All right, our next topic is going to be covering AI at scale. And who better to cover this than someone who stands at the helm of a $150 billion company as the president, COO, and chief product officer of ServiceNow. Please welcome Amit Zavery.
I'm so excited for this chat with Amit. Amit is a legend of Silicon Valley.
Good to have you.
Good to be here, Grace. Thank you for having me.
Thank you so much. As you know, I prepared so many questions, but I'm just going to ask off the top of my head because I feel like you're such an approachable icon. Okay, so I want to start with: Amit, you are the president and CPO of ServiceNow. I want to start with, what does your day-to-day look like for the audience who just arrived here? And I feel like everyone's probably curious, since you are a person with many hats.
Yeah, no, no. I think all of us are in this weird state right now, especially with AI. Things change so fast. I think our days are also changing very fast. It's very unpredictable in terms of how the day will evolve. But as part of my job, no doubt, I run products and engineering. So there's a decent amount of time spent on a daily basis looking at our product plans, reviewing product design, making sure we're building the right things, and taking customer feedback. A lot of other time is usually customer meetings. I usually meet at least five to ten customers a week, either on Zoom or in person, and then some events like this where we get a chance to really interact with a lot of the future-looking things which are going on in the industry, which gets me some input in terms of what we should be thinking about, who we should be partnering with, and how we should be thinking about go-to-market as well. And then just internal alignment, as part of the COO role, is really thinking about how we can operationally become more efficient, how we deliver things faster in a better way, and keep making sure that customers remain successful on our products. So again, think of everything from the customer angle, looking at our days to really ensure everything is aligned to get that outcome for us. So a little unpredictable day, but overall very satisfying because you get to do so many different things the whole day instead of being only on one thing. So I enjoy that part, and context switching is exciting because then you get to rethink things and then take that information and use it somewhere else as well. So that's most of my day.
I feel like one of the things that you didn't mention is AI agents. So I want to start with, since nowadays everyone at work is using different types of AI, what does work actually look like? How can you predict what work looks like a year from now?
Yeah, I think that — see, my view is AI agents should not be something you, as an individual, need to worry about. My view is that products which I use should be using AI to make my life better, as long as they are bringing all these latest technologies and making my days better or my way of working. Yeah, I think that the way we also build products, my goal is with my customers to not really think about AI directly, right? So for example, we're building this idea of an autonomous worker, and it takes equivalent work as humans do and takes care of like a level one support person, or a SecOps person, or an HR talent acquisition person. Using that autonomously, the idea is that you, as a user, should not know it's been driven by AI, what technology I'm using. The same thing with my tools I use daily, right? I am using those products with the idea that it is getting more efficient, is improving its ability, and I don't want to have to do something very differently than what I used to do. Even though the work is changing for sure, I want to understand the way the work is changing, but most humans are not used to completely changing the way their life is. Even though we might pretend in the valley that everybody's going to redo everything they do on a daily basis, that's not true. If you look at the enterprise user on a typical day, they still do 90% of the things the way they used to do. Maybe some things they will keep on tweaking. So that's why you have to bring the technologies into the way they're working versus the other way around. Expecting them to change their work is going to be a big, big barrier to entry for any products. My goal is to really make them efficient. Same thing, that's what I'm expecting from the products I use on a daily basis. It should make my day better without me needing to understand doing things differently. Right.
I love that. I wonder, since many of the people here — we have like many founders and investors, but I'm sure a lot of people are actually working in tech and are curious about, like, what are the skills that they need to upskill on and learn about, as well as, like, how do you think about building — I want to say like a moat, but more of like a competitive advantage for yourself in the AI era.
Yeah, no, I think as product people, right, we have to really think about what technologies can do to improve your products 100%. I think it has happened in many waves. AI is one more wave like that, which is changing what you build, how you build it, what kind of people you need to really have in your teams, what is the process around that as well, right? So those parts, I think, in your way of building things, definitely is changing. The scaling part is really finding the right people in many cases, right? So you need to have a blend of people who understand newer technologies with a blend of people who understand domain, and once you bring those things together inside smaller groups, you're able to see a lot of improvements. We are building a lot of new products on a regular basis, which are becoming much, I would say, faster from a delivery perspective, which is iterating in a different way we deliver to our customers. Our customers are interacting with these FDEs. That's a new thing, which is the forward-deployed engineers, which we used to do when we used to do software development before, where we used to send our product managers or engineers to work very closely with customers to get input and then evolve your product. But now the direct relationship on a constant basis with FDEs is getting the customer voice much faster into a product. So you can iterate, you can evolve and make changes, as well as take things which are done at a customer site back into a product. So that's, I think, the big area of scaling now, because you can now deliver improvements and it creates a lot of innovation in a different way than we used to do before.
For sure. I really like what you said about customers, because I feel like being customer-centric is extremely important for any company. So in terms of, I guess, one of the things — ServiceNow is a huge company, and how do you think about enterprise AI adoption? So obviously, many people can try different tools at their own seat, but at a company at scale, how do you evaluate what is, I would say, whether it's revenue generation or any type of benchmark that you're using to evaluate if they have sufficiently adopted AI?
Yeah. Yeah, no, this is the biggest question. I think most of the customers I talk to — and I talk to a lot of the C-level execs, especially CIOs and CISOs and others, everybody, and CFOs now for sure — are worried about the spend and the ROI. So AI maybe last year would be like, 'oh, let's try it out.' There was a lot of experimental stuff going on. The adoption definitely lacked in terms of the production use cases. The biggest barrier to usage was most of the time around governance, security — I'm talking about enterprise customers, right? Customers who have systems in place, who really want to do this thing properly without impacting their business negatively. So we saw a lot of customers who are reluctant to do mass-scale AI adoption because of worries around security, governance, risk, visibility, discovering shadow AI projects, and not knowing what you're spending on. And then there is an industry which is guilty of kind of doing stupid things like token maxing, right? Which has kind of created this weird structure, weird views inside businesses saying, 'what are you guys trying to tell us? Spend more without any kind of ROI and no idea what the return — what am I going to do with it?' So what we saw, and we built this product last year called AI Control Tower. We launched it middle of last year. The idea was to give CIOs, CFOs, and C-level execs a better visibility in terms of what's going on with AI: what are you spending on, who's using what, what is the ROI around it, what are the security issues around it, and just doing it across the heterogeneous environment in an open ecosystem. And suddenly, that capability, when we went and showed it to our customers, really opened up a lot of interest from our customers, because that gave them this idea that I have some peace of mind, I have visibility, I have control, and now I can do this thing systematically instead of just doing ad hoc random projects because somebody called me up and there's a fear of missing out. A lot of customers last year, there was this FOMO out there, and they're worried about the board telling them, 'are you doing anything on AI? What exciting new things have you done?' So a lot of those questions were coming. They're doing a lot of random stuff. Now I think it's a lot more structured, a lot more thinking behind it, and things like AI Control Tower, where it gives you this ability to really monitor and manage these things in a governed environment, opens up the reality of what AI can do for you. And now I think the projects we are seeing, like agentic workflows, what we're doing with autonomous workers, what we're doing with automating a lot of your business processes around IT, HR, employees, conversational front doors, customer service — now you're able to manage, monitor, and really calculate the value, and then go and invest in it. But it's much more thoughtfully done now than it probably was done before. And I think it's the right thing for the industry and the right thing for our customers, because then we are able to really get the value of the investment we're making for the products. Customers see that technical value and the ROI, and they will buy more. So that's our goal really, is to really create that value and show them the ability. And I think you will hear a lot of pushback otherwise. I mean, trying to go and just show you a feature or something like that without any kind of ROI clearly defined and value creation defined — it's an uphill battle and very, very hard to keep up with.
