The Only Way CEOs/COOs Get To Survive The AI Wave | Stan Hansen, Egnyte

Shamil Malachiyev: Good morning, everyone, and welcome to this week's episode of The AI Floor. My guest today is the Chief Operating Officer at Egnyte, where he oversees revenue, marketing, customer success, operations, and corporate development. He spent over two decades leading companies like Adobe, Domo, Pluralsight, where he helped take the company public, and Wells Fargo, where he ran organizations of over 1,000 people. Under his leadership, a third of Egnyte's revenue now comes from AI, serving 23,000 customers globally. Please welcome to the show, Stan Hansen.

Stan Hansen: Hey, thanks so much. Really appreciate it, Shamil. It's great to be here, and I appreciate the opportunity to be on the pod. Excited.

Shamil Malachiyev: Likewise, super excited. Can we start by sharing with our listeners who are new to Egnyte, who are new to Stan, some of your path and the work that you do?

Stan Hansen: Yeah, look, I guess I've probably taken kind of a creative path to get to this role. I started with a degree in finance and ended up taking a job with Sprint. In fact, it was called Sprint PCS. This was a long time ago. There used to be two wireless carriers, AT&T and Verizon. That was it. And they came out with a C and D block at 1800 megahertz, so the others were at 850 megahertz. Anyway, I started on the finance team, and through a crazy set of circumstances, I ended up as the project manager building out the first 144 wireless sites here in Salt Lake City, Utah. It was an awesome experience. We were one of four markets that launched on time, and we were the only market that launched within our budget.

Shamil Malachiyev: That's a rarity. Usually two times the budget, two times the time, you know, the standard rule.

Stan Hansen: That's exactly right. I had a great team that I worked with. I think what was most interesting is that I ended up in this position because the two other people that were supposed to come in and take it didn't end up coming to Salt Lake City. So I went in as the interim person, and then after several months, we were on such a short timeline that they said, hey, why don't you just go ahead and run it and go with it. And I think the team I worked with was really kind to me, because I did not have very much project management experience. In fact, I bought a Project Management for Dummies book. We didn't have LLMs back then, so I couldn't just search everything and put it together. But it was an awesome experience, I worked with a great team, and it opened up a lot of doors for me. I moved from that role after a couple of years to head of strategic planning and competitive analysis for the market. And then I later ended up at CBS Viacom in a sales role for two, three years.

And through my time at Sprint, that's how I ended up at Wells Fargo. One of the people I had worked with called me and said: hey, look, we're going through this thing called Y2K, man. We're in a crunch, and we need project managers that can go get things done. I thought of you. Would you like to come work with us and solve some of the problems we've got getting these systems and applications remediated and ready for Y2K? So that was my journey. I started there, and after a year or two in that role, I ended up working as chief of staff for a guy that ran the customer service delivery side, the Banker Connection side of Wells Fargo. It was a good-sized organization. And then I later started shifting a little bit out of the procedural side, more into the technology arm, working with a group called Technology Connection that did help desk, different things. And then we moved into role-based access and controls, and some of the merger and acquisition work on the technology side. When we'd go out and acquire a bank, we'd do a lot of the sizing work, a lot of telephony work. That purview expanded, and I ended up with a pretty good-sized group there. Just a really awesome experience at Wells Fargo. I was there about 13 years and went through one of the largest financial mergers of all time, which was the Wells-Wachovia merger. I spent a lot of time out in North Carolina, and later, after all that traveling, my kids were coming up and I wanted to spend less time on the road, stay back home. So I took a job with Adobe.

And that's where I got into the technology and the sales side, leveraging some of the work I had done in procurement and technology and coming over to sales. That was the motion of how I got back in. I started with the SDR team and then inside sales, and made a couple of moves. I worked with some really, really great people at Adobe, had a great time there. Then I moved to Domo, which was an up-and-coming company. I wouldn't say it was a break-off, but Josh James, the founder of Omniture, who had sold his company to Adobe, actually started a new company called Domo, and I worked with him down there for a few years, doing similar work. But at the time I made the break from Domo, I went to a company called Pluralsight. It was a pretty small company at the time, and they were making the switch from a B2C to a B2B company. We had a really small B2B sales organization and not a lot of revenue coming through, but there was just a tremendous opportunity to go grow that business. So I came in and was leading the enterprise sales team, the new business team. That morphed, and after a couple of different changes I ended up running the commercial business, which started small, with seven people. We grew that team significantly, to several hundred million in revenue a year, a large organization. And then ultimately I worked into a role where I had all of sales: enterprise, SMB, commercial, mid-market. It was just an incredible run. We started with under $100 million in revenue; in my last year, our team did $468 million in revenue. That was over the course of about five or six years. We really saw an explosion there. The market was ripe for it, and it was a good move. And while I was there, we took the company public, and then later the company ended up going private again.

That was a good time for me to make a switch. I let the folks at Pluralsight know I was going to be looking for a different opportunity, and it took about seven months from the time I notified to the time that I left. And I came to Egnyte. And, you know, similar story here. We were just about a hundred, in fact just over a hundred million dollars in revenue. We're now in the 370 range and growing at about 20 percent, a little over 20 percent, with a nice 20-plus percent EBITDA margin. So the company has really made a turn there, and we continue to do very well in a very dynamic market, I would say, with lots of changes. It's been a good ride so far. So, a little history on me there.

Shamil Malachiyev: And what I like most is you have such a good wealth of experience from all different parts of companies. Because when it comes to being a good chief operating officer in the age of AI, you need to know the internal processes of different departments, what they consist of, what the intrinsic and extrinsic motivations at different levels are, what the processes are and how they're wired together, to be able to visualize it in your mind and then see what we can automate and improve. And for the listeners that don't know, can you do a short sales pitch for Egnyte?

Stan Hansen: Yeah, look, Egnyte is, first of all, an incredible product. We do a lot of data management; a lot of smart content management is probably the best way of saying it. Egnyte is data infrastructure that focuses on data security, data compliance, data collaboration, data governance, a lot of different areas that companies struggle with. So when you think of the importance of data, unstructured data specifically, it's: how do you extract the value out of the unstructured data in your organization, which is really the crown jewel? And we've been very, very effective in helping companies in various industries extract the value out of that data and go apply it to get an actual return on their investment. A lot of that has to do with collaboration: the abilities we have to audit and track who's accessed the data, the collaboration points where you have multiple people coming in on the same document in real time, working those into workflows to drive efficiency throughout the organization.

Shamil Malachiyev: And I think that's something a lot of companies can use, because even within my company, I always have people coming up from non-development teams saying: okay, those guys have GitHub to manage all of their files and everything. What can we use to drive the collaboration on data and documents for the other departments? And it looks like Egnyte could be that place for a lot of companies too.

Stan Hansen: Yeah, and I'll add one thing. We take a really interesting approach to ownership of the data. With a lot of the tools that are out there, the premise is that whoever creates the document is the owner of that data. Egnyte works a little bit differently. It works more like a file server, where you create folders, collaboration folders, that people work on, and the company owns that folder. You might be a contributor to it, but you're not picking up and moving data if someone moves positions or someone leaves the company. That data is there, the team's collaborating around that data all the time, and you can plug people in and take them out. It's the company-owned data that controls who has access and controls to it, as opposed to the creator of a document. And it's a subtle but fundamental difference that is really important to the continuity of people coming in and changing roles. It's been super effective for us and our customers. It is part of the secret sauce of what makes us great.

Shamil Malachiyev: And over the last two years, I think we've seen some of the most complex parts of AI technology start to make sense, and people have started to understand what it is and what it isn't. The whole crazy scare of 2022, 2023, like, my God, what is this technology going to do, has kind of subsided, and now we're in a position to analyze: okay, what do we do with this newly discovered alien technology that has hit us? What was your experience researching it and trying to think through how it's going to change the way we do business?

Stan Hansen: Yeah, you know, it's funny, and I'm going to date myself a little bit here, but I remember seeing the cover of a Time magazine, this was back in the 80s, and it was like, the Internet is coming, or something like that, I can't remember. And it does feel like one of those moments where a big major change is coming. And I would say that I don't know that we fully understand it, but you can feel it's a tidal wave, and it's coming. In my mind, even what we've seen over the last 18 to 24 months, the changes that have taken place and what's happening, it is going to be revolutionary. You can see it. You can see the applications that are out there. I think we're still right at the very beginning of a very long journey and a long transformation, but it is making people more effective. It is making them more efficient. I look at myself and some of the tools that I use just to drive my everyday work. It's changed the way I work completely over the last even six months. What's happened in the last six months has been phenomenal for me. It's been really helpful, and it's a tool that can make me more efficient and better. And the question is: how are we going to be able to do that at scale, in complex processes? I think that's the next evolution: how do we work this into workflows, and then how do we move into that agentic piece, and then beyond that? I don't know if I answered your question there, but that's how I see it: you see an evolution coming.

Shamil Malachiyev: And when we look at a company, how do we know that it's ready for adoption? Egnyte being right there at the layer where everybody struggles with the data: how do you prepare the data, how do you take ownership, make it reliable, make it usable? What does it take for a company in the current day and age to build that foundation?

Stan Hansen: Yeah, look, I'm fortunate in my role, I get to get out and I'm talking to customers all the time. And I'm seeing this firsthand, real time: what are the issues that come up around companies preparing for AI specifically? How do we get ready for that? A lot of people have jumped in really quickly, and a lot of people have gotten burned, because oftentimes people are leveraging models or systems, their data's not secure, they're accessing data that they shouldn't have purview to. It has caused a lot of issues out of the gate. And that's made a lot of companies actually tap the brakes and say: hey, wait a minute, we are not ready for this. Or: we only want a very small subset of people, or we only want really defined tasks. So they're not able to unlock the power of what's available today.

And specifically what I mean is, look, I don't want this to be an Egnyte ad or something like that, but I will say one of the things that we've learned, and whether you use Egnyte or something else: what we are really good at is defining access controls and permissions to different data structures. That's an area where it's really easy to facilitate within Egnyte, and so people do that. They build their data infrastructure in a way where you structure the data in your system, in folders and files, with controls on who has access to it. And Egnyte layers over the top. So when you bring in these other models, you can deploy the access and controls layers of Egnyte to go get that data. And what happens is, when you're doing a query or you're interrogating data sets or whatever, people are only seeing the actual data that they would have access to. And that is usually the first step that companies need to take. That's where they get into trouble. Because oftentimes, if you don't have a structured data set, you're putting data all over, and when you allow someone access to go interrogate that data, they can find it really quickly. That's one thing these LLMs are really good at. And sometimes they will present data that people shouldn't be looking at. And this could be compensation data. It could be strategic data. It could be merger and acquisition data. It could be HR data. There's a lot of data that we see that gets put in the wrong place, because there isn't an actual structure that companies are working with, or it's not easy to identify it and get that data in the right place.

So we help a lot of companies with that step, and I would say that is step one. Step one is: how do you get your data structured in a way that is actually ready for AI interrogation at scale? That's one area we focus on. We do a lot of that with our tool, but we also have a lot of partners that come in and do specific work with Egnyte. They use Egnyte as their tool to help people go clean up corporate data systems.

Shamil Malachiyev: Yeah. And I feel that everybody who's trying to be on the cutting edge of technology is faced with the same thing every single time: who can control the data? Okay, you've run a Zoom call and you have an AI assistant, and three people didn't show up. Are they going to get the transcript? That's just one example of the many ways in which you interact with a lot of data, and then, where is it going to show up? Are the AI agents that we use internally going to have access to that data too? So what would be the role that you see evolving within these companies to control the confidentiality, security and compliance? Is it one of the existing roles, or something completely new that's going to emerge?

Stan Hansen: Yeah, good question. I mean, different companies are handling it different ways, but for sure it's a data security issue, a data access and controls issue. So oftentimes it's either going to come under the purview of a CISO or a CIO, in that frame, depending on how they're structured. That is a primary risk and a primary concern. One of the areas where you see a lot of data leakage, or even data breach, is your internal employee base. And so getting that right is really, really important. That's probably the number one risk out there: people getting access and taking data that belongs to the company, not to an individual, an individual taking and moving that with them to another company, or using it in some nefarious way. That's a serious problem. And that's one of the areas that we have focused on over the last couple of decades: how do we identify that? How do we help stop that? How do we know when it is happening? What does that look like from different patterns and different things? We can help people with that as they're trying to figure out: how do we control the data, and how do we keep the value that they've developed over years of curating their process and their intellectual property safe and secure and in the hands of the right people?

Shamil Malachiyev: And internally at Egnyte, do you oftentimes have employees that come to you like: Stan, I've vibecoded this solution that can help us do our work quicker. How do we get it past compliance and legal? Is this something the company is okay with? Because with current tools, rapid prototyping is getting easier and easier. Do you get those situations?

Stan Hansen: All the time, yeah, all the time. So look, Egnyte as a company, we've put out some tools out there. We have Gemini, so Gemini is one tool that we have. We have Claude for, I'd say, a large subset of the company. Those are two different tools. And then we've got other people, other groups, that are using other tools like ChatGPT and others, on the enterprise versions. Basically, we have all the major LLMs in some place being tested or worked on at Egnyte. And there are two different things there. One is: how do you use that in conjunction with Egnyte? That's probably something we could spend a little bit of time on, the approach that we've taken with these LLMs and how we see the future. But to answer your question: we're pushing our employees to look at what they're doing on a day-to-day basis and say, how can we use these LLMs to speed up this repetitive work that's out there? That's what I would call step one, or maybe step two, with data preparation as step one. What are some of these mundane tasks that the LLMs are really, really good at, and how do we get them to start doing that work immediately? And it's hard for one person or one team to go in and do that for an entire company. So what we're trying to do is push that out to the individuals at Egnyte, to the teams and team leaders, with a framework that says: hey, this is what we expect, here's how you should be using these tools, this is what you should be looking at. And then as things start to surface, we start talking about: does this fit within the model? Is this something we want to do? Should we pursue it and put some time here? Should we standardize this and start rolling it out broader in the organization? That's the approach we've taken.

You mentioned rapid prototyping. I believe that is the key. And that's actually been one of the keys with our customer base as we've gone out and sold Egnyte and AI solutions. Over a third of our dollars that we sell today in new business and new opportunities are actually coming from Egnyte AI solutions or in conjunction with AI solutions, which is a lot.

Shamil Malachiyev: Tell us about this.

Stan Hansen: Yeah, it's a lot more than we thought it was going to be. And we kind of started on a journey, and maybe that's the better place to talk. I'd say probably 18 months ago, I thought I was probably going to lose my job, related to AI. I don't know whether anybody else has experienced this, but we had kind of a rev one of our AI solutions out there. And I would say that, for sure, they were very immature. We were scratching our heads like everybody else, saying: how can we bring this into Egnyte? How do we make it valuable? And we took an approach where we didn't want to overcommit or over-promise. We had an AI solution, we had a package, we went out, and we were basically selling what we could do. And what we realized is that everybody else was selling something way beyond anything that we could. And look, we've been in the business for over a decade. We have extensive experience with machine learning, in-depth knowledge on that side. I feel as a company we're very advanced. But people were selling these grandiose visions. And I would say that we did not produce anything that was attractive or sexy or whatever. People were like: yeah, hey, that's great, but over here, they're going to transform everything in our business, and they're going to do it in the next 12 months. So we didn't sell a lot.

I've got to be honest, I mean, it was like we were not selling any AI, and we could see that companies were progressing and selling a lot of AI solutions. And at one point there was a serious question, like: hey, you guys, you've jacked up the strategy here. What are you thinking? We're not doing anything. And I would say we were about six months into that, and it was getting tense. And finally, we had our first customer that we had walked, basically, that had picked another solution, come back. And it was a good-sized customer. We were selling AI, but in small pieces; we actually were having success, just not success at scale. But this large customer came back and said: hey, we want to sit down and revisit. We're revisiting all the vendors that we worked with before. We made a choice, we bought into a single ecosystem, and it has been a disaster. We spent a lot of money, and we haven't delivered anything. And we need help here.

And the model that we had decided on is: let's go give quick wins for our customers. Let's not promise the world. Let's just focus on some of these clear tasks that are out there that they need help with, and let's go deploy something, so they can build confidence with their executive teams and their boards to say: hey, we want to spend more on AI, because we can get results that are actually producing a return on investment in our company. Fortunately, I mean, honestly, I think if the companies wouldn't have recognized that, I don't know, three or four months later, maybe I'd be working at a different company. I don't know. I mean, it wasn't that bad, but it was tense. The question was whether we should have developed a different strategy. But that flipped very quickly. We saw that first customer come back, and shortly after, there were several other customers coming back and saying: hey, we're not having success with the AI solutions that we purchased. These commitments were made, and there's no way we're going to hit them. We haven't even gotten anything off the launch pad. So that's what we did. We came in and said: let's focus on small, quick wins. Let's help these companies build confidence with their boards and show that you can actually get a return on investment quickly. And that's been a winning model. We're definitely not straying from that anymore. That is core to who we are. In fact, under-promise, over-deliver is a way better strategy than over-promise and under-deliver. And it's accelerated our business in this space materially.

Shamil Malachiyev: And I don't know how to put it, but I feel like it takes balls to take that stance. Because you look at the industry, all of these startups, all of the money going to companies, what companies are getting funded: everyone over-promising, nobody knows what they're delivering. A lot of AI, trillions of tokens used, look at us, we're so cool. And for a company to be like, well, now we have to compete with that. And you don't know what their churn is like, how long the customers are staying, how satisfied they are with the solution. You just see everybody over-promising on the idea.

Stan Hansen: Yeah, yeah. You know, here's the thing. Over time, and I've seen this play out over and over and over again with many different companies that I've worked with and bought software from: the over-promise, under-deliver strategy is short-term. The only chance you have to win on that is if you over-promise, and then something miraculous happens, and the market changes, and you get lucky, and you're able to build whatever you promised, and it works out. But that's like one in a hundred, or one in a thousand. I don't know. It's not very often.

And the hardest part about that approach is that usually, when you go in and you sell to a company, you're selling to a champion. And that champion is like: hey, I'm all in on you guys. I want to partner with you, and I'm trusting you to help me solve these issues that we've identified in either a proof of concept or a proof of solution. I believe you can help us achieve our goals as a company, and I'm going to put my reputation and my career on the line and buy your software. I'm buying the partnership, the relationship, for you to help us go solve these issues and drive value in the company. If you over-promise and under-deliver, you burned your most ardent supporters. You burned your most passionate people that are all in. And so it is crushing. With your champions that are out there, you cannot take that risk, because they will never back you again. Ever. The majority of our leads, the majority of sales that we get, are from people that we worked with at one company; they pick up and they move to another company, and they're like: we want Egnyte. From a marketing standpoint, that approach has been lights out. That's our number one lead source. And our second greatest lead source comes from someone that calls somebody and says, hey, how are you solving this problem? And they're like: yeah, we use Egnyte. They've been great. They came in, they helped us get set up. We were able to achieve a deployment, an enterprise rollout for the whole company, within a month or two, sometimes weeks or days, depending on the size of the company. It's been a really good experience. And that immediately catches the attention of some prospect out there: who should we talk to over there? They make the introduction, and then we're able to go in and see if we're a good fit. And if we're not a good fit, we'll just tell them: hey, not a good fit on this one. Because we also don't want to burn our champion. If we're not quite a fit, we just tell them. And our guys will still refer customers to us, because they know we're not going to go sell something that doesn't work.

Shamil Malachiyev: Candor and integrity. Basically anything long-term.

Stan Hansen: Yeah, it's the long-term strategy, and for us it continues to allow us to accelerate our growth. In a market right now where the majority of companies are really struggling, growth rates are coming down, I am just so thankful that our customers are the ones that are helping us propel our business. And our business is actually accelerating. We were at 20%, then 21%, 21.5%, 22%. And an extra percent of growth is tough right now in this market. I think anybody in SaaS will tell you: hey, if you're at that level and you're accelerating growth, that feels pretty good. So we're proud of that.

Shamil Malachiyev: And before I ask about the rise of LLMs and how you've utilized them, I wanted to come back to the point of when people come with suggestions about what can be improved. Speaking to other leaders, what I've found out is there is a lot of resistance. If you look at companies like NVIDIA, OpenAI, Anthropic, they've done a terrible job at building trust with the final users of AI. In order to drive the prices of their shares, they're like: AI is going to destroy everything, you're not going to need employees, the AI is going to do all of the jobs, it's a superpowered technology. And what that's done is create a narrative where, if you are the person trying to use AI to improve the processes, basically you're probably trying to get everyone out of their jobs. So why should people be the champions and supporters of AI technology internally? How do you face that kind of challenge and help people understand that it's a tool to augment the work, to help you remove the boring parts, instead of people going: no, we're not going to touch that, because if we do, we're going to be out of a job, because that's what OpenAI says?

Stan Hansen: Yeah. So the first thing: I look at the number of improvements that we need to make in our company, and I wish I could say it was like one, two, three, something like that. But it's hundreds or thousands of tasks long. It feels enormous. And the biggest challenge that we have as a company is prioritizing the most important stuff that we want to optimize, fix, expand, grow, at the top of that list, and putting the right resources on it. And what we're seeing with AI: there's no shortage of work. At this moment, and in the foreseeable future for Egnyte, and I'll just speak for us, we're about 1,300, 1,400 employees, somewhere in that range. If money was no object, we'd go hire another thousand people, or more, because there are just things to go do.

What I'm seeing, and what I'm trying to coach our people on, even my own family, I've got four daughters and a son, and they're all between the ages of 20 and 30 years old, is: you have got to engage with these tools. All my kids have Claude, and they all use OpenAI. We talk about: what are you doing to implement those tools to make yourself more productive and make your life better? And we go through that. But let me go back to your question on employees. In my mind, it's getting people comfortable. And the only way that they truly get comfortable is if they engage and start using these tools. That's it. Because then you can see it's not job elimination. And I agree with you. Even some of these companies that are reducing force, and there are a lot of them coming out saying, because of AI we're going to reduce force, okay, maybe they can do that. But they don't really need to do that. They can reposition those people into other areas, and they will. You'll see a reduction in force, and then they will start hiring again immediately. And the reason they'll start hiring is because they might need to change where those people are. They might need to optimize, because they've cut out some mundane task that can be done with an LLM or an agent. And that's what I would expect as well.

But at our company, it's not like we are looking at "we're going to deploy AI and cut staff." I mean, we have an AI solution, it's a big part of our product, and we incorporate these other models and how they come together with Egnyte. But in our case, all we're doing is getting more work done, quicker and more effectively and efficiently. And for the foreseeable future, I'm not seeing a reduction, a mass layoff or reduction in staff, because there's a lot to do. We're a growing company, we're healthy. As we ramp this up, it's going to drive more revenue, it's going to drive more profitability for the company, and that's exciting for me. It opens up growth opportunities for our employees. That's probably one of the most important things. I don't see a shrinkage. I see an actual expansion, an expansion into higher-quality, more complex work that is augmented by these AI solutions.

Shamil Malachiyev: And I think that's specifically why you mentioned the kids and how you're ensuring that they know the technology. Because it seems like in the future, people are not going to be replaced by the AI, but rather by people who can do three times more, because they've become accustomed to using the AI tools. And the funny thing is, everybody expects it like: oh, it's going to be here, you're going to use the tools, you're going to work less. Do you ever feel like, when you have so many AI tools, you work less? Or do you work more?

Stan Hansen: You know, that's a good question. I would say I work less on mundane tasks. I'm working more on more strategic, more visionary, more complex tasks. And again, my belief is that the people who are going to really change the value of who they are and what they bring to any organization, or to themselves, or to their family, are going to be these people that know how to engage with these models. And I'll say "these models" because it's not just going to be LLMs. There are going to be SLMs, there are going to be other very specific micro-models, that we also think are going to be used in conjunction with the LLMs, that are going to drive the most amount of value.

There's a specific set of skills that I'm seeing, and I can talk about what I've seen personally, because I wouldn't consider myself an expert, but I feel like I'm evolving with these tools as well. People that can go in and converse with the LLMs and frame out what they're trying to accomplish or what they're trying to solve, and then converse with an LLM to build into: what's the best way to attack that problem? And then ultimately build an agent to go tackle it, or a prompt, or whatever that might look like for them. Prompting is kind of level one, then agents, and multi-agent. Those are the people that are going to be successful. So that's one key skill: how do you engage, what does that look like? In the beginning, in the short term that I've been doing this, I would tell the model exactly what I wanted. I would come and ask a very specific thing: hey, I want an answer to this question, or this is how I want you to solve this problem, or this is how I want you to help me write this memo. But now I'm thinking more holistically, to say: hey, here's what I'm trying to accomplish. What should I consider when I'm tackling this task? How would you come at this problem? What are three or four different strategic ways of coming at this problem? And then I can take that information and learn, and then I can go to the next level and the next level and the next level. And I'm getting much higher quality, more strategic solutions and higher-value output. I'm not just solving one small task now; now I'm actually moving into multiple complex tasks. So that's one.

The second skill that I think is going to be really, really important is someone that I would call a BPA, a business process analyst. And this would be someone that can come in and look at the workflow of an individual, or look at a process, and define it really well, define it clearly, and then work with one of these models to go build an agent to go do that work. It's hard to get that right. It sometimes requires multiple versions, because you're like: hey, this is working really well, but I'm getting bad information. Or: it's hallucinating over here. Or: we've got the wrong people accessing this data. So it's fine-tuning that process. But that is another skill that I believe is going to be highly valuable over the next at least four or five years. And there will be some evolution of that in the future, but those people are highly valuable: the people that can come in today and look at what we're doing and put it into a defined process, because then you can automate that process. And it's awesome. So those are two skills, if someone is trying to figure out how to survive in this world, that I'm seeing today, for sure, that are high-value.

Shamil Malachiyev: Yeah, and I'm hearing almost similar things. I've been talking to a lot of heads of procurement departments that have already automated a lot of their processes, and they always say the first step is to understand all of the processes, to map them out, to then be able to see... because it's really hard to give oversight to agentic AI if you don't know which parts of the other processes it interacts with, what the effect is going to be on this thing or the other thing, what the problems are. And when it comes to having an idea, like, okay, we have a procurement department, let's automate this: software costs are going down, and you're thinking, we have a lot of developers, shall we develop something from within, or shall we just buy something that somebody else is building? How do you tackle those kinds of decisions?

Stan Hansen: And that is a tough one. My personal opinion on this is: any time that you're building software, you've got to be careful. We work with a lot of really, really smart customers, and a lot of those customers have built internal systems. And those internal systems, usually for a period of time, are amazing. They're getting awards, they're up on stage talking about how great it is, and it's awesome. But what happens is, technology changes so quickly, and usually these companies are not development companies, they're not software companies. They're construction, or financial, or manufacturing, whatever. Their core competency is, let's just say, manufacturing. And what they realize is that this software system they built, although it was great and customized in the moment, now the market has changed underneath them. New technologies have come in, and they don't have the skills, the capabilities, to keep it updating and innovating. And they get aged out. And their challenge is that they come in and they're like: my gosh, now we've built an entire system and process around antiquated technology, and if we're going to survive, we have got to migrate this thing off. We've got to have somebody else come in and help us. So I think any time you're building software, you've got to be really, really careful, because you can get into a cycle that's dangerous.

Can you build agents, and tie in multiple agents? Yeah, I think you can. But they're usually built on top of a platform that is open, and open to working with these platforms. And that's the approach that we've tried to take at Egnyte. There are some companies that have come in and said: hey, look, our strategy is that we want you to buy our system, and we want everything in one ecosystem, because we can do everything for everybody. And at Egnyte, we said: hey, how's that going to work? Because these different models are going to be very effective in very specific areas. They're not all going to be the greatest at everything, and they're not all going to grow and develop at the same pace at the same time. We've already seen this. OpenAI came out of the gates, my gosh, great thing, they're going to own everything. Now you've got Claude that's come in, they're taking over the world. And then you've got Gemini that's come on the scene, and they're taking a big piece of the business. But they're all different. And so what we decided at Egnyte, we said: hey, let's focus on our core competencies. Let's focus on data security, data governance, data collaboration, that data platform. And let's go out and open our system. Let's go work with every single one of these models, where you can come and deploy any model in Egnyte. We'll focus on things like the data. We'll also work on what we've helped companies do for the last 10 years, which is understand their data: metadata tagging, that contextual and semantic layer of what's in your data. So when you're prompting your LLMs, or any model, even a model that you build, you come and deploy that in Egnyte, and we can deliver up data to make that model more efficient and deliver results from the data that are a lot more effective. And that, in my opinion, has been part of the reason why we're crushing it on our AI products and solutions: because people are getting better answers. And we're not trying to say "only use an Egnyte AI model." No. We're saying use any, bring your own model. Go create your own model. Go do it with Claude, go do it with Perplexity, go do it with OpenAI, Gemini, whatever. It doesn't matter. Go do whatever it is you want to do. But let us help you govern and audit that data. Let us help bring the enrichment of what we've already got, from all the metadata tagging that we have out there, the classification, all of those items. So one, you don't have to burn as many tokens when you're going and querying that data, and two, you get much more contextual and sound answers, less hallucination in the model. Again, we believe that's a winning model. We feel like we've got a good path that's producing good results for our customers, and it's been pretty exciting.

Shamil Malachiyev: And on the product side for Egnyte: you told us that you've worked with machine learning for a long time, and then in 2022, 2023, all of the LLMs started coming out, we got vector databases that store data differently, all of this. How was Egnyte's own transition to making use of that new technology within its product, while at the same time trying to sell it to customers and explain it to them, to help them adopt the new ways of working? Because people like the comfort zone: we've been doing this, it works, don't make us change anything. But the world's changing with this new technology, and you want to help the customers adopt it.

Stan Hansen: Yeah. So first and foremost, we've come at it from a couple of different approaches, and we've talked about a couple of these. One is data preparation and interrogation: actually showing them, either on their data, or especially if they're a current customer, we can go in and show them the capabilities that we've built. And more importantly, we can show them the results that we're getting when we combine these models with Egnyte, and how Egnyte approaches the overall problem of getting the best answers, working with the best models for specific industries, for specific problems that you're trying to solve. That's helpful; that's number one. The second thing that we're doing is what we call rapid prototyping. That's not a new term, but we can go in quickly and solve a piece of the problem right out of the gate. We don't come in and say: hey, if you want to go with an AI solution, in 12 months we're going to deliver X. We don't even like to talk that way. What we like to do is come in and say: okay, what are the issues that you're dealing with today, and what are the steps that would drive efficiency for you quickly? So we'll come in, take a problem, break it down, and go deliver a solution, sometimes the same day. It might be two or three days later. And we're not even trying to do that work ourselves. What we're really trying to do is say: here's the art of the possible, for your team members to go do this work. We're not heavy into professional services. I mean, less than 2% of our revenue comes from professional services. We're more focused on: how do we sit down with your teams and show them what's possible, and then get them using Egnyte solutions? So you're not bringing in a consulting firm that's going to be ongoing into perpetuity, but you can actually go deliver results the same day, or within a week. And as the problems get more complex, it could be a month out there. But they're getting wins really quickly. We help them with the first couple of those, then we sit down and frame out what they're trying to do, help them with the roadmap, and then they're off to the races. And that is what's driving and perpetuating the growth for us.

And they can see it. The most important thing that I've learned is that if you want to go drive success, and this is any AI company that's out there today: you've got to deliver results within days. Do not focus on "we're going to go deliver something in months or years." It's too far out there. Go show them what you can deliver today, and then build off of that. It's awesome, and they get excited. And there's nothing more powerful than an excited customer, when people can see: my gosh, this is going to change my life, it's going to change the productivity in our business, we're going to compete better. They dive in, they work with you, and those are the best customers. The best customers are the ones where you're collaborating back and forth: hey, this is what we accomplished, we did this, this looks awesome, what do you think about it? What would be our next step? Okay, here's what we do. And we can also give some context: some of our customers are really open about what they've done, with Egnyte and with any AI solution, and they're like, yeah, go ahead and share it with people. And when they do, we share that. Sometimes we bring customers together to talk about where they've had success and how that might apply to someone else's business. It's been a fun thing to watch.

Shamil Malachiyev: And when we talk about different departments and the ripeness, in the current day and age, of using AI to boost a department's results: I know you have at least a sales focus, because you've been doing that for a long time, and that's an important part of any company's work. There are solutions out there that help you create your pitch better. How have you noticed the sales department getting boosted with AI and the current tools?

Stan Hansen: Yeah, so look, there are a lot of AI initiatives that we've built specifically inside Egnyte on the sales side. Now, we use a lot of different tools out there: we use Outreach, we use Gong, we use NotebookLM, we use Salesforce. There's a whole complement of different products and solutions that we have, and we use a lot of the AI features in each of those independent products, sometimes. But what we're finding is that when we bring all that data together, in conjunction with the data that we have at Egnyte, we're getting unbelievable results. So let me just take deal acceleration. We'll pull in a lot of the data that we have from Gong, from Outreach: what communications have gone out there? We'll look at what stage the opportunity is at, and then we'll go in and say: hey, have we covered all of our bases here? Is it at the right stage? Do we have a meeting set on the books in the next three days with this customer? If not, did they try to get that on there? We'll start looking at the transcripts, and it will help us on the negotiation side: let's go back through all the transcripts of all the calls that we've had, all the email correspondence. What has come up? What have been some of the key areas of focus for them? We can quickly identify those, and we can build an actual brief. It's helpful. It's also helping us right now on kind of evidence-based management coaching: it will list out a string of items that we should be helping our managers and reps with, in terms of what's going right, where they can optimize, what they need to skill up on, how they've done on message delivery. Are they including the key elements of the Egnyte value proposition? Are they doing the right discovery questions? Are they capturing that information in the right way? There's a lot going on there.

We have another one that's been a game changer for us. We have very, very strict rules of engagement. At Egnyte, we have a direct sales team, and we also have a partner sales team. And the partner sales business at Egnyte is really growing; it's explosive growth for our partners, because they've been able to go and sell Egnyte, have the same level of success that we have, and it's been a benefit with their customers. They're like: my gosh, we're onto something here. We've gone from our partner channel being about 8% of our business a couple of years ago to over 30% now.

Shamil Malachiyev: Wow.

Stan Hansen: Yeah, it's growing. And part of our success there is making sure that the way we partner and engage with our partners is highly integral, so that if a partner brings one of their customers and introduces them to Egnyte and we're doing the work, we want to make sure that partner gets paid. We want to make sure they get compensated for bringing that to us. And there can be conflicts, because we've got a couple of hundred salespeople out there, working hard to go introduce Egnyte to customers.

Shamil Malachiyev: Attribution.

Stan Hansen: So we have a very robust rules of engagement. It's a 70-page document, with examples and different things. And there are two different things we can do. One: we can go in and quickly identify whether we have an open opportunity with a company, which used to be really challenging, because sometimes you'd have to get the name spelled right, you'd have to do all sorts of things. We've got a good tool now that's much more effective, even with a misspelling, or a different country, or whatever. It does a much better job of helping us identify: yes, we have an active engagement here, this deal has been registered with this partner, and there's no confusion. So that's one. The second thing: we might have a partner deal, or we might have a direct deal. This happens often, where we've worked with someone directly, and even though we ask the question, hey, do you work with any partners, any managed service providers, any value-added resellers, any distributors, early on they'll say no, and then later on they'll come in and be like: well, hey, we do work with this MSP. And so you've now got a direct rep that has spent six months, or three months, or two months, working with a customer, and now it needs to get picked up and moved over to an MSP. And sometimes it can go the other way: sometimes an MSP is working with somebody, and maybe the customer wants to switch to a different MSP, something along those lines. We have to be really careful, and we have to have very, very strict protocols and rules around what this looks like, because we communicate those to our partners, and we communicate those to our sales reps, so there isn't a mismatch in expectations. They know exactly how we operate. And getting to the bottom of that used to be painful. We can now just type in: hey, this is the scenario, this is what happened. It will quickly go into this document, it will cite the rules, it will say: this is how I would approach it, this is what the resolution should look like. And if there's a problem with it, they can raise it up to sales operations. We're probably on version three or version four of this thing, and we've probably spent maybe 30 or 40 hours of development on it. But here's the thing: it is saving well over 100 hours of management time going through this, and probably way more than 1,000 hours for reps trying to understand what the answers are. And it just continues to get better and better.

That's one. We have another one on the product side, where you can ask any product capability question that you might have needed a solution engineer for, and we can get all that information quickly. We still bring in solution engineers to piece all this stuff together, but sometimes a sales rep might have a question that they need a quick answer to, and we can get it, with confidence, with these solutions. So that's just a couple of areas where we've been successful. Hopefully that's the area you wanted me to go down there.

Shamil Malachiyev: No, that was perfect. And as one of the questions: let's imagine there are other C-level executives in companies right now, with a lot of pressure from shareholders and other executives to deliver value on AI. And they see a lot of over-promise and under-delivery, and they're trying to find the balance between all of that. How would you suggest, maybe, they find that silver lining?

Stan Hansen: So there are a couple of things I would do. First and foremost, just on a personal basis: what action can they take immediately, themselves? If I were them, and I am an executive, the one thing I did is I got a personal Claude account. I actually have a personal Gemini, Claude and OpenAI account. I've done a lot of work on my personal stuff: my personal email, my personal calendar, my personal to-do list. I own some properties and different things, and some family LLCs, and that has been a hairball, because I'm so busy in my day job that it's really hard to go do my personal life. And when I'm done with work, I want to go hang out with my family; I don't want to go right back to doing more work. So getting a Claude account, and going in and creating my own MCP server, and figuring out what model to use to build whatever task or agent I'm trying to do, has been a great journey for me.

And it's funny, I've learned a couple of things too. I was on Opus 4.7, and I kept getting a timeout: you've run out of credits, you need to wait four hours, or until tomorrow. And I'm like, this is Stan Hansen, I wouldn't say the most technologically sound guy on the planet, but how could I be burning through this stuff? So even just simple things like: okay, well, maybe I shouldn't be using Opus 4.7, maybe I should be using Opus 4.6, or maybe I should be using Sonnet or something else. Even just asking those questions: hey, this is the problem I'm trying to solve, which model would be best to optimize on? Okay, it brings you back an answer. Hey, I want to build this connection into this set of data that I have for a couple of properties. Okay, build that. You're getting this error: okay, hey, I just got this error, how do I fix it? Let me tell you how you fix this. Here's what you do. In fact, do you want me to do that for you? Okay. Yes.

So the reason I'm giving this high-level experience is that I've gone down this journey, and I've probably spent, I don't know, four or five hundred hours now, back and forth with these models, building stuff for my personal self and some of the family things. My kids, I don't know whether they think I'm funny or whatever, with how we engage with some of these tools, because I'm trying to get them into it and have some fun with it. But you start to learn the capabilities. And what that does for you as an executive is that now you can actually have a seat at the table. And you can talk to your developers about what's possible, what you've experienced, and what we should be looking at inside the company. Hey, I've had these issues. Have we solved these? What does this look like? And it will accelerate your company's journey. If you have a basic understanding of these tools, how they work and how they apply to your life, those lessons are applicable in multiple different areas. You just need to connect the dots: this is how these tools work, these are the results they can produce, this is the potential they have. You know what? We have an accounting department where we could apply that same thing. I can now capture all these expenses from email and from text messages, and it puts them into a QuickBooks-ready file. It just comes in, or I just click the email, it picks it up, puts it in a format that's readable by QuickBooks, and then uploads it into my QuickBooks setup. It's cool. It's like: okay, how could we apply that in the company?

Or all of the key elements of emails that are coming in, and action items; that's another one. I now have, twice a day on my personal email, a run that will go out and harvest the emails that have come in. It will categorize them: hey, you need to respond to these. These are meetings you have. This is personal. You've got a birthday coming up, make sure you've got that. It's helping me organize my life in a better way. And it's also helping me understand what the options are: how can we do that with Egnyte solutions? How can we leverage these tools at Egnyte? How do we leverage the Egnyte infrastructure to make it even more powerful? And one thing I have: at Egnyte, everybody gets an Egnyte for Life account, and there's a Claude connector for your Egnyte for Life account. So I can go in, and I'm using Egnyte with Claude, and I'm seeing that interaction the way our customers would be using these solutions. It's awesome. I'm getting the power of Egnyte, I'm getting the power of Claude, I'm solving my own problems, and it opens up a different level of discussion with customers.

That's one thing I would do. And the second thing I would do, if you're an executive: you've got to get that power into the hands of your employees. So figure out, if you've got AI solutions, how to get your AI solutions into the hands of your employees, so they can start using them in conjunction with all of these other models to go produce results. And it will get you to driving value much quicker, for your customers, for yourself. And ultimately, in my case, it's one and the same: I'm an Egnyte customer for my personal life, so I'm seeing what value I can derive from these combined solutions with Egnyte. Where does Egnyte play? Where does it add the value? How does it help me? It's on a simpler, smaller scale, but it's helping me identify the concepts so we can apply them for our customers.

Shamil Malachiyev: I'm loving that. It's almost like: practice where you preach, get your feet wet in whatever you're working at. And I guess that rapid prototyping experimentation gives you a real voice at the table when somebody is talking about these features. You've done that, you've tried it on a smaller scale, but still, the technology, the wiring inside, how things work, is going to be similar.

Stan Hansen: Yeah. Just on that: we had a customer come to us, and at Egnyte we work a lot in the construction industry. We have this capability where, if you're going out and taking pictures on your cell phone or mobile device, we'll geofence it, and then it uploads those pictures into a specific file, so it knows what project you're working on, and it helps with that. And one of our customers came to us and said: hey, you know what would be super cool? If, when you took those images, you could extract key pieces of metadata, or tag those photos quickly. Because they'd be more valuable. Because what I have to do right now is go in and say: okay, I take a picture, construction site, okay, we'll tag it as a construction site. Second thing: it had a bulldozer out in front, okay, bulldozer. Or it's a concrete structure, okay, concrete structure. Or it had safety tape up around something, okay. And they came to us like: look, we're tagging 50 things on these photos. And it's like: okay, let's go. That's something that we can do quickly. So if you want tags on those images: let's bring those images in, same thing, you go take pictures, put them in the project folder, and at the same time, let's make sure that we're loading up simple metadata tags on the areas that are key to you. That's something we were able to develop in like two days. They've got a list of things, and now they're off to the races getting some of the key pieces. Now we're going into V2, V3, which are going to be more complex: okay, well, now we want to know a thousand different things in the photo. That gets more challenging. But at least we've proven the concept on the 30 or 40 things that they want, and now we're working into: okay, how do we capture more subtleties in these images?

That's an example of the rapid prototyping that we've done from an image base. Same thing on document extraction. We have a lot of financial services customers, and there are very key elements in these documents: getting some of the contextual pieces out, the summary of the document, all that. But even more importantly: what are the key pieces of information that you can key off of later? So we can go build these agents, bring them into Egnyte, they can go extract this data really quickly, really effectively. We can test it, make sure that it's performing at the right level. And then it's like: okay, well, let's do V2, V3, V4, V5. And it just starts. You can take a task that someone is spending a lot of time doing, and then you can look at the next task. Okay, well, we've got that solved, that's great. Now let's move to the next one, and you go solve for that, and the next one. And then you build these agents to come work together. It's going to be amazing.

Shamil Malachiyev: It's almost like creating new startups within the company. You look at all these problems, and then you solve them with separate pieces of technology that interact and integrate with the main piece.

Stan Hansen: Every day. Every day there is a startup company starting within a company. Because you're actually building something: you're building a piece of software to go do something, you're building an agent. And every once in a while you hook into something that's really powerful, and you're like: whoa, great, we're onto something here. I didn't mean to cut you off. I could go on and on and on, but we are seeing innovation at a completely different level, in Egnyte, and in our customers' businesses. And I think that's because the art of the possible is now possible.

Shamil Malachiyev: And I think we'll still be able to leave some for the follow-up episodes coming in the future. Because the technology is developing so much that for the episodes we recorded half a year ago, everything's changed so much. There is just so much new context, and things that have been tried out and tested, to talk about. And I think that's something we might do with you, Stan, as well. I want to thank you so much for your time and for all of the knowledge that you've shared with us today. This has really been an exciting and amazing episode, which is going to be extremely helpful, I think, to a lot of people out there trying to navigate their way into what's the art of the possible, and how to stay grounded, keep your integrity and play for the long run when using all of this technology. So I want to say thank you, for myself and for all of our listeners.

Stan Hansen: Well, listen, thank you so much. I appreciate it. We were fortunate we had a chance to have a couple of conversations. Super insightful for me. I've learned from you, so I just want to say thanks. I'm now a follower, loving this thing. It's been awesome. And you've had multiple great guests on here. I'll be surprised if I make the cut for a comeback, because you've got some good people out there that have been on. But I'd love to come back anytime. So thanks so much.

Shamil Malachiyev: I'm sure you will. Thank you so much. And this season is going to be out starting in June. I'm pretty sure this season is going to be even more exciting, because it really dives deep into the trenches. Thanks so much.

Stan Hansen: Awesome. Thank you.