Product Rebels
Product Rebels
Product Leadership in the AI Era
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
What does it mean to be a true product rebel in the age of AI?
Heather Samarin and Vidya Dinamani sit down with Raviv Levi, former Chief Product & Technology Officer at Sift, as they unpack how product leaders can stay focused on what really matters—amid constant noise, data overload, and accelerating change.
From building alignment across competing teams to rethinking experimentation in an AI-first world, this conversation explores how faster prototyping, sharper problem focus, and stronger conviction are redefining product leadership.
I think that confronting people directly through deep care for them, for the business, and for the product yields great results. And I've seen that happen so many times. So that is definitely something that I would suggest people to give a try. Don't beat around the bush, go and confront the topic directly. But building on what we just talked about, you can do that if you don't have a relationship with our person. And you cannot do that if you don't have alignment. Once you have that, then you can move faster, right? You can really get to the topics and debate that the best way to get to the crux of things, to get to the kernel of truth that we care about is through debate and brainstorming. And those things have to happen in a safe and trusted environment that is built again on relationships and alignment.
SPEAKER_01Hey Product Rebels, welcome back to another episode of the podcast. Today we're joined by Raviv Levi, a product and technology leader who has spent over two decades building and scaling security and infrastructure platforms from startup to multi-billion dollar business units. Raviv spent nearly a decade at Cisco building out the product management organization and helping grow the networking and security portfolio to over$2 billion, and then led Cisco's broader cloud security group, a 750 global organization. Most recently, he served as chief product officer at SIFT, the AI-powered fraud prevention platform, where he unified product and engineering into a single operating model and drove major platform investments in data and AI. We're excited to dig into his story and hear what he's learned and about driving real change inside organizations of every size. Welcome, Raveev.
SPEAKER_00Raveev, we're so excited to talk with you today. Welcome to our podcast.
SPEAKER_02Thank you so much for having me.
SPEAKER_00So we're going to start the way we always love to start, which is throwing into the deep end and asking you, what does being a product rebel mean to you?
SPEAKER_02This is such an interesting term. And when you asked me for the first time when I got introduced to what you guys were doing, I thought to myself, huh, this is a very interesting way to frame individuals working through product organizations and businesses. I think that to me, product rebel really means the person that is not necessarily going against the business, but the person that is advocating for the business against the daily grind. And I'm sure the listeners and you guys have seen that as well. Throughout our careers, it is so easy to get lost in the daily grind and to chase the Jira tickets and to chase the escalations and to chase the OKRs. But if you're really a product rebel, then you stop and think, where do I want to go? What markets do I want to disrupt? What problems do I commit to solve for my customers and stay true to that? Easier said than done, for sure. And that's why I think that's a good rebellion to be made.
SPEAKER_01I love it. Love it. Can you take us through a good example of that where you felt like a rebel and sort of went against the grain, but for the good of the company?
SPEAKER_02Yeah, I have a lot across my career. So very briefly, most of my last decade I've spent with Cisco. I had a variety of roles there. And one of my first positions there was with a business unit called Meraki that came out of an acquisition. So the business unit, the company was kept to the side, been given some independence, worked on some product lines that had some overlap with what we used to call the traditional Cisco would be offering those days. And pretty soon after I joined, I realized that for the product and the business that I was running, specifically network security, there was another business unit working on some pretty interesting technologies that we could use. But the thinking on their side was hey, we're Cisco. We know everything that there is to know about security and network security. Those new guys that we just acquired, we acquired them. So they're not gonna tell us what to do. And vice versa, on the Meraki side, we had a top-notch team of the smartest people I have ever met in my career that were just recently acquired for 10 figures for what they're able to do. And the thinking there was well, obviously we know what needs to be done, and we're gonna teach everyone and just build it ourselves. Just let us build the future because we know best. You just paid 10 figures for that. And for me, understanding where the company needs to go as a whole, it was very clear to me that hey, I don't have enough people, I don't have enough expertise to go out there and face the jogger notes of the industry. And I need to leverage those other people that have been spending decades in perfecting their network security capabilities, in understanding customer problems, in understanding the industry and all of that. Easier said than done. I had to work very hard on building those bridges, building those relationships with people across Cisco, ending up being the ambassador of both business units going back and forth. But the end of the story was that we got alignment and we got additional budgets and we got shared resources, which led to huge market success against those joggernauts that we were so afraid of previously. So not easy, but definitely a journey worth taking.
SPEAKER_00Okay, so I think you just set up a scenario that is so familiar to so many of us. If you're the acquiring company or you're the acquired company, and then there's alignment. So let's go to that messy middle because you've set this up where you've got two groups with a very specific point of view, and you talked about building bridges. Take us through what that looked like. What were the first steps? Where were some of the issues? What did you do? Again, because so many of us have been in exactly the same position and will continue to be.
SPEAKER_02Yeah, I'm a very people-oriented person when it comes to my teams and my employees as well as my customers. And I'm a strong believer that there are almost no good people or bad people. There are people in certain circumstances that are trying to optimize for certain things. For me, it's about going back to first principles and starting with the people and having a conversation with those people on the other side. Hey, tell me about yourself. Tell me about your business. What are you optimizing for? What is the problem that you're trying to solve? What is the impact for Cisco at the time or putting your own company there? Understand what is the thing that they're trying to solve for. Then things become more and more clear. Either you realize, hey, we're actually trying to solve for the same thing. Here's what I can bring to the table, how I can help. Tell me about what you're doing. In many cases where I've seen disagreements, they were solved when I had the people involved in the same room being exposed to the same data. Because brilliant people tend to come to similar conclusions when exposed to the same data set. The gaps happen when different people are exposed to different subsets of the same data, and they're just trying to do their best given the information that they have. Right. So if I go back to what is the first step that needs to be taken, this is it. Form a relationship with the person on the other side and align on what we're trying to do. It doesn't mean convince them. Maybe you're gonna get convinced, right? But you're gonna have an understanding of what we need to do.
SPEAKER_01Fantastic. What a great summary of leadership, right? And this is something that we all want to have our product managers strive for, right? This is not easy. And we use a term called shared vision, right? You may not agree, but you understand how we got to the decision. You've all been exposed to the same data and you're gonna commit. And that's shared vision. And just really love that summary. So, Riviv, there's certain traits and skill sets of a person leader in this case that enables those discussions to happen. Talk to us a little bit about what enables you to do that well in a way that helps our up-and-coming leaders and existing product leaders that may be having some challenges. What are some traits or skills and maybe practices that you use to get through some of these challenging times in driving for shared vision that you can expose to others here?
SPEAKER_02This is extremely subjective. I learned a lot from my mentors and other people that have influenced me throughout my career. So there's a couple of things that come to my mind, but it could be different for other people. I would take it with a grain of salt. But as my accent would suggest, I'm coming from Israel. We have a culture over there that is very, very straightforward. Lots of good intentions, but very cut to the chase, straightforward. And I needed to make some adjustments, right? I needed to have a better way to see the person on the other side. And that was a huge learning for me when I first got to the US. Try to build those relationships instead of just jumping right into the thick of it with people in the best way possible. Now, once I figured it out, and work in progress, no one is perfect, so I'm still perfecting that skill, but once I figured it out, I also figured that there's a lot of benefit in being able to cut to the chase in a good way. I'm a great fan of the book and the method Radical Candor. I think that confronting people directly through deep care for them, for the business, and for the product yields great results. And I've seen that happen so many times. So that is definitely something that I would suggest people to give a try. Don't beat around the bush, go and confront the topic directly. But building on what we just talked about, you can do that if you don't have a relationship with other person. And you cannot do that if you don't have alignment. Once you have that, then you can move faster, right? You can really get to the topics and debate. That the best way to get to the crux of things, to get to the kernel of truth that we care about is through debate and brainstorming. And those things have to happen in a safe and trusted environment that is built again on relationships and alignment.
SPEAKER_00Yeah, totally makes sense. I'm going to stay on the subject of alignment because I think that's such an interesting one. You know, you've talked about data, which I think is so critical. Like, let's all look at the same information because smart people will come to a shared decision. We also talked about building, taking the time to build the relationship. But in that journey of getting alignment, I want to hear like maybe some things that surprised you or went wrong that you had to course correct along the way. You've got the data, you've got the relationship, but maybe that's not the entire picture. So can you tell us a little bit when I let's keep exploring alignment? Because I think it's such an important sort of concept to think through.
SPEAKER_02I'll add one more dimension here, and let's just call that experimentation. And I think that it is extremely important, especially in the days of AI and SASPOCalypse or whatever we want to call that, because the evil side of data-driven decisions and alignment is getting to a point where you experience analysis paralysis, right? And I've seen that happen as well, right? So you need to build your self-confidence and you just go and get more data and more data, and it's never enough. You're never gonna have enough data. Or, for example, someone comes to you with an idea or a concept and you just send them to get more data, right? We're a data-driven organization. Go get me this data, go get me that data. And it's a never-ending story because you never have enough data. So at some point, you need to make a decision. And I'm a great fan of a couple of things I've seen happening in Amazon and other companies. Here, the concept of a one-way door, two-way door, right? Things that you can revert back, things that you cannot go back from. And I'm going to attach that to experimentation and the entire MVP process because the best way to drive alignment when you don't have all the data, because you'll never have all the data, is to put something in front of customer and see what happens, right? So you create an MVP, and I dislike the term because it's been bastardized in the industry, and now it's an excuse to put half-baked code that doesn't work out there. So that's not the goal. Let's redefine MVP as the least amount of effort that is solving a problem in a meaningful way. How do you know it's meaningful? You get money for it, right? So that allows you to really put something in front of customers and then they're gonna tell you, right? Because we're in this echo chamber of alignment and data and startup thinking or corporate thinking, and we make up all those personas and all of that. But guess what? I haven't found a single product that stays the same after it was introduced to the market. So let's just make this process as fast as possible. And then when I go back to my peers and I say, Well, I talked to 50 customers, that's what they told me. What's better to drive alignment than that?
SPEAKER_01Love that. Everything about just trial, we call it trial ballooning. We love the idea of trial ballooning, both with other leaders. Here's the trial balloon of our decision. What are you thinking given the data we do have? And then same with customers. So love this. Now in the AI era, this whole thing is exacerbated, right? We moved even that much quicker. And so, even the upfront, you have even less data now, right? Because you're moving so dang fast. And really the learning only can happen once the model is live and you're learning real time, right? Talk to us a little bit about how your sort of leadership and the way that you help your product teams get to shared vision has changed now that we are in this era of AI and accelerated work.
SPEAKER_02Yeah, yeah. The era of AI, I think, really started just a few weeks ago. So obviously it was coming, but I was staying very close to the technology through Teams and through my own experimentation. And there was a promise, but it wasn't really fulfilled. And I'm specifically thinking right now about what's called vibe coding or deeply AI assisted coding, not just autocomplete my function, but create something almost from scratch. There was a promise, but it didn't really happen until after December, I think, of last year. And I recently wrote about that. I built an entire application, multi-user client server application with a proper architecture, using standard frameworks and programming languages that I know nothing about. And I did it in two weeks. My mind was blown. So, Heather, I'm gonna challenge you. It's not that now we need to work on less data, it's more that we don't need as much data, right? Because why do we need data at all? Because the cost of trialing something, right, has to be proportional to the benefit, to the return on that investment, right? We want to go to the moon. This is something we need to plan years and maybe decades for because of the massive costs. But if it's similar cost to catching a bus, let's just go to the moon and see how it goes. So obviously, very extreme example here. But if I translate that into development, if before trialing something would mean I need to get a team of engineers, a squad, some engineers, PM, product owner, designer, and I need them allocated for a quarter or a few sprints at the least. Now, maybe I don't need that. Maybe the PM can use their friendly AI assistant to design and vibecode something not connected to a production environment. Please, God, let's not do that with something that a PM just vibe coded in an afternoon, but something that would really illustrate what we're trying to do here and show that to a customer, right? So instead of having that PM go and collect data for a couple of months before we do that, let's just do it. And within a week, we're gonna see, okay, are we progressing in the right direction? And if so, let's do a little bit more. And by the time that we have a full squad spending a couple of sprints on that, we have pretty good conviction that we're on the right path.
SPEAKER_00You know, it's interesting because there's a lot of that that will agree with you. So this, let's see where this goes. But what we found is that discipline and this really is something that over the last six months, we went down the path of helping people with prompts and with gently trying to get them to use AI effectively. And I think now we've gone full in to it's really easy. You can do this, like you say, fully blown applications in two weeks. So that understanding, and it's not a matter of collecting a ton of data, but a matter of truly understanding your customer problem, we believe is more vital than ever because it is so easy for you to trial. We want you to have conviction that you're pursuing a direction that makes sense and not just something that came up to you in the shower, right? Oh, I can do this, here we go. It might be loosely based on sort of data and some intuition, but I think characterizing and pulling that together, we are telling everyone that this is more important than ever. So I'm interested in your take on that, like your response.
SPEAKER_02Let me give you a concrete example. It's pretty fresh, so I'm not going to mention details of the product and company and all of that, but I was involved in this company, and there's a very nice product portfolio and addressing Fortune 100 customers, meeting with a team that we work with on the other side. Our product team came back saying, Hey, we need a new product. They have a problem that requires a new product. We can solve that. We have what we need to solve that. It's just a matter of putting a team on that for a couple of quarters. That's pretty expensive, especially for the size of the company that I was involved with. So we took a bit to hey, let's collect data. Let's see what other customers may use that, or what is the size of the problem in the industry? What's the upside? And then let's go talk to other stakeholders from other companies. And then let's go and talk to executives. So that's where I got involved. So I got to talk to executives on the other side to make sure that this is really a problem. This is really something that they would appreciate a solution for. So that was before 2026, right? So it wasn't as imbued with the iTool and all of that, like we can do today. And we went through the playbook exactly what you would expect. And we ended up building that product super fast. Two quarters from an idea that was validated to something that's actually working super fast. Put that in front of customers, didn't work. All of a sudden, you put that in front of customers, and they see that their customers don't respond to that the way that they thought. Or that, yeah, it solves that problem. But now that we're really thinking about that and not just brainstorming with you, there are other priorities that we should address first, right? Or now that I'm talking to my CFO, we can actually live with that problem for a bit. And great work. Let's touch on that in a couple of years again. So that's an inherent deficiency of the more traditional process. Those things would happen. And I was trying to describe the process because there are never guarantees, but we were really trying to touch on all the points and build confidence over time and not invest everything on the front end and all of that. But we ended up getting to a place where, hey, you know what? If we could have built a quick proof of concept in a couple of weeks and get all that feedback then before spending a couple of person years on this project, it would have been so much better.
SPEAKER_01What you're saying though is I think we're in alignment, right? Which is how do we learn up front before we invest too much? And how do we make the right decisions before we invest too much, right? It may just be slightly different in how we do it, but I think we're saying the same thing, which is understanding the customer problem, whether that be through quick experiments or something else, or doing interviews or the like, and minimally investing as long as you possibly can to get as much learning as you possibly can before you do the full investment, right?
SPEAKER_02Totally. 100%. I think that again, the ultimate experiment is. Who's gonna pay you money? I don't care as much about money, it's just a proxy to value, right? If it's really valuable, people would put money on the table. So who's gonna pay you money for this product? And traditionally in our industry, it takes a lot of effort to put forth a product that you can collect money on. So we ended up creating proxies. We're gonna look at data of existing users and try to infer if they're gonna have a problem for which the new product is going to be a justified solution that they would pay for. It's not really the experiment we want to make, right? But it's an easy one that we can make. Or we're gonna meet a couple of customers of a dinner and we're gonna ask them what they think, which is great, but it's not really the endline experience that we want to have. People actually putting out the checkbook and writing a check, right? What's changed is the cost of getting to that experiment, right? To that final golden experiment. So on that, I think we're a lamp. We're just in a different place right now that allows us to do that. And in a slide pivot, that's also what's everyone talking about when it comes to SaaSpocalypse, right? For the last decade, there's been a moat that companies were able to create by building very complex and automated workflows, a lot of technology around that in SaaS. And their justification for existence was hey, we're doing all of that. Sure, there isn't something that's clearly patentable there, but I want to see you try build that, right? It's gonna take you a lot of time and a lot of money and a lot of people. And while you try to do that, I'm going to continue to run forward. So now that is not necessarily correct anymore because complexity just means I run some more cloud cod agents or whatever cursor, and they're gonna work on some more tokens, so it's gonna take me a couple of days more. That differentiation and that cost of experimentation is what's changing these days.
SPEAKER_00So interesting. What an interesting conversation this has been. I want to bring it back to the beginning now to ask you given what you said about being a rebel right in the start, put the lens of what's happening now and what advice would you give to someone who wants to be a rebel today?
SPEAKER_02Yeah, that's a great segue. I'm gonna try and tie that to my evolving thinking on the implications of AI today. And I'm dividing that very roughly to three levels. You can think about a business stack. So at the top, you have your funders and your investors. And if previously you have a very sophisticated and elaborated SaaS operation, we just talked about that, that would have been great. No longer the case, right? So those people are now focusing on how can I either leverage the developments I'm seeing with AI to address niche markets that I couldn't address before because they weren't big enough. But now the amount of effort that I need is not as significant. So that's a change, right? You want to be a rebel as a product-minded founder, you think about that. You no longer need a lot of people and a lot of time and a lot of money to sort a company. If you think about it, that's another fundamental principle that has changed. You have to go after the huge use cases. Why? Because you need to raise a lot of money. Why? Because you need a lot of people to solve a meaningful problem. Now, roll all of that back. You don't need a lot of people, so you don't need to raise a lot of money, so you can solve a smaller problem and actually make a bigger impact on the industry. So I'm definitely seeing a lot of that happening. The second thing is that I'm seeing a lot of drive coming from the PE, the private equity side of the landscape. There are huge opportunities in taking mature companies and rebuilding them with AI in the center, right? Because if I'm a PE and I'm acquiring a company, I'm acquiring their business book, and I will connect that directly to relationships with customers, right? And I am buying their brand name and I'm buying their data. In many cases, they hold data that creates some sort of a network effect that they can use. Everything inside, I can now transform with AI and optimize the costs. So increase my margins by driving radical optimization on how the organization is working. So at the top level, if I want to rebel as an investor or a founder, those are the things I would think about. Then stepping down one level, now we're at the product or the service or the thing that creates value within the company, right? So how do we rebel there with AI? There are the traditional ways of doing things, no need to elaborate on that. But then I would split it into front-end and back end. On the front end, what I would usually do is put a lot of dashboards and showcase all the data that I have, right? And some menus and workflows that are programmable and stuff like that. So that's a proxy. Now we have AI that can be used as an interface. So if I'm really thinking about how I want to create product disruption and rebel versus everything else that I'm seeing out there in the industry today, use AI as an interface. Cut to the chase with your customers and help them conversationally get to what they need and solve that problem for them as the user interface. And then there's the back end. On the back end, you have an opportunity to use AI as well, only this time you're using that in order to create agents and in order to have a more faster, better access to data. We're transforming from a model where people are part of the loop, right? Think about I don't know, a person that is using a service that augments what the person is able to do. I don't need to go through all the data, right? Because the system is telling me what's important. I don't need to do all those things manually because I can work with the system and automate and batch do a bunch of things. That's the old age. What I'm seeing emerging today is a man creating the loop, right? So on the back end, think about your customers not as an element that everything needs to go through, but as an element that is setting an agent or setting a system that would run autonomously. So that's how I think about the rebellions there. And I would say finally, and that's probably the most challenging thing on an operational level: go-to-market, engineering, design, support, customer success, finance, legal, every part of the organization needs to be rethought and redesigned using AI. And the faster we do that, the better, right? So that's an opportunity for every individual in the company to become a rebel, not just product people.
SPEAKER_01I love it. It's hard, right? I feel like that's one of the common themes that we hear is my gosh, like how do we stay up to date or how do we even catch up? Super, super helpful and thoughtful in terms of the big strategic implications and how to be a rebel in the age of AI. I want to ask one more question and then we'll start coming to a conclusion. What would you say if you're sort of a classically trained PM or a classically trained senior product leader who is being asked to take on AI in every way, shape, or form and is feeling a little overwhelmed? What's the one piece of advice you'd give them to help them feel confident? They're probably already over the first step, but taking that next step into developing confidence and becoming the rebel?
SPEAKER_02Yeah. First, I would tell them you're doing a great job. Just for the mindset of understanding that something big is happening here, you're doing a great job, and you're ahead of 99.5% of the population on this planet. So acknowledge that we're going through a huge disruption and things are changing by the week.
SPEAKER_01Yeah.
SPEAKER_02Take small steps, small and meaningful steps, but make them directional, right? It has to be directional because otherwise you can get lost, right? Well, one day you work on a chatbot, the other day you create a well, whatever it's called now, CloudBot agent. The next day you create a query system for machine learning. It's impossible to stay on top of everything. So go back to first principles, work backwards. What is one meaningful thing that is going to be amazing if we can gain some progress on? Is it optimization of certain element of our operations? Is it the value that is being created that can be created in a better way? Is it new value that is being created, a new problem that we want to solve? Just pick one, right? Align on that. Don't just pick one and keep it to yourself. Pick one and communicate that to your leaders and to your peers from other functions. Hey, this is something that I'm going to solve. What do you think? Isn't that exciting? And then work backwards from that. What are the small steps that we can take to solve that? And who can be your trusted partner in that? Your AI. And again, take it with a grain of salt, right? I'm very forward-leaning. But one thing that I've learned is that one at least of the best ways to learn about AI and AI applications is through AI, which is not intuitive, but I share that I've been working on my own project just to get my hands dirty and see how things work. I have my team of agents, and that's the development team, and they're building code. And I have a separate AI who's my chief architect, who's guiding me through the trade-offs of the architecture that I'm building. And I have another one, a different AI, that is my product management team. And I'm talking to them. And if you look at my first conversations with them, it's about hey, this is my idea. What should we do? How do we even get started? What are the right frameworks to think about things? Okay, let's build a plan for the next week. What should we do? Step by step. Very easy. We're at the beginning of this huge journey. Celebrate the small steps. Just make sure that you build alignment, people understand what you're trying to do, and that you're incrementally in a way that's compounding progressing towards your goal.
SPEAKER_00No, I love this advice. And all I would add to that with our framework is as you think about all the different things that you can do, and we're talking about focus, is how you think about picking the right one. That is our core job. That is going back to first principles and fundamentals. That is what we do, and that really is the most important thing. Gosh, what an interesting conversation. Just so deep. We so appreciate spending the time with you. Thank you so much, Raveev. We really appreciate your time.
SPEAKER_02Likewise, thank you so much for the opportunity.
SPEAKER_00Thanks for listening to this episode of the Product Rebels Podcast.
SPEAKER_01If you enjoyed this conversation and want to learn more from Product Rebels from companies like Netflix, Amplitude, and beyond, please follow us wherever you listen to podcasts and join us for another impactful interview in about two weeks.