AI Insights: Egnyte CEO Vineet Jain on Future Innovations

0:00What can 160 years of experience teach you about the future? When it comes to protecting what matters, Pacific Life provides life insurance, retirement income, and employee benefits for people and businesses building a more confident tomorrow. Strategies rooted in strength and backed by experience. Ask a financial professional how Pacific Life can help you today. Pacific Life Insurance Company, Omaha, Nebraska, and in New York. Pacific Life and Annuity, Phoenix, Arizona. Welcome to the AI Chat Podcast. I'm your host, Jaden Schaefer. Today on the podcast, we have the pleasure of speaking with Vineet Jain.

0:36He was born in India and is a graduate of Delhi College of Engineering. Vineet has over 25 years experience in Silicon Valley's B2B tech sector. So after roles at Boots PLC, FetchTel, and KPMG, he founded Valdero and later co-founded Ignite, which is a cloud-based collaboration platform, which is doing some really incredible things in AI. Under his leadership, Ignite has collaborated with tech giants like Amazon and Google and has grown into a global enterprise. Vinit, welcome to the show and thanks so much for coming on. Thank you so much, Jaden. Glad to be here. Yeah, it's amazing to have someone with your experience and background on the show today.

1:15Would you mind telling everyone a little bit about what it is that Ignite does, what the problems that it is solving for customers are, and kind of where you guys are going with it? Sure, absolutely. So the idea of Ignite was born by our own experience working in enterprises before of information access, information sharing, especially for a knowledge worker, and the amount of friction that people had to deal with. And the fundamental construct used to be, in a typical sense, a file server that companies had people were accessing either remotely or working in the office remotely being VPN. And so our first foray of how we built Ignite was to say, okay, can we replace this with a cloud-based, multi-tenanted, hosted architecture, which became the underpinning of Ignite.

2:09And that evolved into, in Gartner speak, content collaboration. And then we started adding to that capabilities related to content security and governance. Because on one hand, the end users care about value maximization, i.e., what can I do with my content? Right. I quickly look to the content. But with increasing amount of data leaving the enterprise, i.e. the cloud, the growing concern around risk management, risk mitigation has become paramount. So how do you take these two seemingly divergent constructs and cobble them together and offer as a platform to the customers has been a big part of what we have driven this company's growth with.

2:45That's amazing. That's really impressive. And, you know, you talked a little bit about kind of the beginning and where it's going. I mean, I'm curious, how did you kind of navigate the transition from like a simple file sharing solution to this like, you know, AI powered platform that you are today? So, you know, while, of course, AI is the big topic du jour, I kind of joke it's the flavor of the month, but I think this is going to last longer. Right. You know, I would say for the past 10 years, if I say so, we already have been leveraging techniques related to machine learning and NLP or natural language processing to what today I would call as AI-enabled features.

3:32things like detecting data exfiltration, safety violations, compliance management, and then things like executing data management policies, which run the gamut of how do you identify what is redundant, obsolete, trivial, and still information, and then have a policy-driven engine for retention, archival, deletion. These are the things we had been doing for a while, using a variety of statistical and deep learning techniques since 2013. And then, of course, on July 26, because, you know, this is a topic du jour, we announced some fairly interesting generative AI-driven capabilities, which I'm happy to talk about.

4:09Yeah, yeah. Would you mind telling us a little bit about those generative AI capabilities and what you guys are building there? Sure. So, you know, on one hand, as I was telling you, we've been doing these things related to what has become, you know, the deep topic today, which is AI. just to add to other things we were doing anomalous user behavior detection learning about risk signals coming in from a particular location a particular person looking at their user behavior those things have been baked in but what we announced the other day which was actually on july 26th was how do you do content synthesis beyond document classification but that goes into document summarization and answer extraction.

4:57So it is demonstrated in capabilities like, I'll give you simple examples. There's an internal chatbot that customers can use to query a document or a set of documents. A good example is you have a 26-page report. You could run a query using this internal chatbot to say, what has been the sales strategy in the Midwest region related to life sciences. And rather than read the whole document, it will synthesize that information and present it to you real time. And this becomes very powerful, Jaden, when you're talking about it being pointed not to a single document, but a corpus of data, which could be multiple files.

5:40So in fact, I pointed that to my CRO's weekly sales report for the last nine weeks. And I said, give me the trend analysis of the churn metrics by pointing to these documents. So instead of in the past, you would take this unstructured content, make it structured, push that into a big table or some kind of underlying database, and then throw a reporting engine like a Tableau or something on top. You don't have to do that anymore. You basically can use generative AI, the kind of capabilities that we are building, and being able to do that with all the unstructured content. And so the internal chatbot is one.

6:16Document summarization is another one, which is it gives you the critical information in files without having to read the whole thing. Audio transcription to extract messages out of audio video files. These are some of the things we have already announced and we are making it increasingly available to our 25 ,000 customers. Wow, that's incredible. What has been, that's like you mentioned, this is incredibly powerful technology. It makes a big difference being able to, you know, kind of sit there and query documents and other content inside of a database, give really rich responses. And I've seen a number of different companies doing this in different ways, right?

6:53I believe Shopify came out with one that, like, allows people in their shop to kind of ask it questions around, you know, what's going on in the data within their shop. For Ignite, what has been the customer response since launching these AI tools? You know, that's a very interesting question because you would think that with all the hype and all the noise around generative AI in particular, that enterprise users would be clamoring and jumping up and down to say, I need the capabilities. On the contrary, in the enterprise context, there is a real fear about security issues related to, hey, are you using chat GPT and therefore training the model for public use?

7:36And is therefore my data security being compromised? And so we have to have a lot of explanation to say no. Like we are currently LLM agnostic. So we are some of these with GPT 3.5 on Azure with a state model. So nothing gets persisted in Azure. And we have all kinds of TLCs to say, look, your data has not been stored. It's encrypted. It's a stateless model and whatnot. But the fear that your information might be going into public domain or because we have so many customers, it's a multi-tenanted architecture, that there's no co-mingling of data to give you more accurate response. Those are real concerns.

8:18I would say most customers are excited about the capabilities, but they are a little hesitant to have it turned on on the Git code. They want a lot more validation and a lot more review before they say, now give it to all my users. And we are currently going through that process, Jay. Okay. Very interesting. Can you talk to us a little bit about what a typical business or enterprise's use case is when they're coming to Ignite? What's your typical customer doing? what problems are you typically solving for them? Sure. Even though as we have evolved the product and the platform footprint has gotten bigger, it's become a veritable Swiss army knife with all kinds of capabilities.

8:56But at the most base level, because you and I live in the world of tech and we drink our own Kool-Aid, we think everyone is tech savvy, everyone's working. Surprisingly, the most basic use case of migrating to the cloud, in our case, removing file servers and direct network attached storage and just moving your data into a cloud-based substrate and then being able to access that using your mobile desktop or whatever web interfaces, that primary use case which appears so primal is still a bulk of our new customer evolution or growth story. The most basic. The more advanced people are looking into, okay, now that I moved to the cloud, how do I make the cloud or this repository to be the underpinning of all my data while it's being accessed with my line of business apps like a Salesforce or Marketo?

9:46Or if I'm in a particular industry like construction engineering with my Bluebeam, my PlanGrid, my Life Sciences, my BenchLink. So we think that everyone's moved to the cloud. Only 20 % of the customers or the potential market in North America has already adopted the cloud. the rest is still a greenfield territory so that's the most primal use case but companies that we have which have been with us for a long time which made the evolution to the cloud a decade back they are more focused on elements of cyber security and data governance along with the slew of increasing compliances ccpa gdpr and all these different states are coming out with their own privacy regulations they are more concerned about the risk mitigation the ransomware protection compliances like CMMC.

10:33So depending on your evolution and move to the cloud, the level of sophistication, what you need from a product is also different. And we use those continuums. Okay. Yes. That makes a lot of sense. I can see a lot of value there. Talk to me a little bit about, so on this podcast, we have a lot of listeners that are AI entrepreneurs or starting out in the AI space, creating tools. Talk to me a little bit about the beginning days of Ignite. What kind of inspired you to start this company? What kind of got you motivated? And what did that step look like for you? Jaden, I'm a product guy. I'm an ex-engineer.

11:12In fact, I started my career writing C code on an operating system that most people probably haven't heard on your podcast, VaxVMS. And then I was writing on Unix. So I go deep into technology. So the reason I bring that up is not to say, look at me, or date myself, it's the fundamental, if there's one secret sauce or superpower of Ignite, it's fundamentally a product-centric culture. And four of us who started the company, I and the three others, we are all ex-product people or engineers. And that shows in the amount of effort we put in on the product side. So today, just to be on record, we are 16 years old.

11:53Even today at a late stage, and we are private yet, 30 % of our OPEX goes into R &D. Wow. Which in a company at this stage is 70 to 21%. So that shows you that we keep investing on the product side. And going back to the crux of your question, we've always believed in building a fantastic product that the end users love to use. And if you build a great product, you know, more upsell will happen and therefore your dollar-based retention will go up. Churn will be single digits, which is all true in our case. but the most important thing is end users love using your product that to me is the most satisfying thing and we have striven and we continue to strive to say if i bump into a customer if i go into a cio's office to talk about ignite the first thing i want to hear is we love ignite it makes a difference it delivers the value i'm not being altruistic man look we are as capitalist as anyone and we have to create value for everybody, including the effort and stockholders.

12:58But I fundamentally believe in my core heart, don't chase valuation. Create value for your own customers. The value will be created in your company's valuation, but value versus valuation. So we have never gone around saying, look at us, we are a unicorn. In fact, I don't like using the term. I say, if you want to call me something, I'm a stallion, not a unicorn. I have written to my own drumbeat just to make a point here. And this is on record. This is audited numbers. ENY is our auditors. For the last seven years, we have grown a respectable 25 to 30 % year over year. Wow. That's impressive.

13:37It's not a 70, 90 % growth with insanely high losses or very high losses. Five out of seven years, we've been cashflow positive. We are today at scale. So, we are well north of$200 million. We have 1 ,000 plus employees. But the thing which is, I take a lot of pride and therefore my bragging rights. We have been EBITDA positive for the past several quarters with improving EBITDA margins. We are adding gas to the balance sheet and not eating off it, which is becoming more and more coming back into fashion. But four or five years back, people would challenge me to say, man, you're not growing 70 % or 90%.

14:14Growth is what you should focus on. And my answer used to be a polite F that I want to build the way I want to build, which is a total antithesis to the Silicon Valley approach, which is growth at any cost model. I never, ever pursued that. Yeah. I think that's, that's really, uh, that's amazing that you were ahead of the curve on that. I mean, I guess this comes from, you know, 25 years experience in Silicon Valley. This isn't your first rodeo or first time around, but yeah, you, it definitely was, you know, 2020, 2021, everything was so frothy in the markets. People were, you know, giving away money like crazy.

14:50And you saw people raise crazy amounts of money. And I feel like a lot of really unsustainable business practice. So that's to your own credit, that's amazing that you had the foresight to build in such a sustainable way that, you know, would make this a long-term play and really sustainable company. And I think that is really good advice for anyone building in AI, anyone entrepreneurs in tech right now. Talk to me a little bit about when you wanted to put together Ignite, what that core team looked like, how you found your core team members when you started this, and I guess the importance of having a really solid team for a company.

15:28This is one area where even today I feel very blessed. Obviously, there were four of us to start with, and I knew all of them from my prior life. In fact, my previous startup before this, which you referred to in my introduction, Valdero, the two out of the three were co-founders there. I was always the main driving force that would put people together. Besides these four, when we started hiring people, which traditionally is engineering and products, less in marketing, because you're trying to build the product first. I am very happy to tell you, I still have some of the core team members with me.

16:04They've been with me for the last 10, 11 years. Wow. The value creation for them individually, their own growth, their roles and responsibilities have grown. And I find it so gratifying that, you know, we've created a culture and an environment where people, of course, everyone has to make more money as you go through life, expenses go up, family, kids, college, whatever. So you have to support that for sure. But the level of excitement, the joy of when you were a 10 % company, now that you're a plus thousand person company and people feel that there's a personal growth, that's what keeps people motivated.

16:38So my unsolicited advice to anyone would be is find people who are having a certain core set of values that align with yours. and one of the things I keep repeating is there's too much of a focus on valuation versus value I find it silly that people go on to LinkedIn announce their funding which they try to do every 18 months funding is not your bragging right it should be more about business growth I do understand you have to create credibility to say to attract talent and therefore you have to show my hundred million dollar round which are even a million dollar round but I feel that people congratulate themselves to say, I raised a hundred million.

17:17I'm like, dude, that's somebody else's money. And you put it away. Why are you congratulating yourself? It's like, what did you do with that money is what you can brag about. Yeah. What are the results of the actual company? Right. And you know, this is tech in general, but things are changing to some extent. People are associating more with value creation than just insane growth. And I think in the current climate where valuations are getting crammed for subsequent rounds for private companies, public company valuations are coming back to some reasonable multiples, especially in the tech sector.

17:51I feel that some common sense is coming back. That's at least my opinion. I sound quite a contrarian in the past, but not so more anymore. I feel like people are saying, okay, this is the thing now. Yeah. Yeah. 100%. And I guess on that vein, I would love to pick your brain on where you see this going because you kind of alluded before to the fact that AI is kind of like, it's kind of the trend of the day 100%. How do you think AI and like some people call it the AI bubble, but let's call it like the AI boom or the AI revolution or whatever you want to call it right now. How would you call this current kind of AI cycle that we're seeing?

18:27How do you view this compared to perhaps previous ones, right? Like with the previous crypto cycle or the Web3 decentralized, crazy valuations, crazy amounts of money. Where do you see those two on, I guess, like on a long-term play in tech? How do you see the integrations and the sustainability of those and comparing those two kind of bubbles or movements? It's interesting. We have that discussion quite a bit internally. And I've been here in the Valley for 30 years. So I've seen 2001, 2008, I've seen IoT bubble, or well, IoT hype, then Net3o decentralized, DAOs, crypto, and of course now AI.

19:05So it's interesting that in tech, we always have something new and then it has to be the thing that's going to solve world hunger. Right, right. But this one in particular will have a long-term long tail effect in terms of the value creation because I do think that if applied correctly, these existing AI techniques and the evolving techniques, I don't believe we'll ever get to AGI. I don't see that happening. Forseeable future where AI can be almost like a human, I don't believe in that. We're a long ways to go unless we get into computing or something. But I do believe that the amount of productivity impact, the amount of knowledge synthesis that can be done using the current and evolving AI techniques, they are real.

19:55So I don't think that this is a hype cycle that will abate. And the valuations are beating like the VCs were clamoring over each other to get into some AI investment. A lot of AI companies were nothing but wrappers on chat GPT in my opinion. Seriously. But I think this will persist for years to come. But yeah, there'll be less hype around it. Where do you see the biggest opportunity in AI? I mean, you're currently integrating AI tools in your platform. You're seeing the ways the enterprises are using them. You mentioned that a lot of startups were wrappers on OpenAI's API call. Where do you see the most sustainable or the most opportunity in AI?

20:37Is it the AI models? Is it the infrastructural technology supporting them? Where do you kind of see that? My belief is that, look, the LLMs that you're getting from Google or from Facebook or where the big guys are you know besides chat chat gpt or sorry uh open ai i believe that the the generic large llms will only go so far and there's a lot of value in them don't i'm not underestimating their power but where i believe the more exciting thing's gonna be is going to be in vertical specific uh ai uh like we are working on and even though we haven't come out with cool buzzy marketing terms. So it sounds hokey, but we're calling it Ignite AI for life sciences, Ignite AI for AEC architecture, engine construction.

21:30We are building our own language models there. And one of the biggest challenges in building a model is to train the models with the amount of data you have access to. The biggest problem is not building a billion attribute or a 500 million attribute model. I mean, that itself is a non-trivial problem. Don't get me wrong. It's training the model and the data that you need access to. Now, fortunately, in our case, because of the nature of our business where people's content resides with us, we have petabytes and petabytes of data. So we can leverage that to refine the models. I think the real power of AI will be felt in industry-specific AI models that people will build, including companies like Ignite, and then extract the value out of these models to deliver the document synthesis to actionable activities that you can automate using elements of AI.

22:25But I think it'll be tied to industry-specific models, in my opinion. That is at least my core beliefs right now it leaves for change yeah and i think that you 100 are going in the right direction there and that's impressive that you guys have the first site over at ignite and you're building those out i i remember um it kind of reminds me of a conversation as stanford trained pubmed gpt and essentially they grabbed all the publicly available medical journals and they made a gpt that was good at answering medical questions they said this is great to answer medical questions but if you wanted this to be even better you would train like a heart disease gpt or like a breast cancer gpt on very specific areas um and so yeah i think you know right now like you said we got the googles the the meta we have open ai and stuff with kind of these big large general ones but i see the space as well like you where this is going to be fragmented to thousands of really specialized specific models um and so yeah i think that's really impressive that over ignite you guys are are kind of building that out and on the forefront of building you know you have the data and and that's really where it comes down to who has that data and is able to to build powerful things from there.

23:28I do have a question because you did mention kind of GDPR and compliance and stuff. With regulations like those kind of coming into play, how has Ignite adapted its AI solutions to help businesses stay compliant with all of that? See, right now, from a standard industry compliance is like for instance you know when you're looking at uh generic uh pii or sorry uh you know identifying uh information like pii or other critical identifiers generic ai or nlp can help you do that even things like you know if you had your social security number in a picture using all techniques we can extract that and say hey this is your social security number and depending on what kind of heuristics you've defined, it's treated as a high priority or high risk item.

24:25Then you look at who's got access to it to say, hey, only HR should have access to this, why somebody in engineering is able to access. So those things to me are relatively mundane. Okay. But I think that, and a lot of these compliances, you mentioned GDPR, I mentioned CCPA, there's CFR Part 11 in life sciences. there's all other things in different industries. Those things can be delivered using NLP and data classification quite easily. I think couple that with the fact that there's still a growing sensitivity about how much information synthesis you can do and therefore extract things that can be presented to the end user.

25:11Does it go in front of the right eyes? Does your scanning not go outside the boundaries of what you're allowed to see in the context of an enterprise? Those issues to me are more relevant and more concerning for enterprise IT, enterprise CISOs. I'm not sure right now if I'm seeing that compliances and regulations and AI are sort of, let's say what do you call them, they are at cross purposes. Now, of course, as you see AI becoming more pervasive, I can bet you even the data sovereignty issues about EU data versus U.S. data versus data in Canada and Canada provinces, those issues will start coming in context of AI.

26:02But so far, we haven't run into that that often. I think we're in all the innings, to be honest. Okay. Okay, yeah, that's really interesting. Talking about a little bit of your journey, I know you guys have done a bunch of really interesting things. I was looking over at your website and I noticed that you guys kind of mentioned invisible apps. I was wondering if you could talk a little bit about the concept of invisible apps and how they enhance users' experience, what kind of feedback you've received from users about those. So, you know, it's interesting. I'm glad you brought up that topic. so in my opinion as you call invisible apps they enhance and optimize the user experience quite seamlessly if they're done right so to me it's having software that works non-inclusively alongside the user for example and i may have mentioned that before how can you automatically prioritize incoming content based on learned risky signals.

27:07Why is that important? Because that can allow for faster identification of anomalous content and user behavior. So we have introduced several of these AI-powered invisible apps without requiring user intervention. One such app is the content recommendation generated within and around the search queries. The other is what we call is the smart cache app that i hate to use buzzy terms but that does employ deep learning techniques that will preload the large files locally on a hypervisor device based on anticipated user needs so when people are in the office they'll be able to access the data at land speeds on a local cache without they needing to pin that data explicitly so you have to factor in things like users roles past activities, current place within the application, to recommend what I would call as knowledge-based entries and prioritize recent activity within the search results.

28:04These are some of the examples of invisible apps that we have already provided capabilities around within the product. I don't know if I answered your question. Yeah, yeah, for sure. Yes, that was a great explanation. Vinny, thank you so much for coming on the podcast today. I know we have to wrap this up, but if people want to get in contact with you, or if they want to learn more about Ignite, where can they find you and where can they find more about Ignite? So even though I might be inviting a lot of spam, I'll give you my email address. It's vjain at e-g-n-y-t-e dot com. And oddly enough, on Twitter, which I still use, by the way, despite signing up on threads and it was just signed off and nobody really is on there, on X now, my Twitter handle is such an oxymoronic, it's cloud not enough.

28:50So you can DM me on that or you can email me at vgen at ignite.com. Okay, I'll add you on Twitter and yeah, ignite.com. People can go visit there to find out more about what you guys are building. Yes, yes. I'll leave a link to that in the show notes. Thank you so much for coming on the show today and for the listeners, thanks so much for tuning into the AI Chat Podcast. Make sure to rate us wherever you listen to your podcasts and have a wonderful rest of your day.