ServiceNow, Inc. (NOW) Earnings Call Transcript & Summary

All right. Every year, there are a lot of disruptions, which actually occurs in the tech world. Before the pandemic, it used to be every 3 to 5 years. But during the pandemic and after the pandemic, it just got accelerated, especially after the pandemic and when our world got took over by the generative AI, it is now monthly or quarterly, to exaggerate, it's even yearly at the moment. Taking advantage of those disruptions and be able to flex with those changes and the times is what actually differentiate the leaders versus the rest of the pack. I'm really excited to be here presenting some of the amazing research our team has done in terms of low-code, generative AI, automations, all of it. In this session, we will cover leading trends, including gen AI, hyperautomation and some of the best practices, which we have seen our customers embracing every day. This is a safe harbor, something we all do. And some of this information which we are sharing today may be forward-looking and the time lines, features, et cetera, are subject to change. Now I'm super excited to have one of our amazing research leader join me today, Anna. But prior to that, let me introduce myself, Jithin Bhasker, I'm Vice President, General Manager at ServiceNow. Anna, over to you.

Thanks, Jithin. Hi, everyone. My name is Anna Byers, and I'm a Senior Manager of UX Research here at ServiceNow, and my team leads the research for our developer experiences here at ServiceNow.

Excellent. Here is the agenda. We're going to cover -- you've seen the intro, we're going to do a quick intro. Then we will talk about the landscape, how the demand for apps are growing, the evolution of generative AI, low-code and the perception of AI from a developer point of view. And we are also going to do an exciting Q&A as we speak about some of these in depth as we go forward. Now with all the technologies changes happening every day, seeing our customers innovating some amazing things, leveraging our platform is what truly inspires us. As we see some of the examples here, it's not just about the cost saving. It's about the real differences and the impact. Every one of these creative ideas, which translate into process automations, business process optimizations, all of it, this is what inspires us as a company, as a team, as a product leadership within. Now as you look at it, some of the examples, which we are going to cover today you can see very evidently that there is a huge amount of demand for digitization. It has grown significantly over the last pre-pandemic, pandemic and the post, especially with all of those things, which I just shared earlier, those who actually innovate fast gain that significance or the early mover advantage in the market. And as you see, some of the forces which truly drives the market today or this is what keeps a lot of the leaders, I have the privilege to talk to a number of CIOs on a monthly basis. And everyone has some of these on top of mind. It's about, how can I accelerate the productivity? How can I be more agile? How can I actually reduce the cost? And most importantly, how can I actually continue to build my talent and the workforce within the company to be able to compete fast, to be able to get to market faster in terms of the innovations and products they are building. Now all in all, that digitization or those apprehensions and cautiousness is what truly driving the evolution and the explosion of low-code in the market. Effectively, what we have seen through the creativity of our customers and what we see them apply are basically immense amount of productivity gains through low-code, almost like 66% reduction in terms of time to deploy or develop and increase in some of the key metrics like revenues and the pipeline and also time to market or time to ship from a delivery of an idea from a concept to ship from an application itself point of view. Now while we speak about all these amazing innovations and the excitement about the low-code, there is an apprehension which a lot of CIOs have, too. It's about how do you make sure you have the right guardrails and governance in place so that you can now leverage the skill set and the capability of the broader organizations. Often when I'm in front of the CIOs, I speak about leveraging someone from the sales operations or the marketing operations to be able to truly build those next level of process automation on optimizations you could actually deliver because they are the one who are in the forefront of the day-to-day fire fighting, they are the best one who can actually identify what need to be automated or optimized. Now the challenge is, how do you make sure you enable them and that's where the low-code truly comes in. And -- but on top of it, how do you then make sure as a CIO or an IT organization, you have the right controls and governance in terms of who gets access to what data, what process, what can they build and how you actually deploy and what are the security and the privacy data guardrails, which you put overall in order to make it highly adopted application more or less in terms of how the usage actually grows. Now every company and every leader is expected to have an AI strategy. This is what drives every one of our product leaders, CIOs, technology leaders in the market. According to McKinsey, generative AI's impact on productivity could be in trillions of dollars of value added in the global economy. It's just the beginning. And we see it every day in our own life, starting with last years of introduction of OpenAI and ChatGPT and a number of other new models, which we are now starting to test, understand and deploy as a part of our product capability and the portfolio as well. Now as the generative AI transforms the market and in general, how that's going to effectively drive the growth in the market according to Bloomberg, generative AI, more or less, the market is going to grow up to like a $40 billion to $1.3 trillion over the next 10 years to come. And fundamentally, it's not just about the developers in terms of how generative AI capabilities is going to help and support. It's also going to immensely help some of the non-developers as well. You can imagine the complexity of getting to once you have a pain point in terms of a process which you have identified, there's a lot of steps involved in between for you now to be able to actually come up with a process map, design the right optimization within the process map before you now start to work with IT to be able to build automations on an application which now can automate. Now all of those things are cut short with the help of the combination, especially with generative AI and the low-code application. And the way I explain it normally is with a single English prompt or a statement, the generative AI or we call it in our world, app generation capability, which fundamentally is text to app kind of a capability where in a few minutes, you have the ability to build an application almost 60% to 70% done. Now the last 30% or 40% is where truly the power of low-code comes in. Instead of scripting or coding, you now have the ability to click, drag and drop, change the fonts, change the headers, all of it with simple clicks and simple editing options, which comes through a local platform. That is going to truly disrupt and change the way how generative AI and the low-code comes together. And gen AI is actually going to dominate the way every app is developed as the years to come. Now on that note, I am really excited to now invite my amazing coworker, Anna Byers, who will walk us through some of the latest research we have done with our developers. Anna, over to you.

Yes. Thank you so much, Jithin. So as Jithin mentioned, my team has been doing a lot of work on understanding our developer population and understanding their perceptions of these tools. So today, I'm here to share insights from the study that we ran late last year, where we conducted a survey of developers to understand their perceptions on the use of AI in their development work. We received 254 responses from developers across geographies, industries and job roles. And importantly, we wanted to understand how developer's skill level might impact their perceptions of AI. So we categorized our respondents into buckets based on their experience with development activities. So we grouped no and low-code together into an entry-level group. And then we have 2 additional buckets for our mid-skilled and highly skilled developers. And another important note is that these findings are not just relevant to ServiceNow, but across development platforms. So you can see we had only 16% of the respondents to this survey indicating that ServiceNow was their primary platform. So these results can really help us understand the technical direction that many development tool providers are bringing to market and really help them understand -- help us understand how to target those tools to the developers appropriately. So let's jump into it. So across the board, the perception of AI by developers is overwhelmingly positive. They are generally just excited about AI for their work. So nearly all of our respondents, 95%, said that they're already using AI in the development work. And then an additional 86% say that they're excited about that potential of AI and then most importantly, over 80% have high levels of trust in AI, but it's really worth noting that a lot of that trust comes from prior experience using AI. So really giving them the trust that they need by having positive experiences with AI. When we asked what they think the benefits are to using AI in development work, about 40% cite the productivity gains that AI can provide to them, particularly through things like auto-generated code or code suggestions and functions. So it is also important to understand what developers' concerns are when it comes to AI, so that we can make sure that we're addressing those in our AI tools. The biggest area of concern is around privacy and security. And we know from our other research that security is a huge concern for development teams. So it's unsurprising that, that would also be true for AI. We also see that about 1/3 of our developers question whether AI will even really improve their code quality and they think that AI tools might not be accurate. And again, that might not be that surprising if we think about the new stories we hear about tools like Chat GPT generating incorrect information. So we really want to make sure that as we're developing these AI tools that we are making sure they are as accurate as possible and giving those developers the confidence that they need to continue using them. For me, one of the most interesting things that we uncovered in this survey and have validated with other research we've done is that the developer's skill level impacts how they perceive AI tools. So with our high-scale developers, they tend see the most value in AI tools. They're more likely to have already used those AI tools and report that it increased their productivity. They're also more confident in the accuracy of AI. And from qualitative research, we hear that they are less concerned when the AI output is somewhat inaccurate because they know enough to figure out how to quickly fix that and still see the productivity benefits. With our mid-scale group, I'd like to describe them as cautiously optimistic. So like our high-skill developers, they also report those efficiency gains from AI particularly in terms of being able to complete those repetitive tasks more quickly. However, they do express some more concerns around transparency, in particular, knowing when they're using AI. And then finally, for our no and low-code developers, those are the most wary. They tend to have more concerns around transparency. But for them, it's about not knowing how the AI works and not understanding it. They also report some more concerns about job loss to AI, which kind of makes sense if you think about folks who are maybe early in their careers and are primarily responsible for that less complex development work that might be able to be replaced by AI. But there is a silver lining, our entry-level developers are also optimistic that AI can help them grow their skill set and allow them to focus on that more meaningful work. And so we can kind of assuage some of those concerns by helping them grow their expertise. So overall, that means that our approach to AI tools for development work really must take into consideration the types of developers that will be using those tools so that we can ensure we're maximizing the benefit and minimizing the risk for each of those groups. So let's talk a little bit more about these top perceptions around pros and cons of AI development tools. So first of all, it's worth noting that the positive far outweighs the negative. Across the board, you can see that over 80% believe that AI tools are valuable and will have a positive impact on them and their company. In terms of benefits on their actual work, they believe that AI will allow them to be more innovative, more productive and then free them up to solving more complex development problems. On the flip side, we see that 3/4 say that AI-generated code will require more rigorous testing, which comes back to some of those points about security concerns. We also see that more than half believe that AI could lead to job loss for developers, although remember, that's largely driven by our entry-level devs. And then there's also a sentiment that AI is just another automation tool, which is true and not necessarily a bad thing, right? Like we -- automation tools are helping to improve productivity and accelerate app development. So it's just worth keeping that in mind and that perception in mind as we think about tool adoption. And then our last 2 points here relate to the ethics and security concerns, which are both critical for the adoption of AI tools. We need to make sure we're giving our development teams peace of mind that using AI development tools is not going to lead to either irresponsible development or sacrifice the security of their data. So as we think about AI-assisted development, there's really 5 considerations across 2 key areas. So starting with the ethics, we really want to make sure that we guard against risk. And that means having baked-in testing for both AI and developer-generated code. We know that AI-generated code tends to be buggier than developer-generated code. But on the flip side, AI-generated code is also easier to debug with testing. And so because AI excels at testing developer-generated code, we can incorporate AI-assisted testing into our tooling, and that will even add further to our productivity improvements. Next, AI recommendations can provide really great inspiration, but that user control is essential. So like we talked about earlier, developers are comfortable with some degree of incorrectness as long as they can control whether or not to accept that AI recommendation. And we commonly refer to this as human in the loop. So for even our more seasoned developers, even those incorrect recommendations can serve as that inspiration for future tasks. And then finally, in terms of ethics, developers also are going to need explainability and transparency on how AI assistance works. So reminding them when they're interacting with AI and also how that AI works. We see from some research that they sometimes can have expectations of what AI can do that might not necessarily be accurate or even applicable to that particular situation. And so without proper explanation and having transparency, that could lead to loss of trust. So internally, ServiceNow has actually developed and is utilizing a set of human-centered AI principles across all our AI products that include these listed here and several others. In addition, we recently joined the AI Alliance, which is an international community of leading technology developers, researchers and adopters who are committed to advancing open, safe and responsible AI. So if we align on these responsible AI principles, then we can unlock the benefits of AI. Use of AI in development can lead to measurable productivity gains by reducing the amount of time it takes to complete development tasks. It's also a valuable learning tool as long as we ensure that the content is there and that the recommendations are explainable. Both skilled and novice developers can benefit from the guidance that AI assistance provides as a way to upskill and adhere to best practices, again, as long as those things are clearly explained. And finally, for our entry-level developers, that in-context content on how or why an action is recommended, can help them learn as they go and build their skills. So to conclude, I'll share a little bit about what this means for ServiceNow. As we create our AI experiences, in particular, with generative AI, we want to focus on maintaining transparency, providing guidance and control and ensuring adequate guardrails for safe AI usage. Embedded generative AI also allows ServiceNow to have much tighter control over what gen AI does and how it performs. It starts with how we build and maintain our large language models or LLMs. We can control the training data sets, and so we're able to detect and remove bias and also eliminate the possibility of false information from external sources leaking into our AI. More broadly, as we follow these human-centric AI guidelines, which are critical for responsible AI, we have that control. Our embedded gen AI approach also addresses those security and privacy concerns since gen AI is constrained by our access control lists or ACLs and protected by the security on the ServiceNow instance. And because gen AI is integrated into the Now Platform, it's under our control, we're able to make it reliable and enterprise-ready. So now I'll turn it back over to Jithin for some questions.

Thank you, Anna. What a fascinating set of data, research and statistics. So I have a lot of questions, but I'm trimming down in terms of what those are. Really great insight in terms of what you've just shared.

I think the first thing which comes to mind, Anna, is about what really excites the developers? Because I see that according to the research, 95% of them are using generative AI. That is probably the fastest way any technology was adopted by the developer community in the history of computer science probably, right? So say more.

Yes. Thank you. It is really interesting to me, and we were talking about this in another conversation earlier that generally, we find developers to be pretty skeptical as a group and so it's interesting to see the level of excitement here. And I think it really comes down to kind of the thing that motivates developers at their core, is that they are problem solvers and they like to go fast. And so if there is a tool out there that can help them move more quickly and be more efficient, they're going to be excited about it. And I think that's what the possibility of AI is to them.

Exciting Yes, for sure. How about the other way in terms of what the concerns you've seen or anything else go beyond what you shared earlier?

Yes. So certainly, I think trust is really the biggest concern area, right? We want to make sure that we are creating tools that the developers can trust, both in terms of accuracy but also in terms of knowing what it's doing, right, knowing where the data is coming from. We know that our platform teams and particularly our system administrators, right, they won't even roll out new technology to their teams until they feel confident that it's not going to disrupt their current system, that it's not going to present any potential security risks and keep their data safe. So we really want to make sure that we're creating things, first of all, that are accurate and as trustworthy as possible and then also sharing that information as we release those products to our customers so that they know, hey, I can trust this. My data is safe. My platform is safe.

Excellent. Yes. After hearing all the pros and cons, what is actually driving this high amount of adoption in general across the community?

Yes, absolutely. It's a great question, right? Because we do -- we see that [Audio Gap] already starting to use AI tools in their development work. This one bit that comes to mind, right, is that code suggestions, right, or entire functions. So that was a really early and easy [Audio Gap] of a large language model is, I don't know this code snippet or I can't remember this code snippet, AI can help me. And even, again, if it's not totally right, it's good enough for a skilled developer to know, oh, yes, like I couldn't remember that syntax, but now I've got it, and I can build from there, right? So again, it helps them move more quickly than having to go to like Google and like be like, what was this syntax for this thing? So really in line, allowing them to work more quickly. But I think it's also important to think about how that might impact, not just pro coders, right, who are scripting, but folks using our lower code tools. And again, I think there, it's really about helping them learn and helping them navigate their build, right? And it comes back to maybe they don't remember how to do something or maybe they're trying to figure out the best way to solve the problem. And if AI can make a suggestion for that, that helps them get a really good starting point that they can build from.

Interesting. Yes. On that note about the non-pro developers or entry-level developers. My -- somehow I assume that they'll be the most excited about this, but you said there are certain concerns they have. And what do you think -- what can make it optimistic for them?