Perfecting your digital assistant

Aaron Levie | Co-founder and CEO, Box

By Casey Newton | Nov. 17, 2016

Aaron Levie spends more time thinking about the future of work than most. When the 30-year-old entrepreneur co-founded Box, it was as a place for average people to store their files online — Dropbox before there was Dropbox. But years ahead of Dropbox CEO Drew Houston, Levie saw that the real money in online storage was in the workplace — and so the company turned to making products that let people collaborate at work.

The result is a company that successfully went public and beat Wall Street’s expectations in its last earnings report, though it still has yet to turn a profit. (Levie says it will be cash-flow positive by the end of the year.)

When Levie is away from Box, he’s surveying the tech landscape for other innovations. He has been a savvy investor, making early bets on companies including Stripe, Zenefits, Instacart, Authy, Robinhood, and Color Genomics. His work has given him an optimism about the way that new productivity tools will make working life better for the average cubicle dweller — making organizations more transparent, and good ideas easier to spread.

Levie is also an evangelist for artificial intelligence, which is slowly but surely working its way into the office. The future of work is in many ways a story about the future of AI — about what it can predict, what it can prevent, and what it can do on your behalf.

The Interview

Your company exists, in part, because mobile devices enable us to demand consumer-quality experiences in our work software. Five years from now-ish, how close will we be to parity between work and consumer software? Will we even recognize a distinction ultimately between these two things?

I think you sort of nailed it, which is mobile and cloud sort of conspire to redefine what enterprise software looks like and how it runs and how it operates. I would say we are already on the cusp of that happening, where our enterprise technology should resemble and be as simple as our consumer technology. What we actually have is an issue where there’s just a disparity of the tools that the most early adopting types of companies are using and the tools that the rest of the world uses.

If you were to look at maybe how we work, we’re probably already using mostly technology that feels like, and is indistinguishable from, consumer software. But if you go to large corporations all around the world, whether it’s DOW Chemical, or Ford, or these big businesses around the planet, there’s a separation of the haves and have-nots of modern technology. I think five years from now, we will have converged. Companies won’t be implementing old technology anymore.

Subscribe to our newsletter so you don't miss the rest of the Verge 2021 interviews. Terms apply.

How will the tools themselves changes?

I think what’s going to happen is enterprise software will become more and more deceptively simple.  It’ll feel like we’re using just a consumer application. I think we would agree that a product like Slack, or what Facebook is doing with Facebook at Work — these just resemble consumer applications.

And how does that change organizations over time?

If you study modern management science of just how organizations have been built up over the past 100 years, you have a very clear hierarchy. The hierarchy has been built up to empower the flow of information up and down the organization. We had to structure our organizations in a very specific way because we didn’t have technology to help us communicate in real time with anybody, anywhere.

You have a lot of companies that are very hierarchical, where the way you get ahead is you hoard information, you hoard power. The best ideas don’t necessarily free flow to the people making the decision. You get only a small subset of the ideas to work off of. Most organizations, the way they’re designed is not necessarily the right organizational structure with the right people at the top, because it’s the people that have gotten ahead using a very specific set of tools that were built up in a non-digital era.

Now, imagine an organization that can share [information] in real time with anybody, anywhere across the company. You essentially have a virtual space that flattens the organization, where now it’s actually not just about the people that are able to obtain and retain the information, but the people that put forth the best ideas and can execute on those. I think you’ll start to see that work just changes in a pretty material way, because it’s not about how much information you have versus everybody else. Everybody has the same level of information. Now it’s about who can execute on that information and who can continue to put forth better and better ideas.

The speed at which our companies are going to run and operate is going to change pretty dramatically. The kinds of organizations that will win will be those that can adapt to this in the best way, and so far we’ve seen that it’s companies like the Ubers and the Facebooks of the world that are able to operate and work in this way. The legacy companies just cannot move and change fast enough to act on this.


Hearing all this, someone who works at a more traditional, siloed organizations, might think, wow, it sounds like work is going to take over my whole life, and I need to make total institutional awareness of every interaction inside my company.  All of these changes that you’re describing, which do have many advantages in the organization — do they have real advantages for the employee?

There are a lot of people around the world that actually want to do awesome things. They want to do their best work, but they’re not recognized by the organization. Their ideas don’t bubble up to the top, because of the complex and archaic way that we built businesses. I think that some people will lose in this future — the people that got ahead hoarding information and working in a way that was very specific to that model of an organization. You’ll have a new era of people that are able to win on their ideas, and able to win on what maybe we would consider to be actually better reasons, that open up opportunity for a lot of people that previously their voices were not necessarily heard.

I did a couple internships before we started Box. It was just so obvious that due to the hierarchical structure of an organization, if you were 21 or 22 entering that company, it’s going to take you 10 or 15 years for your ideas to be valued by anybody at the top of the organization. That’s just how it was structured. Actually, nobody at the top of the organization wants you to get ahead, because the whole design of the system is that way you can have a clear hierarchy, people can keep their jobs, etc. We actually do want disruption to that structure. We want organizations to be able to get the best ideas so they can change more rapidly, so they can perform in the best way possible. I’m generally optimistic.

For many people who do this kind of knowledge work, email has been the foundation of their day for decades. Will that still be the case five years from now?

Email is to communication as newspapers are to media or news, which is to say that there’s going to be multiple channels over time. Email has been the dominant channel, but we now know there are other experiences, there are other paradigms that we want to be able to have with our communication. We want a real-time experience. That’s where a Slack or a chat tool comes in. You can’t fit real-time in email, and any time you try, it creates just a very complex Frankensteinian product that nobody really wants, like Google Wave or something.


Five years from now, do you think you will send fewer emails than you do today?

I actually send not a large amount of emails, so I don’t think I could send fewer emails. I think that I already am using different tools.

How do you think advances in artificial intelligence and machine learning will reshape work over the next several years?

We think about it in a bunch of ways. Most of the most interesting AI machine learning examples that we’ve seen so far have been on the consumer side. All of this stuff serves to make our daily lives a little bit more convenient. It helps us engage with the people we want to connect with, tells us when there’s a birthday coming up. That’s essentially what it’s doing. Siri helps us interact with information in our device in so many ways.

We’ve never actually seen real deep machine learning or narrow AI applied to business problems. It’s just never really happened at any meaningful scale. There’s been some kind of pattern recognition and anomaly detection, but never really have we applied actual advanced intelligence and machine learning to business problems.

What happens when we take all these business applications and we start to apply real intelligence to them? That, to us, paints a very, very different picture of what could possibly happen in the workplace. There’s a whole range of outcomes, but at the starting point you just think how much time we all collectively waste doing things that literally require no thought. When you can  have the computer do those things for you, now maybe we’ve started to save 20, 30, 40 percent of the time that we have in a given day. Technology is helping us just focus on the things where we fundamentally have to interact with the tool or system, but the technology is doing the rest for us.

All the way to the other end of the spectrum, which are things that people weren’t doing in the first place. Now it’s possible for the first time ever to have computers aid a person in a way that wasn’t previously possible. That’s where the power of AI in healthcare or life sciences or financial services starts becoming incredibly powerful — which is, how do you take the experiences that tens of millions of patients have had on a specific drug and apply that to just one individual patient that a doctor is interacting with? How can we take all of the data that we have and be able to apply it in a highly precise way that no individual could have previously done because they weren’t able to process across all of that information?

It’s going to be everything from saving time to completely changing business or human outcomes because of this technology. Sundar [Pichai] had a pretty good analogy of AI is the new mobile — which is to say software that isn’t mobile today just feels out of date and clunky. Well, software in five years from now where you’re having to do anything manual is going to feel out of date and clunky.

The technology should be telling you what to do: if you’re in sales, it should be telling you who to sell to. If you’re in marketing, it should tell you who to market to. If you are in healthcare, it should give you suggestions of the potential prescription or the diagnosis that you should have. If you are in manufacturing, it should instantly collapse your supply chain so you know where the potential problems are going to be so you can go right after them.

I’m probably more in the optimistic side of the spectrum. There’s a lot of fear that it will disrupt jobs, but I think it’ll probably more have an impact, at least for the foreseeable future, of raising of the level of work that we’re all able to do because we just won’t waste our time doing the stuff that is entirely redundant to what a computer can do.


When it comes to the future of work, does it feel like we’re on the verge of kind of a fresh wave of really generational companies that may have something to do with AI and machine learning?

I’m probably more on the side that AI will have a much bigger impact on existing modern technologies than sort of an AI-first company emerging that does a bunch of these new things. I think there are some companies that are pure AI-first that are building powerful platforms that I think have a strong chance of being successful, but because you need a large amount of data and interaction already to be able to make use of that, the companies that should have a strong benefit or advantage should be companies like Slack and Box and Workday and Salesforce. Because they already have the usage, they already have the data, and now AI is just a turbocharger for what people can do.

AI will be something where you will get disrupted if you don’t have it, but I believe it’s an enabling technology as opposed to a disruptive technology because it’s intended to dramatically enhance our existing tools and services. Now, the one caveat I would say, just so I don’t look stupid and not like somebody claiming that phones need keyboards or something, is if there’s a fundamental shift in where AI is in five or 10 years from now and it does things we simply can’t even contemplate today, potentially you would want completely new types of experiences to be built off of that that existing incumbents couldn’t adapt to for business model reasons or whatever. That would be what makes it disruptive.

Five years from now, are there technological tools you hope you have that you don’t have today?

I think virtual reality will probably have more impact in the enterprise space than it does in the consumer space at its mature peak, because I think the implications are so enormous that it at least has the makings of being one of the very, very fundamentally disruptive technologies.

We sort of see VR as maybe being to the non-office worker as iPad was to the knowledge worker, which is the iPad was the first thing that was like, okay, I can kind of digitize my entire mobile experience around documents and email and communication. There’s the perfect form factor for being able to go and do that. I think VR has the ability to bring that same level of immersive experience, but in a world where we haven’t necessarily enabled people with best-in-class technology. Whether that’s manufacturing, when you’re designing a product or you want to go look at the building that you’re creating, or in education, or if you’re in healthcare and you want to be able to instantly learn about or interact with multiple patients and different experiences.

We always overestimate the change in two or three years and underestimate it in a decade. It’s probably in that type of category, which is things that we think about VR being useful for today are just based on us just imagining a very linear shift. It’s going to be more of a function change.


How will Trump’s election change the course of business — either yours or Silicon Valley’s in general? Are you making contingency plans you weren’t before for the next four years?

This election exposed serious divisions throughout the nation. At Box, we’re committed to our values of openness and equality and will do everything to fight for them, including being an advocate for racial and gender equality, LGBTQ rights, fair immigration policies, and more. There are many policies we’ll need to debate in tech, but equality should be a non-negotiable principle for America.

Second, this election showed that many communities have been left behind as we’ve entered the 21st century. In Silicon Valley, we need to be more honest about the negative consequences technology can have for people — both in America and abroad. This isn’t to say we shouldn’t aggressively push forward with innovation. It simply means that we have to work better with each other, the government, and our communities to help the benefits of innovation be more broadly shared and advance policies that help people weather the disruptions technology can cause. This is only going get to get more important. Technology is on the verge of re-shaping industries like healthcare, transportation, and manufacturing, just to name a few.

Finally, we all want America to remain a leader in innovation and the global economy. We have to recognize that it’s not a given we’ll end up ahead. America’s role in the world of technology is unmistakable today, but maintaining our leadership is going to require making smart policy choices, cultivating partnerships between the public and private sectors, renewed investment in education, and fostering a culture of entrepreneurship. We can’t take anything for granted.

This interview has been edited and condensed for clarity.


Editorial Lead: Michael Zelenko; Design: Frank Bi, Yuri Victor, James Bareham, William Joel, Georgia Cowley; Photography: James Bareham; Development: Frank Bi, Yuri Victor; Illustrations: Slanted Studios.