I think the product leadership and the engineering leadership kind of just came organically after that. I got into a couple of companies where we did not have those systems in place around product leadership, and I just found myself naturally trying to fill the void — doing business analysis, trying to understand the market, trying to look at different tools — and because I was an engineer, I was actually able to connect what I was finding, like business needs, to what we should actually make, what is doable, and what the level of effort is. So yeah, the first two or three companies in my career, including Sony, I ended up playing the product role quite a lot.

Shobhit: The entrepreneurial bug — let's talk about that. When did you start to realize that you're ready to lead companies and you want to start something of your own, not just leading teams?

Karan: I think that's been since a very young age. I started my first company when I was 26, just settling down in the US. I did not care that immigration laws or visa laws would disallow it — I just wanted to see if we could make something. And that was my first real touch with starting something completely from scratch and seeing it to an exit — a minor exit, but an exit. After that, it just became fun, something I knew that I could do, and it became something I wasn't afraid of anymore. When I started my next company we still made a ton of mistakes, but I also knew that this is something people do around the world all the time, and it's something I enjoy — I enjoy moving fast, I enjoy making decisions, I enjoy connecting with customers directly. So it just naturally fit into the role of "you've got to start something to be able to do all of those things."

Shobhit: You've gone through large companies, scale-ups, startups — what's been some common threads or things that stayed the same?

Karan: You always underestimate how difficult it's going to be. You always underestimate how long it's going to take. And people are going to be people — they're not going to show up, or they're going to show up in different ways, sometimes they're going to be slow, sometimes they're going to be super excited and move faster than you need them to, sometimes they're going to go down a rabbit hole because they're so driven by a particular idea that's got them. Because people are going to be people. Big or small, no matter how many resources you have in the company, those things are kind of relatively common. End of the day, you have to have a deep focus on the customer. There was a period of time in my career where I used to listen to a lot of Jeff Bezos videos and things he's spoken about — his leadership principles, customer obsession as a key tenet of Amazon. I don't think it's unique to Amazon, but I think they defined it so well that it's easy to absorb and easy to follow as a company. So I think as long as you're doing that, as long as you're focusing on the customer, things are going to be okay.

Shobhit: Let's now come to the present — tell us the story behind Alice.ai. What's Alice?

Karan: Alice is a secure voice intelligence company. I started thinking about voice intelligence actually many, many years ago — I built a first speech-based product in my undergrad, so quite a few years ago. It's not something that just came new to me or was just in the air and I thought "this is something worth trying because it's hot." It basically started with honestly trying to make a better recorder. I was getting all of these ideas while I was driving long commutes in the Bay Area — I had just sold my company, my brain was suddenly free, and I was seeing problems everywhere and wanted to capture these ideas as I was having them. I tried Apple's voice memos, I tried recorders, a whole bunch of apps that were available, and I always found something lacking. So I thought, you know what, we can just do better — these products all look the same. I spoke with a lot of customers, thought there's got to be something better out there, spoke with journalists, engineers, sales, lawyers, people in many different fields, and realized no, something is missing. So the company actually started from that idea of "let's make a better recorder" and then "let's make it extremely secure so that people can actually talk freely" — because that was something I learned in customer interviews, that people have access to recorders, they've had it for decades, but they don't use it because they're afraid someone might be listening, there's a bit of implied mistrust of the whole ecosystem. As we kept talking to more customers, we realized we need to put high-accuracy AI, we need even more security guardrails, you need order tracking. And it very naturally became this enterprise voice intelligence company. Now we have two or three products that people across businesses use all the time, and we still keep talking to customers — some of them we just know by first name. That's what we do.

Shobhit: I just watched a YouTube spoof the other day where a woman tells her husband, "I would love to talk about this, but I'm afraid Alexa might be listening." Her husband laughed, then Siri laughed, then her Tesla laughed. I mean, someone is listening, right?

Karan: Yeah. I don't think most of these software companies are trying to be malevolent, but I think it's just how companies are wired — technology just gets built a certain way, to track everything you're doing, every behavior, every action, every browser, every device, and then never ever delete that data. It's just wired into how tech companies are built. So we ended up going a very different extreme — what can we actually build without tracking anything, without giving access to any third party. We don't use Google Analytics, we don't use tracking pixels, we don't connect with TikTok or X or anything like that. Most companies have 80, 100-plus vendors attached to their software — we have like five, and it's been difficult to build that. But now everything is in one system, it's easy to extract information, it's easy to answer questions, it's easier to build products on top of all of this information that we have. And so we end up giving a lot of value to the customer, and the system's easier for us to manage.

Shobhit: You mentioned you started with this interesting insight — tell me more, what are some of the misconceptions you came across as you started building voice-intelligent products? You mentioned one about privacy — any others?

Karan: I think there is a certain assumption that speech-to-text is everywhere — that you can even do it in Microsoft Teams, record a meeting or something like that — and when we talk to customers coming to us, we realize they tell us, one, even though their company has sanctioned Teams and they discuss a lot of things in Teams, they're very scared of actually speaking and getting a Teams meeting recorded. Number two, even if they get over that hump, the quality of the transcript is pretty poor, so they're like, okay, we need a better solution if we're going to actually use this technology that's available now. So what I'm trying to say is it's not that difficult to be 80% good, maybe 85% good — it is very, very hard to be 99% good, crossing that final leap. Now, that being said, speech-to-text has been around for years and years — dictation, Dragon software — it's not new, it's more that the technology is becoming more widely available and much better. I think the real value is how you apply that technology, just like other technologies that have existed — databases have existed for a very long time, and when you started building web apps on top of them, suddenly they grew in popularity because you needed them to maintain state. I think that'll happen with speech-to-text technology too — what applications do you build on top of it. That's what we've done with our newest product, which is called Voice Box, a feedback inbox for retailers to collect the voice of the customer. It's very straightforward, extremely simple, works on every device, extremely accurate, and the reason the intelligence is so good is because the accuracy is so good — if you're upstream, the quality of the recording is high, the quality of transcribing is high, everything is secure, then downstream, building products becomes easier and better and higher quality. It seems to be a space where people would just assume "yeah, we've seen this, we've done it like this," but there's so much possibility that I think those are the kinds of spaces people miss pretty often.

Shobhit: Let's bring it back to trust. You mentioned user privacy, but this is a completely different ballgame when you're talking about earning trust with large enterprises. What's the hardest part of earning trust with large enterprises when building with AI?

Karan: I mean, you've probably heard the story about how they chain up baby elephants, and then even when they become really big, they feel like they're still chained even though they have the strength to just break it and walk away. I think there's a lot of distrust that has been built over the years, and so even though a CIO might say "we're going to use XYZ software" or "the entire company will use Microsoft or Google," and the CIO's direct team has done a lot of research on the security aspects of it, day-to-day people still have that underlying fear — either of the corporation or of big tech — that someone is listening, or it's not safe, or if I say something or type something it's just going to be there forever, it can be misconstrued. And now there's a new fear that it can be used by AI to publish something in a totally different context. A lot of these fears are unfounded because I don't think many of these companies, or the product leaders or technology leaders, are trying to be malevolent — they're not, they're people like you and me, just trying to do a good job. But somehow this does happen — data does get retained forever, data does get abused, data does get leaked, and there are instances where, say, ChatGPT is publishing something it read from somewhere else. So I think the biggest thing when it comes to establishing trust is we have to really show these companies and the executives we work with that we are different, and we have to literally talk a lot about our technology and architecture and explain that. We tell them we built this from the start for companies like them, which really need extreme privacy. One phrase I like to use is that our infrastructure is near air-gapped — which means most of our infrastructure has never connected to other third parties on the internet, and that's how it is safe. You cannot retrofit privacy, you have to build it from the beginning, and luckily we did that. But we still have to keep educating people and really explain to them our reasons and how we started.

Shobhit: Karan, let's now switch a bit and talk about product managers, the topic of the podcast and something everyone's interested in. I noticed a post by you about product managers that really piqued my interest. You've built organizations on almost every continent and hired product managers for seed-stage startups and post-IPO companies. If you were to say the top few signals you look at when you're hiring a product manager today, what would those be?

Karan: I think by far the number one thing for me — and I was saying this morning — is being able to talk to customers. I want to know, as a product leader, the last time you spoke with a customer. What did you talk about? How long did you talk? How often do you actually talk to customers? Do you use email, the phone, do you get on Zoom calls, do you go visit people — depending on the main side of the business. There's a couple of other ones — I want to make sure a product leader actually knows KPIs and is able to talk in the language of KPIs, knows industry terms, but more than that, actually knows they have to measure their success, their product success, in terms of real metrics. When I'm talking to people about their professional growth, that's what I'm asking — "hey, okay, what metric did you move?" And sometimes, surprisingly, people don't even remember their metrics. I'm like, you should remember the metrics — I'm not saying remember 25 metrics, just remember two or three key metrics in your career that you know you helped move. Because then that shows they are data-driven and customer-driven. The final thing I guess I look for in everybody, not just product managers, is just humility — know that even though you're a leader or a manager, you work in a team, and everybody in the team has a role to play, no matter how senior, no matter how junior — just have some humility that everyone brings something to the table.

Shobhit: We talked about hiring — let's translate that. When you've hired someone and you want to coach them, what's your coaching philosophy, and how do you grow those product managers?

Karan: Most of the time, honestly, what I've seen is people get so caught up in the drama of building a company or building a product that they forget to actually build the product. Having too much process has been one of my pet peeves — I'm like, reduce meetings, reduce the number of steps it takes from "we heard this from the customer" to "now it's delivered," or "we saw this in the data, this is the thing we should solve" to "now it's delivered." Just reduce as many steps as possible. So when I talk to product managers about performance and career growth, that's one of the things we talk about — what are you spending a day on? And if they tell me they're spending their day on, I don't know, six one-on-ones and writing up two very large documents, I'm like, you're just wasting your day, you don't need to spend so much time doing these things. If you're delivering, there's going to be excitement in the team, customers are going to be happier, the business is going to do better. End of the day, that is the goal — that's why we come to work, to have an impact, to bring value to customers, to bring value to the business. So when people get caught up in all this other drama around process and how we do things, just cut it out, put some simple systems in place, and then cut out the rest. But it is surprisingly hard. Surprisingly hard.

Shobhit: Sorry, you want to say something? I was just going to bring back the point I feel like I keep harkening about — talking to customers. When did you last talk to a customer? Any medium, I don't care — talk to customers, and take a couple of engineers with you on that call.

Karan: I love that — for me, trying to be the go-between never worked. But when I brought the engineers with me and they had that firsthand experience, everybody was energized to do something about it. Their eyes light up when they hear things directly from a customer — this might be things they've been debating about internally for like two weeks, and then they hear the customer actually say their reasons for wanting a particular feature, or needing a particular problem solved.

Shobhit: Let's combine some of the things we were talking about — AI and product teams. I saw your LinkedIn post which is about "AI won't kill product managers, it will create more product manager jobs" — a bold and refreshing take, especially with a lot of doom and gloom. What inspired you to write that post?

Karan: I think it was partly what I've seen at Alice and partly what I saw in previous companies. The typical ratio I saw in previous companies was maybe one is to six, sometimes one is to four, sometimes one is to ten — one product manager for that many engineers, maybe one designer, one QA engineer thrown in. What I realized is the way we're working now, you give an engineer a spec — even though they might have been really fast pre-AI, now it's crazy, the velocity at which software can be delivered is insane. But one of the concerns is if you don't have very good specs, you'll deliver a lot of crappy software, a lot of hacky features, without holistic thought applied to them. I'm not saying every engineer needs a product manager, that's not true, but I do think the combo of a business person and a technical person is great — you're basically dividing up the work of what it takes to build something. One person is talking to customers, looking at the metrics, looking at the impact; the other person is building the technology. Add a designer in the mix and the three of them can deliver a lot more than was possible before, when you'd have a sprint meeting every week or two weeks, sometimes even once a month, and then you'd go away and do your own research and tinkering, and after a couple of weeks some work is ready to be delivered. Now the speed is insane, so you kind of need both parties working hand in hand.

Shobhit: Let's explore a few different things — what does it mean for product managers in terms of the skills they need to bring, how they need to show up, any changes there?

Karan: Oh yeah, exactly right, because if the team is able to move so fast, then the business part of the team also needs to move fast — that luxury of having weeks to decide what to build, I mean you still have that, but you don't have it for everything, for every feature or every requirement of the business. So I think you have to become better at extracting data, become self-sufficient. Tactically speaking, I think product managers should know enough SQL to put into ChatGPT or Claude, whatever tools they use, to really extract information from databases — or if you're using Google Analytics or Amplitude or something, you need to master those tools yourself, because if you're going to depend on a BI team, a data analysis team, an ETL team, you'll basically have engineers sitting for three, four days while you try to figure things out. So I think you just need to become really fast — being KPI-driven, talking to customers, quickly designing stuff, turning on a dime — those things tactically become important now. Gone is the day of writing, you know, how Amazon famously has six, seven page PR/FAQs — I'm quite sure, and I know because I have a few friends at Amazon, a lot of them are now being written by AI, and then they're also being read by AI on both sides. So I'm like, just become faster — put down bullet points, simple language, so that the other side can also just read those bullet points, simple language.

Shobhit: So how do we hire, structure, and support these product managers through this shift?

Karan: I think part of the coaching has to be — okay, the strategic part of business, I think fundamentally that still stays, how you do strategy, how you think about the long term and business models, and that has probably stayed consistent over years and years of companies being built. But what is changing is the tactics. So I think a lot of tactical education has to now happen — around extracting data, but also actually becoming a lot more technical. You should know how a design team is working in Figma, how they're able to move designs from Figma to code, and maybe there's something you can do yourself as well. With so many coding tools available, if you need to massage certain data to see certain patterns, to extract certain value, maybe you just need a script, and you don't need an engineer to spend a day figuring it out, because you'll spend half the day telling them what you need and another half iterating on changes. But if you were able to learn some of these things yourself, you might just solve the problem yourself. And that doesn't mean you need to install an IDE, doesn't mean you need to become a coder — it just means most good product managers are already a little technical, it just means become a bit more technical.

Shobhit: It seems technical, like becoming a super user of AI in some sense.

Karan: Yeah, exactly. I don't think you need to learn how to read code or write code, but you need to know how to talk to AI that does that for you.

Shobhit: I do want to address this point — a lot of product managers who've been product managers for a while are feeling a little anxious, nervous about being left behind by AI. What would be your advice to them?

Karan: Look, it's an exciting time for sure — it's like the discovery of fire or something, it's pretty crazy what's happening right now. ChatGPT is the fastest growing software ever — even people I know who are not very technical, not adopters of technology, have been using ChatGPT for months, it is everywhere. And ChatGPT is one example — now it's showing up in Google, there's Perplexity, there's Claude, all these things — it's everywhere. So like it or not, anxious or not, this is happening. I think product manager, engineer, anyone, you have to accept that yes, this is happening, and your career, and probably your next generation's career, is going to be very different from the careers we've seen so far. So what I would encourage is, actually go and learn stuff, become a super user of AI, so that you bring a lot of value to the company where you work — whether it's a job or a company you start — so that you can move fast, so that you can just deliver actual, commercial value.

Shobhit: The thing I took away from that was acceptance — accept things are changing and then act toward it, rather than feeling nervous and getting behind because of that.

Karan: Yeah. Or worst case scenario, just bury your head in the sand and continue to do things the old way. Things are going to change whether you like them or not. It's happening.

Shobhit: As we come towards the end, I want to wrap up with a lightning round — just say what comes to mind, don't overthink it, one-sentence questions. What's the most underappreciated skill in product managers today?

Karan: The amount of crap that they have to take from so many different parties. I think they are very good at dealing with it, from senior executives all the way to everybody else complaining and whining to them.

Shobhit: What would be an interview question you would ask every product management candidate?

Karan: Oh, when is the last time you spoke with a customer? And then of course you dig deeper — what did you talk about, how did you do it, how did you schedule the call, all those things.

Shobhit: What's a belief about product that you've completely changed your mind on?

Karan: I used to think that maybe you only need business analysis, something very basic. And then I learned along the way that no, of course it's a much more involved skill. But I also realized there are a lot of, just like in every field, a lot of average PMs who are maybe more like project managers. A really great PM can just really change a company.

Shobhit: What's one thing product managers get wrong about working with engineering?

Karan: When it comes to timelines and difficulties — I once had a product leader who said, "Why is this going to take so long? Isn't it just some HTML that you have to put together?" I was like, are you crazy? That's not how it works — you have to think about so many different things. Thankfully it's not something I've heard often, but I do see there's a general clash around timings of deliverables and not understanding why things take as much time as they take.

Shobhit: What's the most interesting AI use case you've seen recently?

Karan: I've seen it talked about — I still need to prove it actually works, I haven't seen it work myself — that an AI agent can scan the internet, find a whole bunch of leads for you, automatically figure out what to email them, email them, and then they will respond, and it will respond to the response and schedule a meeting. That's crazy — a lot of things an AI agent is supposedly doing. I've heard about it enough that it feels true, I just haven't seen it in action, so I'm really curious to figure that out.

Shobhit: Last but not least, how do people get to learn more from you and follow you?

Karan: I don't post a lot, and honestly this has been great, to be here and chat with you — it's also a chance for me to reflect on some things. But when I do post, I'm posting on LinkedIn. I have a presence on X but not as much. So mostly on LinkedIn — my handle is Karan MG pretty much everywhere.

Shobhit: I learned a lot, this was packed with insight. Thank you for sharing your thoughts, your experience, your forward-thinking take on how product leadership is evolving in the AI-first world, and tell us more how people can learn more and try out Alice AI.

Karan: Yeah, of course. Look, if you're trying to get feedback from your customers, you really got to hear what they have in mind — so we've created a very easy, fantastic tool for you to use. Just go to vbx.to — that's Voice Box, vbx.to, it's the simplest link. Sign up, we have a team, we'll help you get started, onboard you and everything — a bunch of companies are using it now. For Alice, head to alisapp.ai — it's all one and the same company. And for me, like I said, just find me on LinkedIn, hit me up, easy to get connected.

Shobhit: Karan, first of all, thank you so much for taking the time for this podcast. I really appreciate it.

Karan: I really appreciate it too. I think it's great that we've been connected, and I think your podcast is great, and people get a lot of value from it. So thanks for having me on.

Shobhit: And for all the listeners, if you're a product manager looking to step up as a strategic leader, you know where to find us — Intentional Product Manager, let's build your career, your influence, your impact. And hey, last thing I'd ask of you, make sure you leave a review and share this podcast with others — whatever platform you're listening on, YouTube, Spotify, Apple Podcasts, please leave us a review. Thank you and see you next time. Be sure to check out our website at intentionalproductmanager.com to see how you can level up in your career.