How Adobe is leveraging AI

SHANTANU NARAYEN: I wish I could look at you and say, “Yeah, yeah, we knew all of that up front.” We didn’t.

REID HOFFMAN: AI models are upending the market for tools that help people make things like illustrations and animations. But this isn’t the first time Adobe CEO Shantanu Narayen has had to lead the iconic software company through a moment of monumental change. He spearheaded Adobe’s transition to a cloud-based subscription model more than a decade ago.

NARAYEN: There was a lot of opposition from existing customers as well because they were a little skeptical and/or cynical about why Adobe was doing it, but there was a lot of excitement.

HOFFMAN: This transformation wasn’t just about the sales model for Adobe. It also revealed a more powerful way to respond to customers’ needs.

NARAYEN: People would sit around the room, you’d say, “Hey, I think these three features are most important.” I would say, “No, I think these three features are most important.” Perhaps the person who had the loudest voice won that particular battle, right? Today, it’s just the opposite.

HOFFMAN: Shantanu is as passionate as ever about helping customers unleash their creativity. This week, we’ll hear how he stays nimble, curious, and competitive while leading a global team of more than 30,000 people.

I’m Reid Hoffman, your host.

[THEME MUSIC]

Adobe makes essential software for creative professionals. There’s Photoshop for editing images, Illustrator for creating designs, and now its own generative AI imaging tool: Firefly. All together, Adobe’s offerings brought in more than 21 billion dollars in revenue last year.

I wanted to talk to CEO Shantanu Narayen about how he ensures that his long-established brand stays innovative.

Shantanu, good to see you. Welcome to Masters of Scale.

NARAYEN: Thanks for having me, Reid. Excited.

Copy LinkThe journey to becoming Adobe’s CEO

HOFFMAN: I’ve been looking forward to this a while, and one of the things that even with people I know that I love about doing this show, this interview, is to be able to ask some questions that we actually haven’t covered in our various dinners and other things. So, what was Adobe and the current state of software like when you first joined the company back in 1998?

NARAYEN: When I joined Adobe in ’98, we were hard at work on Indesign. Adobe had this vision for: how do we create a brand new platform from scratch that allows people to do professional layout? And that was one of my first products that I worked on at Adobe.

HOFFMAN: Less than a decade after you joined Adobe, you became its CEO. Was it a goal of yours? What was the process in which that happened? Was it a surprise?

NARAYEN: It was a surprise, even now when I look back upon it. Reid you sort of have to pinch yourself because it was magical. I think the company went through a couple of big transformations. I joined in June of ’98. There was an economic issue in Japan. The company hit the wall. And we had to have a very significant layoff. And the company sort of said, “Hey, we know we hired you as a, I’m not a quote unquote, GM, but you know, we want you to head up all layout technology.” And so hopefully the company saw something in me which said, “Okay, I’m a product guy at heart, and we’re a product company at heart, and I will do whatever the company wants,” which was important.

And I think it was that flexibility. I would also say you have to be the right person, right time. And Bruce Chizen, who is the CEO before me, and in many ways gave me a lot of these opportunities, he was more on the sales and marketing side.

And so, I was lucky in that I complimented him with a lot of the business side. And then he decided he wanted to retire, so I can honestly say it wasn’t like, okay, I was plotting how to get from point A to point B. I think I always showed initiative. I love what I did. but when that final tap on the shoulder came and said “We’d like you to be the next CEO” was one of the most professional moments of my life.

Copy LinkThe leadership lessons that Shantanu has learned as CEO

HOFFMAN: Well, obviously, it’s been nearly 20 years, so it’s been an awesome run. Let’s start a little bit with your leadership approach. When you came into the job, what was your kind of theory of being a CEO leading an organization like Adobe, a scale public company? How has your approach changed? What would you say are the key things that scale public CEOs should do in their leadership of the organization?

NARAYEN: I was fortunate in that I grew up within the company, as I said, and I was CEO and I was head of products. And so I would say I made the mistake of thinking that even in this new role, I should just continue to do the kinds of things that I was doing.

And I think all executives have to really sort of take a step back. I try to do that now every year and say there’s so much, sort of, table stakes of what you do, but where can you really have impact, and where can you really make a difference? Because when you have as vibrant an organization as Adobe has, or the other companies, for a lot of that, you have incredible employees who can take that on.

And so, I think initially I was like, okay, let me just continue to do the things that I was doing. And in retrospect, that was a mistake because you always have to think about how you’re disrupting yourself. You always have to think about: Where are you uniquely either qualified or expected to impact change? And impacting change — the larger the company gets the more difficult it potentially is.

And so I would say the management style has changed a lot more as it relates to thinking about every year. Okay, are there one or two areas that I should — AI, for example, and trying to understand AI over the last year has taken up a disproportionate amount of our time because we as leaders, our most valuable commodity is time.

So I would say that’s one learning. The second learning I would say is that when I first took over, I was like, “Wow, these are all my peers. And let’s just keep doing what we did.” But I think as a leader, you have to be way more comfortable with ambiguity and way more comfortable with providing direction.

I’ve been surprised to some degree how much people want you to opine. And I’m very comfortable with it, but initially I thought I would say, “Wow, this is a good group of people.” And so I think that way also my management style has changed a little bit.

HOFFMAN: And how do you, kind of, balance between that setting of strategic direction, the coordination of the exec staff, the deliberate creation of culture — what are the kind of the elements that are your style, and what are the things that you would recommend other leaders pay attention to as you put those together?

NARAYEN: I think most leaders, and it doesn’t matter at what level you’re providing leadership, you have to create a vision. And if you’re a first line manager, you’re creating a vision for the group. If you’re CEO, you’re hopefully creating a vision for the company. You have to have the team in place. Who is the right team? Are you leveraging the superpowers of the team that exists? So I mean, again, whether it’s a small group or a large company, and third, I would say you have to have a cadence for execution because having the strategy without the associated execution is not going to be effective.

And so I think with scale, how you do it and where you do it changes. But at the end of the day, I still have eight or 10 people reporting to me, right? And I have to make sure that I’m empowering them to do the right thing. So there is more commonality than I think people sometimes appreciate.

Copy LinkAdobe’s big pivot to the cloud

HOFFMAN: One of the, kind of, biggest pivots the company has made under your leadership is the shift to the creative cloud subscription model. How was it thought about, how are the risks managed? What made you conclude that this was an important thing to do?

NARAYEN: If you go back to where we were. I mean, the reality was the company was doing really quite well, but when the recession of 2009 hit, we were a considered purchase.

And so given we were a considered purchase and potentially a discretionary purchase, the revenue dropped quite precipitously. And when the revenue drops, in a company that is software, which, as you know, the biggest expense that you have is people-related expenses. And so it really impacted me when we had to let people go because we were doing that not because of anything that they hadn’t done, but because of the external environment.

And so trying to achieve financial stability and financial growth was a key priority on the other side. The product teams would say, “Hey, we have this 12 or 18 month product cycle. Innovation doesn’t happen in 12 or 18 months. It’s happening at a far more rapid pace. Look at what’s happening with the cloud. Look at what’s happening with mobile.”

I think people look at it and say, “Wow, Adobe was the first to pivot.” Maybe we did just a better job of listening to what the incredible employees at the company were saying. “Hey, why aren’t you allowing us to innovate faster?”

You mentioned the word risk and sometimes I say, Reid, I didn’t view it as risk. For some reason, risk seems irresponsible to me. I would say we made an investment. Sometimes those investments work and I, of all people, I mean, you’re the leader in thinking about venture and how you do investments. And so we made that investment. If it didn’t work, we would have had to adapt and transform.

Copy LinkImportance of being transparent with customers

HOFFMAN: This kind of transformation must have also been met with some uncertainty, concern, reaction from customers, the creatives. How did you kind of navigate through that?

NARAYEN: It’s a great question. And as I think about it, and I think that’s true even now with AI, you have always two sets of customers, one of whom is going to embrace that immediately and see the benefit of it. And there are others who are probably going to have a little bit more inertia as it relates to change.

So, candidly, there was a lot of opposition from existing customers as well because they were a little skeptical and/or cynical about why Adobe was doing it, but there was a lot of excitement.

The strategic decision to move to the cloud and the financial model in retrospect, that may have actually been the easy one executing on it because, in parallel, we actually allowed people to buy the perpetual version or the subscription business.

We had to transform how the company thought about delivering software that was expected to be always on. We had to make sure that we had compatibility between each of the versions of Photoshop that were out there.

So the heavy lifting on executing this flawlessly, in many ways, I think that’s the under appreciated aspect of what the Adobe employees did.

But I think that also stuck with me, which is ‘let’s be thoughtful. Let’s be deliberate. Let’s communicate with our customers.’

That was the other unintended benefit. Everything that we were working on, we had no reason to hide from our customers, right, Reid? I mean, we could basically, even today, we can say we’re working on these 10 features. If you want to vote on which ones are the most important, great, give us that feedback. When you had this secretive, ‘I’m working on this for 12 or 18 months, and I can’t talk to you because if I talk to you, you may not buy my current version.’ So, you know, it was self reinforced by all these incredible benefits that we got that weren’t completely obvious. I wish I could look at you and say, “Yeah, yeah, we knew all of that up front.” We didn’t.

Copy LinkAdobe’s ‘data-driven operating model’

HOFFMAN: When Adobe switched to cloud-based subscriptions for their software, Shantanu’s teams not only had to sell the product in new ways, but also measure success differently.

NARAYEN: So at the same time that we created this new model, we actually instituted within the company what we call a data-driven operating model. And the data-driven operating model talks about discover, trial, buy, use, and renew. And it speaks to the customer life cycle. So we said, “Hey, in this new model, how are customers going to discover us? How are they going to try out the software? What do they buy? What’s the right version? What’s the usage?”

And then hopefully they’re new. And this focus on usage started. How our products developed prior is that people would sit around the room, you’d say, “Hey, I think these three features are most important.” I would say, “No, I think these three features are most important.” Perhaps the person who had the loudest voice won that particular battle, right? Today, it’s just the opposite. People are like, “Hey, here are some ideas.” Either engage with your community or even better, put it into the software, market it, and see what gets used. And we had real surprises of what people were using.

And it empowered product people because they finally could, in these meetings, tell us, “Hey, we know exactly what features are being used.”

So we’re going to continue to invest in that. And ironically, you have less debates when that’s what the usage reflects.

HOFFMAN: Still ahead, how Shantanu is leading Adobe through an AI-fueled transformation.

[AD BREAK]

Welcome back to Masters of Scale. You can find this conversation and more on our YouTube channel. Next, I wanted to talk to Shantanu about Adobe’s approach to AI.

Copy LinkHow Adobe is approaching AI

NARAYEN: One of the things I think we do reasonably well at the company, Reid, is when you have such a profound technology shift that’s happening, we really try to take a step back and sort of say, “What are some core hypotheses that we believe in that will guide the next few years of development in this particular area?” And I’ll give you an example of some of the things that we said in AI.

I mean, we represent the creative community. So we said, “If we create models, one of the things that we have to do is create a model where every piece of data that went into the creation of that model, we can testify to the fact that we have license to do that.” We created a hypothesis where we said, “What do people fear the most when you’re trying to do creative expression? The blank page.” So if AI can enable you to take the idea that’s in your head and express it in forms that are different from just using a menu, but you can have a conversational interface with it, why wouldn’t that further make our products more affordable, more accessible, more fun to use?

And so we have to embrace AI despite people saying potentially that there’s opposition to it because it may replace humans. We said, “People who use AI, this is an augmentation tool, and it will potentially replace people who don’t use AI.” And then if we don’t do it as Adobe, somebody else is going to do it, and we will find ourselves disrupted. And so I think as we started to think about it, we said on the ideation of creativity, on the prototyping, at the front end of the entire process, this is going to be the biggest boon to creativity because people can use other modalities to talk about what they want to do.

Within the creation of the models, let’s be clear. Let’s be clear on how we build our own models and why that’s important for us for controllability. Let’s understand how we support third party models, because, as you know, all of these models will have slightly different characteristics, profiles, behaviors. And then we said, “But at the end of the day, it’s in the interfaces that we build that humans will actually derive the value associated with it.” So I think we had really good discourse and conversations around what is our stack, what’s our framework, and what are we doing in each one of these spaces across customer segments. And that still guides us.

Copy LinkWhy experimentation is so important for Adobe

HOFFMAN: One of the things that I think you’ve done a really good job at that a lot of tech companies and others are going to have to figure out is like, “Okay, so we’re going to be using an AI.” How much did you decide “We’re going to train our own?” How much did you decide “I’m going to leverage open source,” how much did you decide “we’re going to partner?” What were the balances of those considerations? Where are you now? Where do you think you’re going in terms of the models and the various kind of generative capabilities that are part of their our current world.

NARAYEN: Part of what I’ve always admired, Reid, is Sand Hill Road and how Sand Hill Road says as the biggest creation of companies and capital in the world is like: Run multiple experiments, right? I mean, how can you say yes to as many of those things as you can, because they’re all learning. And so I think it’s probably fair to say at Adobe, we’re running experiments and innovating across each one of those different approaches. On the marketing side where we have one of the largest marketing technology businesses in the world, that’s where we process so much customer data there. The customers are like, “Hey, we don’t want our data used for you to commercialize, but if you can learn from our data and others data in an anonymous way, great. Then that improves the state of the art for everybody.” So I think they’re different.

We just recently introduced AI assistant for PDF for Acrobat. You can have a conversational interface now with a PDF. The next step clearly is how do you do it across all of the documents, whether they are on your desktop or whether they are in the knowledge management system in the enterprise? And then the third step on that is how do you do it for vertical documents?

We announced Contracts AI, which understands contracts and the semantic meaning of it. So I think like all good companies, you experiment in different ways, and then you sort of bring the teams together and say, “Hey, how do we learn? And what’s the assertion that we’re going to make? And what’s the path forward?”

But I think shutting down options too early in a company of our size and scale is actually a mistake. And so this is all about learning and experimentation. And then at some point, you have to make a decision. And you have to say, “Hey, okay. This is the path that we’re doing, but why shut it down prematurely?”

Copy LinkHow will AI impact jobs?

HOFFMAN: I’ve been here in New York launching my book, Super Agency. Part of this is that AI is augmentation, amplification, intelligence, it gives us superpowers. Obviously, one of the things that people have is a worry, as workers, is where is it replacement?

NARAYEN: I think the reality with all technology, all automation, is it introduces way more jobs, and it disrupts some, and I don’t think AI will be that different. And a lot of the automation or menial things that people do, it’ll force individuals to rise to a higher plane in terms of the work that they do for creativity.

We think it’s that the focus on that emotional story that you’re trying to do and focus on the customer segment. But it will change a lot of the, I would say, grunt work, and some people will find that threatening, but you’ve spent way more time. And you know where AGI is going and where all of this stuff is going.

But so I think you’ll have both in parallel. And every single industrial revolution or, you know, change that has happened, has simultaneously disrupted some people, but created way more opportunity. And I hope this is the same.

Copy LinkGo to where your customers are

HOFFMAN: Yeah, I obviously completely agree. When you’re looking at the future, all kinds of things are changing. Customers are changing, technologies changing, competitions, start-ups are changing in Sand Hill Road, throwing lots of different companies out and figuring that out. How much of it do you think it’s kind of like a “Hey, we’ve got this true north of how we’re building, we’re doing it internally,” and how much of it will be like M&A or partnering or licensing things?

NARAYEN: I think we take an approach where ‘let’s understand the stuff that we do well.’ I mean, with Acrobat and with AI Assistant and the pure tech stuff, we focus more on the applicability on that. With the large language models, I’m sure you’d agree there’d be four or five people who have the capital to create those large language models. There’ll be maybe one open source and, so I think we take a very prudent, pragmatic approach, and the reality is, also, if you’re an enterprise customer, some of you are going to standardize on. It’s no different, for me, in how Adobe has thrived by saying if there are different operating systems or platform, it’s in Adobe’s best interest to support whatever operating system or platform becomes the platform in which people want to accomplish what we want.

We supported the Mac, and we supported Windows. When iOS and Android came, we supported Android and iOS. I think in many ways, and I’d be curious to your take, all these models are going to be a different form of platforms. We should embrace that because you know, you have to go to where people are right? And so, at least, I think some of these models are similar to how I’ve thought of operating systems. And it’s like, hey, where do we leverage the functionality that’s there? And where do we augment it by doing something? And so we’re like, hey, let’s go find another hill to climb where we can add differentiated value. So yes, we will partner, we will support the models, and we want to leverage. I mean, this has also become such a scale game in terms of dollars of investments that we’d be crazy not to leverage some of that investment.

Copy LinkHow AI is changing Hollywood

HOFFMAN: I 100 percent agree. And how are you finding your interface with the Hollywood industry? When I go down there and talk to them, they all are like, “Oh my God,” and I’m kind of like, “No, no, you should say, ‘Oh my god, AI, there’s things you can do now…'” How are you finding that? And are they beginning to realize what the kind of creative potential is?

NARAYEN: People understand the cost of producing these movies. The production and the post production costs. I mean, take animation, right? And think about how much cost goes into the animation. So, I think on the one hand, they’re all incredibly excited about the prospect of dramatically increasing the amount of media that they can create for the same cost. And accelerate the time to production. So that excitement, I think, exists. At the same time, to your point, the apprehension of what it means for certain roles, whether you’re a scriptwriter, whether you’re an actor, what that likeness is, frankly, it’s the large studios that are going to be at the forefront of determining what they do. But I think the tide is turning in terms of the fact that the genie is out of the bottle right now, and people recognize that this is something that they have to embrace. And to what degree and at what pace they embrace it, I think depends on who you are in that chain.

Copy LinkHow AI has impacted the job of a CEO

HOFFMAN: And how has AI changed your job? Like, what are the things that kind of change in the way that you operate as a CEO, as Adobe CEO?

NARAYEN: Well this is actually not Shantanu talking to you. It’s my AI avatar because the real Shantanu is out. No, no, it’s a great question, and I’m not sure it’s changed dramatically yet, honestly, I think you’ve seen the software development teams talk about how much more productive it’s made them. You hear the marketing organizations talk about how the agility in the marketing campaigns has gone up dramatically.

For me, I would say the biggest profound impact is I get a lot of mail and a lot of them have PDF attachments, and those are always the most thoughtful ones that I have to spend time on. And using AI assistant for me nowadays, every email that I get has the PDF attachment and it has the Acrobat summary, so that I can look at it. And so I think hopefully becoming more efficient, but if you have any tips on what else I should be doing to make myself more productive, I’m all yours.

HOFFMAN: I tend to think every professional will deploy with multiple AI agents. So almost everyone’s doing mini management and that then composes up as you go through a company. And so part of the thing that I’ve been anticipating is the future of how organizations are going to evolve is this notion of how agents are in the mix for individuals? And you can put that in a kind of a joking way. It’s well, my agent will talk to your agent but in the joke, there’s some interesting questions around information flow, decisioning.

NARAYEN: Well, I’m aligned in your thinking associated with, hey, agents at the end of the day have to understand roles and tasks and personas that they’re trying to represent and how they engage. I mean, I think the ultimate proof of how effective an agent is is when an agent can interface with an agent and an agent can interface with a human.

Copy LinkBroadening creativity with AI

HOFFMAN: Obviously part of the other amazing thing that AI brings is enabling creativity across even a broader range of human beings and humanity. Where do you see that kind of space of creativity opening up, and where do you see some experiments that Adobe might run or things that might try to broaden creativity across the human population?

NARAYEN: Super, super excited about the prospect of that Reid. And enabling anybody to be creative is something that’s a passion of ours, and more people are perhaps able to describe what they want to do in terms of creative expression, then use the tools, and why isn’t this the most empowering way for everybody to tell their story? And on another dimension, why is it in education right now that all education and reports are still primarily text and numbers? Why isn’t there an animation and a video and an image and a link to something else and feedback associated with every piece? And so when people talk about STEM and the importance of STEM, we’d like to talk about the importance of STEAM. The world without arts would be a really boring place.

I mean, can you imagine everything being one color and one font? And so we think that it’s going to be a massive unlock and one that we’re super excited about.

HOFFMAN: AI obviously creates huge amount of opportunity for reinvention. What are the principal challenges you see with it that you then need to help Adobe and the world navigate?

NARAYEN: Part of our values is how we think about technology to transform and unintended consequences of it. And so, are we scouring the earth also to understand where there are good ideas and learning from them or buying them as companies? It requires us to be way more agile. It requires us to be way more vigilant, and it requires us to embrace our own products. I’ve also always said people build their best products when they use them themselves. I think everybody is trying to think about every company and now they’re going to be an AI secular winner or an AI secular loser, right? And it’s incumbent on Adobe and us as leaders to demonstrate why AI is going to make us even more of a leader than we are today.

HOFFMAN: Shantanu, thank you very much. It’s been an honor and pleasure as always. And I look forward to the next time we have dinner.

NARAYEN: Really enjoyed it, Reid. Thank you so much for having me on your show.

HOFFMAN: I always love talking with Shantanu. I’m inspired by his optimism and curiosity about what our future holds. His decades of leadership at Adobe are an unparalleled example of how curiosity can keep your business competitive and your culture innovative.

I’m Reid Hoffman, thanks for listening.