June 15, 2022

Building an AI-Driven Product With Gagan Gulati of Behavox

On this episode of Product Chats we dive into building AI-driven products with Gagan Gulati, Chief Product Officer at Behavox. He’s built a wealth of product knowledge during his decades-long career that features product leadership roles at Microsoft, BEA Systems, and Oracle. Gagan shares his deep knowledge of AI-driven and AI-infused products in this episode, so be sure to check it out.


Time Stamped Show Notes

The importance of storytelling in your career [02:33]

Advice for transitioning from engineering to product management [04:15]

Ensuring you understand users’ challenges [05:49]

Avoiding “what” and focusing on “why [06:41]

Building AI-driven products [09:00]

AI-Driven products vs AI-infused products [10:54]

Researching to build an AI-driven product [13:55]

Difference being a PM in a large vs small company [15:15]

How wide a span PMs at startups have [19:36]

Importance of aligning with company goals [22:21]



Product Chats is brought to you by Canny. Over 1,000 teams trust Canny to help them build better products. Capture, organize, and analyze product feedback in one place to inform your product decisions.

Get your free Canny account today.


Stay Connected!





Kayla: Thanks for tuning into Product Chats. On today's episode, I talk with Gagan Gulati who is the Chief Product Officer at Behavox and we talk about building a AI driven product, also the transition from engineering to product. So enjoy the show and don't forget to leave us a review.

Kayla: Hey Gagan, thanks so much for coming today.

Gagan: Thank you so much for having me.

Kayla: Awesome. In a minute or less, can you tell us about yourself?

Gagan: Yes. My name is Gagan Gulati. I'm the Chief Product Officer at Behavox. Behavox is an AI driven company that helps organizations root out the bad actors. I've been in this company for 18 months doing amazing work. Before this, I worked at Microsoft and I spent almost 16 years at the company. Grew up along the company, worked in Office 365, worked in Azure Windows. So a lot of experience there, that's pretty much the professional side of it. I spent more time in a couple of other companies as well.

Gagan: On the personal side, I live in Seattle, very happily with my wife and two kids. And it's just been a great journey. So thank you for asking the question Kayla.

Kayla: Yeah. So you kind of talked a little bit about your more recent years, but I want to learn a little bit more about how you actually got into product.

Gagan: I was an engineer for the majority of my life. The first 10 years of working, I started working on, I was actually a scientist for the first year. Became an engineer, after that worked at Article, worked at BEA Systems. We should build a Java, the web server. Oh, sorry. Java application server that pretty much the whole word used.

Gagan: Then I decided that's enough of Java. I went to Windows. Six years, building Windows Vista, Windows 7, some of Windows 8. I think at that time the realization came that I'm looking for that. Yes, I'm pretty decent there, but at the same time, I'm more interested in understanding core customer problems.

Gagan: What they're trying to solve. And I gravitated towards that very, very quickly during my days in Windows 7 and Windows 8 and the more I gravitated, I realized that's what I most passionate about. Defining why, what and yes, how, but a little less of how you know, how you want to solve the problems.

Gagan: So that was one thing. My friends tell me that I moved out of engineering, being an engineer to be a PM because I was never a good engineer. I take that, but it's been, what's funny is that the last 10 years have been fantastic, both from a learning and group perspective. So I'm very happy that I changed my profession halfway from being an engineer to a product manager.

Kayla: And so to tap back into that, around you started as an engineer, right? What are like, let's say, there's someone listening to this that is an engineer and wants to get into product. What are the skill sets that you've really leaned in on that kind of are strong as an engineer and also strong as a product leader?

Gagan: I think at some point in your career all of us realize that the most important thing is being able to tell a story. And I think that's the most important, whether you're telling the story to your fellow engineers on how you build something right, or how they should go build something right. Or you're telling the story to a customer or to other people who are not in your department.

Gagan: So I think if you are, if you believe that you're a good storyteller, you can weave a story together. You can explain the problem well. You can work the words trying to answer the core problems, simplify them for your audience. These are the core capabilities that every leader should have. Whether you are an engineer, whether you are a product manager, it could be in any department for that matter.

Gagan: For the shift specifically from an engineer to a PM, you need to be passionate about the spending time and discovering the problem that the customers are facing, not what they're saying, but what's the intent behind what they're saying and I think that was the key thing for me. And I think that's one of the key things that I believe, you know, if you are trying to move away from being just an engineer to become a more technical product manager, I think these are some of the core things that you want to look for, more storytelling, understanding you know, taking the time to understand the real problem that the customer is facing.

Gagan: Because a lot of times somebody will come and say, oh, you know my fever is high, high fever. That's just a symptom. There's something else that's going behind that's causing the body to react that way. And I think that's the, that's the capability that you're looking for as a product manager, because you literally are defining why you're going to do something and what are you going to do to resolve the problem?

Gagan: So if you are an engineer and you have, if you believe that that's something that you would want to do, that this is a great discipline to be in. There's another thing that you need to understand if you're an engineer trying to be a product manager and that is that you will soon become jack-of-all-trades because when you're an engineer, you basically focus on your code.

Gagan: You write it, you test it, you ship it and you feel pretty good about it. But when, once you become a product manager, you have to go through the entire life cycle of building product, like from conceptualizing it, right. Our feature back from conceptualizing it to figure out whether it's a good product market fit that you want to build it.

Gagan: Your customers are happy about it, working with engineers, telling them what to do. So that they can actually build it for you. And then when it's done or it's being built, you got to work with everybody else, whether it's marketing, whether it's sales, if something sellable or marketable, whether working with support, working with account managers, like there is an entire, so you basically have to be that tie in.

Gagan: So it's a very heavy Bread first job, and if you like Bread then this is the perfect job for you in a sense of, you can put the whole thing together. So there are there are elements, I think to the question, just to answer your question for the last one time, I've answered twice is instead of asking, what would it take for you to become an engineer from an engineer to be a PM?

Gagan: And the question, what should be asking is would I be a good product manager? Like, are these the qualities that I possess or I want to posses and if you are the person that there is nothing stopping you, because I have gone through this one from being an engineer to a PM and actually going from an engineer to a PM. You can be very successful in a technical organization, a software company.

Kayla: So I think on that, right, something you mentioned was understanding the customer, right? And that's a core piece of that. So talk a little bit more about how you do that and how you understand your customers. Like what frameworks are you using? What are you making sure that you're doing so that you understand the kind of problem versus the bandaid and just solutioning.

Gagan: Yeah, good point. And this is where I think there's a lot of I mean, there is an opinion and everybody will have a different opinion and I have an opinion as well. The key thing as a product manager, I think the first key thing as a product manager is to not have bias.

Gagan: When you go in and ask questions. Number two is have data from a variety of customers, not just one. Number three is try to ask questions or try to interview or ask questions, where you try to get to the depth of the problem rather than saying, oh, you're looking for two reports, right? Sure, I build them for you.

Gagan: These are the reports that you want. Are these the reports you're looking for, rather than asking a question like this, the focus should be on, so why do you want me to build these reports? Or why are you asking that I should be working on something like this? A lot of times, product managers focus on the, what, what do you want me to do?

Gagan: I'll give you a good example is why I'm building these reports? Oh, because I want to build a dashboard. Why are you building the dashboard? Oh, because my, I need to present this dashboard once a week to a manager or to my management. Why are they looking for the dashboard? Oh because this is how we measure progress.

Gagan: Progress for what? So once you start on the why journey you get to realize that they're trying to solve a very particular problem, which is that their managers or their management is trying to make progress on a certain point. And this particular piece that you may be working on adds to the big picture and also opens up opportunities.

Gagan: Right? And I think these are some of the techniques that some of us have used over a period of time to really go to the depth of the problem that the you're trying to solve or the customer's trying to solve, and then how you can go help drive that. So, I mean, there are a lot of tools that can be put in, like the simplest tool is a spreadsheet that you can basically start saying oh, I met with this customer and here's a tab for that and here's a tab for that with asking specific questions. All of that is for med in terms of how you want to present the data that you've collected, but the key here is the content that you're really, really going after and that I think is the crux of I think every product manager should go after.

Kayla: Awesome. And I think on that, this kind of gets us into our next subject of like building products. Right. And so I kind of want to dive in a little bit about focusing in on building a machine learning based product. Can you talk a little bit about that?

Gagan: Yeah. So as I was saying earlier, I worked at Microsoft for very long, and then I came to Behavox and one of the reasons why start came to Behavox was when I was working on different projects at Microsoft, we were trying to infuse AI or machine learning into the existing products. So I have a workflow or I have a problem, I have a solution that I'm trying to get our customers to adopt. And then somewhere in the solution, you basically are saying, oh, I need to do some automated detection or, you know, wouldn't it be great if we automated this and the way to automate, automate this would be to run a machine learning model, which could basically tell me the probability.

Gagan: It could be something else, and I can then infuse that to make the life of my customer easier. My user easier. Right? This is a very different model of building products because there data science is an afterthought, machine learning is an afterthought. Afterthought, in a sense of, it plays a small role in the big picture.

Gagan: The world that I've lived in the last 18 months at Behavox is a completely different world. So instead of going the AI infused or machine learning infused world. You're talking about an AI driven machine learning driven world. Right. Which is the whole concept is around the problem you're trying to solve can only be solved if machine learning was sitting at the core and you're trying to figure out and you're trying to build from there.

Gagan: So just a one line spiel on Behavox, I think it will be easier to explain the rest of the conversation is so at Behavox we have multiple products that try to where we try to help our customers protect their organization by rooting out the bad actors, bad actors equals users who, or employees, or employees with immoral, illegal, or any inappropriate behavior.

Gagan: And that can do itself into multiple different products. We have a compliance product that primarily is for financial organizations. We have inside a crime product, which is all predictive and catching the bad actors before bad things happen and then conduct, which is also very similar. So when you have products that are focused on predicting, you know, who, for example is the highest risk individual in your organization.

Gagan: So that's our insider crime product whose job is to use all the data that's available for that person or that individual. And come back and say, is this person a high risk individual for the organization? And there could be multiple different indicators. For example, does this person, is the person going through financial distress?

Gagan: Is this person going through a depression? Is this person being bullied at work? Is this person being, there could be many, many such factors and there's a set of, we call them indicators that put together. The outcome is literally, is this person high risk individual or not, that you should be watching for literally.

Gagan: Now this outcome can be, sitting on a dashboard as a yes or no. We could generate a watch list and send you an email, a list and send you an email. We could send you a report. We could have sent you. We could send a pigeon with this answer, right.

Kayla: Highly unlikely but yes.

Gagan: This kind of a product, if you're building this kind of a product, it's not AI infused. It's AI driven. So therefore, as a product manager who works closely with the data science team, and the rest of the engineering team to build a product like that, the focus is a lot less on the user experience in the beginning. You know, of course those operations will come and will be focused a lot more on, are we building, do we have the right indicators? Is our machine learning working right? Do we have the right data sets that can help us detect these indicators through all the communication data out to all kinds of other data that you may do? So when you are building a product like this, you have to take the conventional product building hat out of the picture.

Gagan: Let me go to 10 customers and ask them where their problems are, as an example. Let me understand how they would run their workflow. Would they like a dashboard? I can tell you that if I could come to your organization and with certainty, tell you, here are the top five individuals that you should be watching out for because of all the work they've done, you would say, write me, write it for me on a piece of paper and that's good enough. I don't need to go to a dashboard. I don't need to. So like, that's the mindset change that I've tried to infuse and actually infuse into myself and then the rest of the team that yes, we will work on the operating model.

Gagan: Yes. We will work on a beautiful user experience so they can investigate really, truly. Is this the case and what have they done? But there is a lot more focus on getting machine learning, right? A lot more of the, as you say, data driven AI. Teaching the AI on how to detect some of these indicators is the key.

Gagan: And that I think is the biggest learning. If you are working in a AI driven organization of products versus AI infused products.

Kayla: And I think with that, right at the end of the day, it gets down to what problem are you solving for your customers? And maybe they don't care as much right now, right about, Hey, I need a shiny dashboard. What they really care is are you solving my problem? And you're are you solving it in an easy way that I can digest? And so with you, right, it's not a shiny dashboard. Maybe it, maybe you guys have those at the end of the day with any product, right, it's all about solving the pain of your customer and understanding them. And yes, a pretty UI does well, but also if it's a pretty UI and it's not solving their pain, it doesn't bring any value and you're not going to grow as a business.

Gagan: Exactly. That's exactly like what you said is 100% right and that is the key, but that understanding comes doesn't come very easy. That understanding doesn't come because 95% of the world still runs on an AI infused model, system and it's not wrong.

Gagan: It's just, that is also very well required getting the workflows right so you can solve a problem for a sales guy or a marketing person or, you know, depending on what your audience is. This is very important and very complex. It's just that the world that I'm living in, the question you asked around, you know, what does it mean to build a AI driven product is what's the core value that you're trying to build that can solve the customer's problem.

Gagan: And the core value in this case happens to be making sure that you're finding the bad actors. And then there is a presentation layer, and then there is all the other things that you can do on top. So I think that's one of the biggest learnings that I had as I transitioned from building AI infused product to AI driven product.

Gagan: That's one of the big reasons why at Behavox we have some of the best scientists on the planet who, data scientists on the planet who are helping us do this. We still have to go do a lot of work as a product manager because you need to really go dig deep, understand what kind of model do we want to get built?

Gagan: You have to go research on a very different level. Your research is driven by like for our insider crime product, FBI, like all the research that FBI has done on how they go after criminology. There are degrees in criminology, so literally some of our product managers are now experts on criminology.

Gagan: You can literally go ask them. And that is not something they signed up for, I guess, as they came in to the company but it's been an amazing amount of learning in terms of the amount of research that you have to go do. Talk to the behavioral scientists, talk to the professors at Harvard, talk to the FBI agents, talk to the criminal. So that's the difference.

Kayla: I think it's really interesting because it's just about a mental shift, right. About how you're building out a product and AI ML and data science are really, really hot words. And so there's just like kind of breadth of how people use it and it could be an AI infused product, or it could be an AI driven product.

Kayla: So I think as people are transitioning into different industries, figuring out is the product AI infused or is it an AI driven product? And do I want to work for a company that is AI driven or do I actually want to work for a company that's AI infused, right?

Gagan: Yes. And that's the, to the point of hot topic I think this is where the buck's going. Although the buck has already gone there and following the buck.

Kayla: So something I kind of want to transition to, so something you mentioned as you started at Microsoft, now are the chief product officer. So I want to kind of get your insight on like what things are different at a bigger and at a smaller company and how do people set themselves up for success?

Gagan: Yeah. I love Microsoft and I guess I would equally love other companies as well, like a Google or a Facebook. They're all big now. Right, Amazon. What they give you is an amazing sense of security and you're still working really hard. Right. And you're introducing an impact that's at a much, much broader scale than a small startup would do. However, the big difference that you see is, you know, how would you measure your impact? And that was a huge learning as I moved from Microsoft into Behavox.

Gagan: The number two is how fast-paced a startup can be when it comes to decision making. I'm not saying that Microsoft was not fast-paced. We've all seen where the company has gone. And some of the other companies that make decisions fast but not as fast as a small company would make. And I think that was another big thing, which is the decision making is much faster.

Gagan: You have to make decisions. You have to commit to those decisions and move fast. You cannot have a three month review period. You cannot have a, okay. I'll sit here. I'll find something that I'll work with my cohort. And then I'd go to my manager and my manager will agree with me. And then I'd go to the next person and the go to the next person.

Gagan: You cannot have seven layers of decision-making that will typically go through in a big company to start a new feature or start a new product. Here the decision making is a very, very different volume. I think that's a big thing. So for PM's who are professionals who want to move from big company to small company, that'll be a shock on how fast things move, how fast decision gets made. That's number two.

Gagan: I think the third thing also is depending on where, which company you're moving from, like a big company, a small company, at least from my experience was a lot of times when you are in a big company, you have big responsibilities. So as a, as a leader, you don't have that much time to go and work on every decision point.

Gagan: You trust your team to make the right decision. They will deliver it a lot more and they will go for further because the span of control is usually pretty big. When you come to a small company, which as in the case of Behavox 200 something people and we grew massively over the last year, but we still have 200 something people. There are a lot of decisions that we have to take, including which needs to include the CEO of the company. Right? So that's the other big thing, which is because the impact is company-wide, the impact is on the products. There is an impact. For example, if I decide tomorrow using the same report example, which is very tiny, but I'll use it anyway probably the wrong example, but we decided to build a couple of reports. New reports or new report infrastructure that you put in place. One, if you did it at Microsoft, you have to go through seven levels or five levels or three levels of agreement to say, Hey, this is the reporting that we want to put in place. Okay. And yes, you will get through and you will say, yes, that's one.

Gagan: And then you'll take a longer time to build that because that's how the world works. There are milestones within a milestone, then you kind of try that and once you put it into, once you built it, you put it in production, then you will get towards your evangelizing that customer deployment, there may be an upgrade from a licensing key point of view and all those things take time. And you can basically say, you know once my product is in beta or some other state I will start the conversation.

Gagan: But when you are a startup, all of those things have to be discussed right at the beginning. First of all, you want to make sure that revenue team is fully aware that this is coming on day one so they can shift their motions if required. Number two is you want to make sure that your entire, you know, or at least the people who are the most impacted like CEO that he's well aware of what's going on because if you're a founder led company, you know, like I am, the founder is very heavily interested in the direction of the company.

Gagan: Ideas are very welcome. You should absolutely go make, do with them and work through them. But at the same time you want to make sure that they're well aligned. Very, very important. And then therefore the key point here is that I'm trying to make is while the decisions are made much faster, why we need to align and yes, you need to align at the top level on decisions much faster and the execution. You've got to over-communicate as well in a much better, much faster way than you would ever do in a big company. For example, in a big company, the report conversation. Sure, when in six months when the beta happens, yes, you can start conversing with the revenue team.

Gagan: You can make the marketing team aware. Here the communication starts on day one and you've got to keep over communicating on the changes you're bringing in so that they get deep. They get entrenched into the entire marketing revenue organization, very, very early that this is what's coming. So over communication is the key. And I think some of those, so there is faster and better. You've got to do both, and that is a lot of talk but these are real life instances of things that I have seen that the things that I would say this can wait. Easily, we don't worry about talking to the go-to-market team yet. Don't worry about talking to the sales team yet.

Gagan: There will be a time. Here, there is no time because they're always pushing, you know, driving forward in a much, much faster pace. So, I mean, I just kind of spoke through a few of them.

Gagan: And the last one I would definitely add is if you thought that you were in a big company and a product manager and you thought that you have a span of, you know, you have a big span. I'm working with this and I'm doing that.

Gagan: You don't even know what you're getting into in a startup, because you will have to have the end to end ownership of your product. Way more than you would do in a big company. In a big company, there's a dedicated department for marketing. There's dedicated department for doing something else. There are frameworks in place built over the last 10, 15 years. When you come to a small company there is nothing. You have to start from ground zero. Every time. There is nobody who's going to help because by default, you are cash crunched. You are resource crunched. So while you may be building something, a new feature trying to get to your marketing team, they may be saying, no, we are just two people or five people.

Gagan: And we are working on something that is super priority. So why don't you do it yourself? We can help you. So there are way more, many hats that you will be put in now. You will be a product manager. You will be a product owner. You will be scrum master. You could be half marketer, like getting case studies ready to give to marketing team.

Gagan: So they all have to spruce up. You may be writing playbooks for your product. You may be working with the support team yourself. You may be like doing way more. You don't even realize how thin you could be spread going in, but that's the joy of it because you are the actual owner of your product. You are the stakeholder, your the primary stakeholder.

Gagan: You joke about it that I am the CEO of my little thing, like 40 individual product managers, and the CEO has no responsibility actually speaking. He's not supposed to do anything, but he's the busiest or she is the busiest. Right because they have to make sure that the entire ship keeps floating so these are two or three learnings that I feel foster better and full ownership, end to end ownership of your product that you just somehow in big companies get used to.

Gagan: You say I'll just do it and I'll give it to somebody else. There is no somebody else. There's nobody to catch it on the other side. It's you. So that I think I just stop now, but these are two or three big learnings that I had to take in in the first six months. And I can keep going.

Kayla: So on that note of big learnings, right? If there's an aspiring product leader and they could take away one thing from you and one piece of advice you would give, what would that advice be?

Gagan: A thing that we didn't talk about speaking was hiring and how should we go hire? I think the two things that I think is very important. One as a product leader, you need to be aligned with the mission of the company.

Gagan: Because if you are not aligned with the mission of the company your, team's not going to get aligned with you and your team could be people directly reporting to you, people you influence. So if you ever read the book Extreme Ownership kind of walks through that little philosophy.

Gagan: So that's number one. I think the second thing is, especially if you're in a smaller company and the same thing's true for big company as well. Culture as well. Is that people you can trust. And are you working with the people you can trust because if you're aligned and your teams are aligned, you're over-communicating all the time to make sure that they are always aligned with you.

Gagan: You have to make sure that you're aligning the people that you can trust and they will deliver. That even at some point, let go and say, okay, you know what? You will take care of it. I know really well, so that you can focus on the next big thing or the biggest, you know, next thing that the company needs, a team needs.

Gagan: So as a product leader, especially as a product leader, I think it's super important to be able to align with the goals of the organization very well. Commit to them, influence, have the right influence that you can bring in with the other leaders, because you are the product managers, you are the end to end owner. With the people who report to you so they can align with you and have the level of trust with everybody around you, especially the people who work for you so that you can continue to lead on next, you know, on the big things and drive your team forward. And you are rest assured that your teams are performing at the level you want them to perform and you have the trust in them. So, and that's one of the things that I've tried to learn from the past at Microsoft, at Oracle and other companies, and I try to bring in even at Behavox not just to myself, but to walk through the bottom.

Kayla: So on that note, you mentioned hiring what roles are you hiring for right now?

Gagan: Haha yeah so in the in the company that I'm in right now, we are hiring for product managers.

Gagan: We are hiring for the documentation team reports to me. The UI design team, the design team report to me. So we are hiring across the board. Pretty much. We have built a pretty decent team over the last year and a half so now we're focused on bringing in more fresh, more fresher talent into the organization because we have a pretty decent leadership bench with us right now, but the company is growing and it's growing fast. So, you know, there are always opportunities. If the strong product manager or your design person or others who are looking to work in a company that's AI driven, cutting edge all the time. And working on products that I believe are going to be game changing in the coming future then I'd love to have a chat.

Kayla: Awesome. And then on a note of chatting, where can people reach out to you?

Gagan: The best thing to do is LinkedIn. Just search for me on LinkedIn, you know, the easiest and just connect with me and then we'll go from there.

Kayla: Thanks again to Gagan for joining us on today's episode of Product Chats. If you want more product management resources, head over to Canny.io/blog