Episode 31: AI at Box with Chief Product Officer Jeetu Patel

Tune in for Jeetu's take on how they are thinking about AI at Box, updates at Box regarding AI and machine learning, advice for executives on evaluating AI as a strategic advantage or differentiator, how to tell the difference between the snake oil, and what really works, whether or not there is time to wait for the market to mature before you jump in on AI, and much more. 

Subscribe to AI at Work on iTunesGoogle Play, or Spotify

Rob Circle Headshot

Rob May, CEO and Co-Founder,
Talla 

jeetu patel

Jeetu Patel, Chief Product Officer, 
Box


Episode Transcription   

Rob May: Hello, everybody, and welcome to the latest edition of AI at Work. I am Rob May, the co-founder and CEO of Talla. Today, I am very excited to have the Chief Product officer of Box, Jeetu Patel, to join us to talk a little bit about what they are seeing with respect to AI around documents and in their part of the market. Jeetu, welcome. And let's get started by telling us a little bit about your background, and how you got to Box, and what you do now that you're at Box.

Jeetu Patel: Thank you for having me, Rob. I'm really excited to be here. I joined Box about 3 and 1/2 years ago, maybe 4. Time flies. I joined here from a company called Syncplicty, where I was the CEO, which was a competitor to Box, oddly enough, and was owned by EMC. When I joined Box, we had just gone public. We were a couple hundred million in revenues.

If you think about what our challenge was at that point in time, it was that we were a single product company. We had achieved pretty good product market fit for a single product company. We then had to scale our business, to being a multi-product company. And, Aaron Levie, our CEO, and I, we're friends, and we were having dinner one night. And he's like, “why don't you come over here and help us build out a platform business”, which was going to be our second act where we could be the platform that could help take this to a multi-product kind of offering.

I joined the business after a few conversations with him. It seemed like a great track. Then about a year and a couple of months ago, initially we had incubated the platform business, and one of the things that's pretty important as you start to incubate new businesses is you want to make sure that there is a level of insulation that that business has as they start to incubate a new idea when there's already an existing business. So, we kept that as a separate business unit.

Then as platform got to scale, and we started having it grow, and there were great customers on it, it made sense to kind of fuse it all together into one. I took over all the product and strategy a little over a year ago. Now we have about 20% of our revenues are spent on R&D. We spend hundreds of millions on R&D, and it's largely to make sure that we can have a very, very potent kind of business around helping organizations fundamentally change the way that they work with content.

So, how do you collaborate with content, manage content, share content with the people inside and outside the organization, empower business workflows, and become much more of an intelligent enterprise over time and go through your digital track. We can help you go through the digital transformation. That's what we're trying to do. We are now about a little over 90,000 customers, close to 62-63 million users, and have about 69% of the Fortune 500 that we enjoy calling our customers.

RM: That's awesome. I'm really interested to know, with respect to AI, Box is really interesting to me, because a lot of the companies-- I've kind of thought of the world very much in sort of two pieces. There's AI-native companies, which are sort of a lot of the companies that started circa 2015 or 2016 and later that AI was the thing that they were built around. You have a bunch of older companies that are big, they have a lot of data. They're sort of older tech companies, and they've had machine learning and AI divisions that are doing things for their products.

Box falls into this really interesting category. You guys came up in the sort of cloud collaboration area and were one of those leaders there. From a public company perspective, you're sort of mid-tier in terms of size and particularly age. I'm really interested in a company like this, like, when you had dinner with Aaron and you guys started talking about this, was AI even on the roadmap then? How did AI come into Box, and when did you guys realize, hey, this is not only something we need to look at, but something that could be really important for what we do?

JP: The way that we think about this is the way that, typically, companies can get to hypergrowth and scale is when you see a major megatrend in a market, you've got to make sure that you're riding that wave and using it as a tailwind. You have to do it in a way that's not just building technology for technology's sake. You have to have a deep understanding of how that market shift is going to help your customers solve some very meaningful problems.

We've been lucky enough where when we first started the company there was this move to the cloud, and that was a big megatrend. It completely transformed an opportunity for us where we said, oh, wow, this is going to give us the propelling boost that you need. It was a shift that was happening, and people were going to move to the cloud one way or the other. They wanted to move their most sensitive intellectual property and content to the cloud, and we happened to be at the cusp of that intersection.

The second wave of a megatrend that came about shortly thereafter that really made our business take-off was the mobile wave with smartphones, and iPads, and so on and so forth. You started seeing that people wanted to access information from anywhere at any time, so on and so forth. And now what you're starting to see is this third wave with artificial intelligence and machine learning innovations that are going on. But you have to be pretty careful of what do you apply that to, and is there a real problem that can be solved?

And content, it's a pretty interesting market in the space that we're in, because one of the challenges that you have with content management and collaboration in general is that the more you use these systems that have content, the harder they get to use.

RM: Right.

JP: That is because the more content that's in there, it's harder to find. It becomes pretty hard to navigate through it and so on and so forth. That just seems counterintuitive. It's the wrong thing to do. You should have a system that over time the more you use the easier it gets to use and the more value that starts to get compounded.

So we've been thinking about this problem for a pretty long time. And one of the ways that you think about solving that problem is, say, well, once you have a lot of content in your system, we're pretty fortunate where our customers double the amount of content that they use us as custodians for every year in our repository. So our data doubles every year within Box, roughly.

RM: Wow.

JP: So, when you start thinking of value, you say, your data is growing at a pretty fast pace. As the data is growing at that fast pace, what do you need to do to add more intelligence to the content so it gets easier to use and you can actually be able to get to the content and have the right level of recommendations and insights come out of the content as fast as possible? There's this whole notion of, how do you put more intelligence around the content, where we talk specifically about adding metadata to the content that's already there. If you have a file, or an image, or a video file, or a contract, or a document, how do you put some metadata around that you have some ways to describe the content and get more intelligence around that content?

Now, metadata is not a new concept. It's been around for a while. However, the challenge with metadata has always been people have always assumed that metadata gets entered in manually. So if I have a contract and I want to enter in the account name of the account for which the contract has been designed, and the amount for the contract, and the contract expiration date, those are all things that I manually enter in. Well, that doesn't scale when you've got billions and billions of files in your repository.

What we saw was a perfect opportunity to say, well, artificial intelligence and all the innovations that are happening in machine learning, with the new ML models that are being built by some of the major players who are spending billions and billions of dollars and perfecting these models, why don't we bring that intelligence to content in Box, because an AI algorithm or a model is only as good as the data set that it's applied to. And we happen to have a lot of data that our customers trust us with, and they want to make sure that they can extract more intelligence out of this data. And we want to make sure that we provide them mechanisms to do that with an ecosystem, while keeping the data secure, and protected, and managed within our repository.

What do we do to make that happen? We created this product called Skills that literally allowed you to automatically add metadata and extract insight from content. If you happen to have images, you can automatically tag the objects in an image so that if it's a digital asset management use case you might be able to say, well, show me all my images that have a tennis ball, a tennis racket, and a female playing tennis with a green background, because that's what I need for my marketing campaign as an image. An of my 10,000 images I might have, it'll show me the 30 images that have those kind of objects in the image.

That was an easy way to automatically tag that metadata, rather than having to have people manually enter it. You could do the same thing with contracts, or with documents, or with video files, or with audio files, or what have you. And there's some pretty neat things that you can start doing once you have these machine learning algorithms that could be applied to it.

That's where we saw a very natural fit with machine learning tied to something that we already do to make sure that you can extract more meaning and compound the value. The way we think about this is any piece of content that's in Box should be infinitely more valuable than that same piece of content if we were outside of Box. And the way that you create and compound more value is by adding intelligence to the content, and then having it embedded into a business process. And the way you add intelligence to the content is by applying a machine learning skill to the contents so that you can automatically process that content at scale.

Now, if you only had five documents to do this with, it wouldn't be that interesting. But when you've got billions of documents, people just don't have time to go out and tag them individually. So this is where, through a machine and going out and conducting a task at scale, that really goes out and delivers a lot more value from content. And as a result, people want to use Box more, they want to protect their content more, get it part of an embedded workflow in a more effective way, and so on and so forth.

Does that makes sense?

RM: Yeah. That makes a lot of sense. We were talking before the show started about, definitely, Box Skills is one of the newer initiatives that you guys have had. But are other things going on at Box with respect to AI or machine learning that you can talk about?

JP: We have this product called Box Skills that we've launched. And what Box Skills essentially does is takes all the innovation that's happening with machine learning models from the world and applies it to your content. Then you can go out and have very interesting use cases that get served, like I want to do better discovery with content, I want to go out and trigger a workflow, all of that metadata can then allow you to do all of those things.

What we also found was we have this wonderful technology that we're building over time with machine learning called Box Graph. Box Graph is the ability for the system understanding the inner relationship between a piece of content and a person, two pieces of content, and people and people. And so what can we do with that information to actually surface more meaningful insights for our customers?

For example, could we use machine learning, not just for going out and automating workflows, and not just for making sure that there's better collaboration or better discovery, but could you also do better policy enforcement of security with content. We've announced a product called Box Shield. What that does is allows you to go detect anomalous behavior and detect anomalies within the systems. So if, for example, Jeetu, his last day at work is on Friday and on Thursday he starts downloading two terabytes of data, that seems like anomalous, malicious behavior. Let's flag that and make sure that that gets reported out to the administrator that this is not normal behavior.

If salesperson starts downloading 500 product specs on their system, that's something that they don't typically do. Why are they doing that? That becomes an anomalous behavior. What we're starting to now do is use machine learning for things beyond productivity, beyond process automation, but also in areas of security, compliance, governance, which then completely opens up the aperture and providing value to customers in many, many meaningful ways where every single piece of activity that's happening in the system is compounding and adding more insight to what you might be seeing as normal behavior within your system.

RM: Interesting. So, you're right you're a product guy, and you're in the middle of Silicon Valley at one of the key Silicon Valley anchor companies talking about AI. And so do you have a framework that you use of the way that you think about for, let's say I'm an executive at an insurance company, I'm at an e-commerce company, or whatever. I'm not that technical, and I'm trying to figure out, how do I evaluate AI as a strategic advantage or differentiator in my business? Do you have a framework for thinking about what that person should say, oh, Box should do that for AI. I should do this myself for AI.

Do you have advice for people about when they should outsource AI to the tools they use and the different providers for things and when they should do it in-house?

JP: We have a pretty strong view on that, which is what you have to do as a business is-- the half-life of business models, and products, and businesses in general and the velocity with which they move, the half-life is shrinking down quite dramatically of business models. Every seven to eight years you see that a business model completely resets within a company. As a result, every seven to eight years, you as a company have a choice of disrupting yourself and moving to the next wave or being disrupted and being an irrelevant company.

And by the way, the half-life of a product is actually going down even faster, where within 18 months a product gets dated. So you literally can't take three years to build a product and ship it, because by the time you shipped it would be dated. So then you ask yourself, well, if your half-life is shrinking, what do we as a company need to do to make sure that we can succeed at the velocity that the market is moving at. And the way that you do that is by being very disciplined about what you pursue, versus what you pursue to partner with.

The framework that we typically use is anything that is meaningfully core to your business, where you would differentiate your offering based on that capability, you would want to make sure that you're building yourself and you maintain that IP and that process. But anything that is an essential component but not core to your differentiation you might not necessarily want to focus on building yourself, but go to someone whose core business it is to do that thing and partner with them. And we do that a lot on our side with things that are not core to us that we would partner with.

We actually encourage our customers to do it, because if there are certain things-- I'll give you an example. If content management is not core to someone's business, but it is core to Box's business, you should not, as a financial services company, think about building a content management system. You should go and make sure that you use someone like Box for that piece of it and then focus on doing what you do best, which is serve your customer, issue a loan, have an insurance policy, whatever it might be. And that's your core, and you should focus on your core and partner for the rest.

We think that that formula works in every step of the value chain. We do the same thing. Our suppliers do the same thing. Their supplied do the same thing. What that does is creates an air of specialization and openness of an ecosystem where the ecosystem just interoperates with one another.

That actually is just a different way to operate in this new economy, compared to the way that it used to be, which is you have a vendor or a company deciding that they want to go out and build out the entire stack from top to bottom. That's just not how the world is going to work in the future. So, you have to be extremely adept at partnering and knowing what's core versus what's not core.

RM: On top of that, when you think about people who are thinking about adopting AI now, and they're looking for their vendors to start doing some of this, how does a new person, coming into the space, educate themselves in a market where we're still trying to sort out-- maybe less so with Box, but particularly with a lot of startups-- what's the snake oil, and what's the real stuff? And what really works and all that kind of stuff?

Kind of a two-part question here-- how do you get smart about AI if you're not an AI person? Are there frameworks you can use or resources you can look to? And then number two, can you wait for the market to mature, or is this technology moving too fast that you have to jump in and find ways to embrace it no matter what business you're in?

JP: Both great questions. And let me answer the first question by before you get smart about a technology area, I think what's really important is to get extremely smart about the problems that you want to try to spend your time solving. And I think this is an area where especially in companies-- the reality is, every company eventually is going to have some kind of a technology differentiator, and they're going to be a tech company. They're going to be a digital company of some sort.

Even if you look at a financial services company, or you look at a retail company, or you look at an insurance company, you look at the health care company, you look at a life sciences company, they're just operating in a way that's much more digital today than the way that they used to be. The business processes need to get more digital. The way in which you operate with your people needs to get more digital.

The thing that's most important is don't start from the technology and work out upwards to find problems, start from really important problems that are going to be central to your defining yourself as a unique provider of value, and then work backwards from that problem. And once you've identified what the problem is, one of the things we talk about over here is we're very obsessed about building painkillers not vitamins. It's nice to have nice-to-have technology, but it's really important that you actually build some technology that's critical that alters and makes a difference. Do something different so that you can make a difference.

And so in that vein, what we tend to do is identify what are the different kind of areas where problems need to get solved within your domain. You're going to be the best judge of what those problems are that are going to be the most potent problems. Go validate those problems of whether or not they are in fact problems for which people are willing to part with their money regardless of the industry that you're in. And then work backwards and say, if I want to do this in a very effective way at scale, efficiently, and do things in a very discontinuous manner, where the solution I build is going to be 10x better than the solution that currently exists, that's the only time that people are going to move over to you. They don't move over when you have a 20% better solution.

AI can be a pretty big contributor to that in saying, what can AI then do to make sure that it can provide you a 10x differential rather than a 20% differential. And so that's the way that at least we think about it over here. We try not to get into things that other people could have replicated just as easily. We try to get into things where we have unique value and a perspective where if we didn't build it, the world would look different compared to if we built it. And then in those areas where we choose to build it, we want to make sure that we are very, very focused and kind of laser-focused on using technology that's actually going to provide us a tailwind to propel forward in a much more accelerated fashion.

RM: That's a great answer. Are there things, products, or technologies that Box isn't working on with respect to AI that you wish somebody would work on or solve that would make your life a lot easier for the things that you're doing but maybe aren't core to your business?

JP: I would say if you think about-- one of the things that's fascinating about our business is our infrastructure scale is 10x every 3 to 3 and 1/2 years. That's a pretty kind of core property of any of the cloud vendors is, how does your infrastructure scale? So if you think about an area where that has been massively innovated on with cloud providers, like the Amazons, or the Microsofts, the Googles of the world as providing those kind of infrastructures that allow you to do these things at scale so that there is much, much more innovation that can be there, I think in the area of AI what we are still in the early days, as far as the kind of machine learning models that are being built for going out and extracting different kinds of intelligence, we are not going to build every single intelligent machine learning models for identifying all the intelligence out of every class of document, or file, or unstructured piece of content.

We'll have people that'll get specialized in really understanding a contract, or really understanding a lease, or deeply understanding how a video can be tagged more effectively. And more and more of those kind of innovations and machine learning models, as they get built, we will continue to keep benefiting from that. And I there is billions of dollars being spent on that, but that's an area where we actually find it to be super useful.

The second area is international language processing. As that gets better and better, you will actually start to see that the way in which people will engage with systems that have information might be very different. You're starting to see a lot of innovation happen around voice, which has been pretty exciting to see with what Amazon's doing with Alexa. I think that voice innovation is going to move over to different systems.

None of those are going to be our core where we build those capabilities. But as those capabilities evolve, we can just sit on top of them and make sure that it makes our user and our customers' lives better. And so if you start thinking about all the innovation that's happening-- the way I'll leave you with this example is, imagine if Amazon was built-- and this was a great example Jeff Bezos gave-- before the railroad was built, and before the airplane system was built, and before the US Postal Service was built. It would have not been a viable business model if you didn't have the core underlying infrastructure.

I think in every phase of innovation you need the underlying infrastructure so that you can build the next phase. And right now, the underlying infrastructure around AI is just getting started. There's a lot of maturity in certain areas. There's a lot of areas that still needs to mature. And so in all of those places, we want to make sure that as quickly as we can apply that to our content, the better off we'll be.

RM: I think that's a really great insight. Last question before we wrap up the podcast. There's a lot of talk-- it's interesting, because we talk mostly on this podcast so much about AI and how it applies at work and what it enables people to do. But so much of the news cycle around AI is driven by killer robots and should we be worried about AI-- whatever. As somebody who kind of works in the field, are you concerned about this at all?

Do you think we should be afraid of creating a general, terrible AI? Or is that all overblown? What's your perspective?

JP: I think this is probably one of the most important questions as you ponder the impact of not just AI but impact of technology on the markets. And so I think there's a couple of things here. One is, Silicon Valley in general, and in general the tech space, has had an overly simplistic narrative, which is technology is generally good, and disruption is generally good, and you need to just make sure that you move forward with assuming that there is going to be net positive with disruption, therefore, disruption is always good.

I think there is a deeper responsibility that technology companies are going to need to have in the next era compared to the way that they've thought about this in the past era. And it's going to be based on a few things. One is, people's privacy is going to be of extremely high importance. Every technology company is not only going to have to think about making sure that they prioritize privacy, but in addition to prioritizing privacy, they need to make sure that they are keeping in mind the ill effects that technology has while they're building the technology, just as much as they're thinking about the positive effects.

In the past, what's happened is people build technology with only the optimistic view in mind. And when something which the downside occurs, they address it at that point in time. I think what we're going to need to do with AI is actually start to foresee what the societal implications might be and to make sure that you're addressing that ahead of time.

Then the second thing is, in general, when disruption occurs and you actually disrupt the steel mill industry or a coal mining industry, and the coal miner doesn't have a job and they're 50 years old, their next job doesn't become that of an AI developer. So I think we as a society have a responsibility of re-training the workforce in a more effective way, so that when you do disrupt a market and an industry, we have to make sure that as a community we are retraining those people so that they don't actually have-- imagine someone who's actually been in a job for 30 years, and now that industry goes away and they haven't really re-trained their skills, and now they're in the last decade of working life. It becomes very difficult for them to go out and adjust to society.

That's when people had rock bottom, and there's kind of inequality, and there are people that get too much, and there's people that have too little. I think there is a deep responsibility in how we build technology that looks at both sides, not just the good but also the bad, and ponders that deeply. And also make sure that they are responsible when there's disruption occurring, that there's some way that they're giving back to society with retraining of the workforce and with making sure that they give people opportunity from different parts of the world.

RM: I think that's a great answer. And that's a good answer to end the podcast on. So Jeetu Patel from Box, thanks for joining us today. And then those of you who are listening, If you have questions that you would like us to ask, topics we should talk about, guest we should have on the show, please send those to podcast@talla.com, and we will see everybody next week.

JP: Rob, thank you for having me.

Subscribe to AI at Work on iTunes, Google Play or Spotify and share with your network! If you have feedback or questions, we'd love to hear from you at podcast@talla.com or tweet at us @talllainc.