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

Welcome, and thank you for joining us for today's event. Before we get started, we have a few housekeeping tips that will help make your experience more enjoyable. First, today's session is being recorded and you are currently in a listen-only mode. [Operator Instructions]

Welcome, everyone. In the morning and the afternoon and the evening. I'm Mike Hasenkamp, I'm a Senior Program Manager at ServiceNow as part of our digital technology team working in the emerging technologies group. And today, we're going to be talking about Now on Now Unlocking the Organizational Power of GenAI for employees using Now Assist. As I said, I'm Mike Hasenkamp, and I work as part of our Emerging Technologies Group. And part of my role there is to work with our product teams as they innovate and build new solutions on the Now platform, and this is how we're transforming our business, and of course, that includes GenAI. And today, we're going to be talking about Now Assist. Now Assist is how our customers and our employees are experiencing GenAI with ServiceNow. So I'm really excited to introduce our presenter today. This is Rajeev Sethi. He's going to join us in just a moment. He's our Group Vice President for Emerging Technologies. And as I said, as part of ETG, I know firsthand how passionate Rajeev is about GenAI and how it's transforming our business and how our customers are transforming their business. So welcome, Rajeev.

Thank you, Mike. Thank you, and welcome, everyone. As -- and thanks for the introduction, Mike. Before we get started, I want to make sure that you are -- this presentation contains some forward-looking statements or would currently things, which we may be doing internally within ServiceNow. So don't make any of those buying decisions based on what you see here, please connect with the account team or any other ServiceNow representative you may have connection with them and help them to understand what is available from a product perspective. So before we get started into our topic today, I just want to give you a little bit understanding about when we say Now and Now. Now and Now is this -- we have customers internally and part of ServiceNow BD team, some of you may call it IT. We closely work with our product team. We are -- we are, as I said, customer zero, so we are the first one to use our products internally implemented. In some cases, we are co-designing, co-developmenting with our product team before it gets up to you. So really passionate about our ServiceNow platform. And obviously, our products as part of the Emerging Technology Group, my team also looks at what is happening outside ServiceNow and bring that inside our 4 walls to make sure our -- it is driving value for our employees and then for customers. So that's a little bit about Now and Now and being Customer Zero. So jumping into Generative AI agenda for today at ServiceNow. We're going to go through like what are the challenges we encountered and some of them, maybe you are also seeing it today and the approach we have taken. How we see the Now platform is transforming into becoming an AI platform for business transformation. We're going to look at some of them Now and Assist that how it's unlocking some of the productivity. And in case you're interested, how do you get started with generative AI? So like how did we get started? I'm sure there will be some product questions. So if you have any questions, please post it into the Q&A panel, which you may see on your screen, and we will get to as many questions as possible as we wrap up the session today. So as we get started, I just need one -- another favor from you. Would like to see -- we'll give about 30 seconds to just get some feedback from you all what best describes your organization approach to the journey with generative AI. So are you just absorbing what's going out? Are you exploring out there? Are you evaluating? There are some use cases, doing some POCs out there or we call it -- some of them is pioneers there. They are early adopter and actively integrating GenAI and processes and products, and doesn't need to be with ServiceNow platform. It can be anything out there because I know there are a lot of generative AI solutions and platforms available. So just love to see your feedback. We'll probably give another 10 seconds there and see the responses. We'll share the responses with you. So please go ahead and click any of those A, B, C or D. All right, let's see how we are doing. So it looks like majority of you are in the exploring. We are currently experimenting with generative AI, looks like some of you are still trying to monitor what's going on but actually not engage in it. So hopefully, this is one of those things about -- from learning about what's going on in the industry out there, what ServiceNow is doing, probably are exploring out there and it looks like a good 15% of you, 16% are already on your GenAI journey out there. So congratulations on it. Really excited to also learn from you what you are doing. So if we get an opportunity, we'll do some kind of follow-up with everybody who's trying to have some use cases already deployed. So when it comes to ServiceNow, I had the opportunity to be -- I'm part of ServiceNow AD organization. But as I'm doing this webinar, I've been doing a lot of customer conversation in terms of how we are leveraging our platform, what we are doing with Generative AI. I've been with ServiceNow for more than 8 years now. And what I've been hearing about the C-suite priorities is, is the -- the technology landscape just overall has been changing a lot. So how do we manage that? How we're going to help employees create impact there? How can I improve the customer experiences out there? How can we make developers more productive? And then obviously, every C-suite member and the Board member out there are talking about what is the Generative AI agenda? What's your AI strategy? It is pretty -- like we are doing an AI day next week internally within ServiceNow where we're going to have like a 2-hour session about talking about AI, so we do that very frequently. So again, this is top of the mind for lot of C-suites out there. But what is more important to understand is the fundamental side is that our business objectives are the same. And what you mean is there are only 2 key business objectives. We want to grow the customers. We all like, it doesn't matter which industry you are, what -- how big or small your enterprise is, and you want to make your employees as productive as possible. And for each of these objectives, we feel like AI is one of the way. It's not the only way because AI has been there for a while and -- but it's one of the ways to get there, to make our -- to grow the customers and to help the employees to be more productive. But when it comes to when we are looking within ServiceNow, how AI is helping? We look at it from 3 different dimensions or a vector, we call it. How is the experience on the screen like how we are engaging with any technology is changing and with AI, how is going to improve. We, in fact, on that consumer life outside our world, AI has been there every day of life, like whether you use LinkedIn, Facebook, Instagram or any of the social sites or if you look any of the sites where you go to Amazon or something similar to that or you're buying anything online. There is already AI in that experience is because what you see on that is AI generated a lot. Speed is a critical part in Generative AI and is definitely helping here behind the scenes, how can it do things on your behalf or augmented and improve the -- just the overall speed of getting things done out there. And lastly is about decisions. Decisions is like because we -- as humans, we have to sit through so much of data, how AI can help you to take action? What is the next best action? What is the next thing I need to work on it? So helping we feel like AI is going to profoundly improve the decision aspects of that? With that, we've seen within ServiceNow, our vision is AI is going to be everywhere, and we're going to give that into seamlessly into our work, which will transform, as I said, any of those experiences, decision-making or making things faster. And we see -- we have already seen that every employee will have an AI-powered assistant, AI, we don't believe that it will replace humans for anytime soon, but it will provide the assistance to humans or employees to do things more efficiently and faster out there to help to do that work out there. So more from an Assist point of view. So that's our product. Also, as we call it Now Assist, and that is how we look at it, generally, not only in the product but also internally within ServiceNow. And I wanted to give this context to you all to set the stage. Before we dive into Now Assist, because I'm sure you have heard about Now Assist, if you're a ServiceNow customer, you might have already seen some demos, but more important, which we saw that a lot of our customers I was talking to, they wanted to just get an overview about their overall strategy and our approach towards the generative AI. So for ServiceNow internally, our AI journey, we have been on the journey for the last 6-plus years. And this is not like sequentially, you have to go through foundation and do real time of predictive, but that's what we experienced. We started by -- generally, it starts with data. Obviously, we know that AI without a data is not going to -- doesn't do anything and especially building machine language models in the past, and which we continue to do that. We have more than 100 ML models internally and especially in the areas of sales and finance, we started that. Then we brought in search and Virtual Agent. Search like how do you drive the relevancy, Virtual Agent using the natural language understanding all that data was also embedded not only internally but into our product also. Then we started looking at predicting. Some of the business events out there are which we are predicting in terms of our financials, our forecasting and things like that, our headcount forecasting. And then we started doing prescriptive AI, whereas we were providing the AI recommendation. So to our employees, whether it is for learning or the next action they need to take. And obviously, as generative AI came our way, we started embedding that into our workflows out there. So that's been our internal journey on the AI because this is, again, important for us to share that we just not just jumped into generative AI, not only from our internal use point of view and also from the product point of view. And when it came from a product point of view, we have been on the AI journey for like more than now almost 7 years, back in 2017 through some acquisitions, but mainly in terms of getting the talent on board. And then through those, we have been making AI available to you as a customer through our family releases back in Kingston and then onward from London, New York, Paris, Rome, San Diego, Tokyo and then in the time where we did was in last year 2023, you started seeing the generative AI coming in and in 2024, we have already shift them Washington and this [indiscernible] release is coming, and more and more, if you can see AI capability in the platform is growing over a period of time. So again, not new for our internal use. We have done, obviously, more than what our platform offers today. And then obviously, within the platform, there's already a lot of capability available from an AI point of view. So when Generative AI came, all of us, we saw ChatGPT, how it has absolutely changed our world in terms of you get to ask a question and you could get answers, but these were all in the public domain. And then they said like, okay, how do we bring this kind of experiences inside within internally with ServiceNow? We know that it is going to impact our business outcome. It's going to make things faster, and the quality of the answers are going to get improved better and better. We see that work will get automated. In some cases, we even -- may even get eliminated out there. And this is what we expected from a business outcome and the experience and the adoption of this technology and experiences are going to drastically improve over a period of time. So this is how we see that how generative AI is going to make an impact out there. But how do you bring generative AI into an enterprise now. This is the question everybody is, as you said, you have some of your exploring, some are evaluating, and most of the reason maybe because of the questions you may have, like what's going to be my cost? Is it going to actually give me value out there? What is my ethical considerations, right? We talk about responsibly AI, like how do I drive chain management? Do I have skills to deliver this capability out there for my stakeholders? How can I trust the output? Do I have the right data to get that. And then these questions, and I'm sure there are many more than this but this was the general questions we had when we started and as we were talking to some of our customers, this kind of became real that these are the questions we need to get answers as we embark on the journey with generative AI. So as we look at it from an approach point of view, we came with a clear strategy for us internally. We -- our goal was like, let's use in-platform or in-app generative AI capabilities. Now what that means is, yes, we offer generative AI capability, like Now Assist, we started shipping about 7, 8 months back or the earlier one was the GenAI controller back in last May at Knowledge '23. And then if you have been following, we have been continuously shipping a lot of that capability. But then there is other platforms, which are offering GenAI capabilities like Zoom has GenAI capability if you use Zoom. We use Microsoft, so the Microsoft and Copilot. So our goal is like let's look at all those capabilities, what will really drive the productivity of our employees, obviously, addressing all those concerns, we have that. So we are not expecting that there will be like one solution, which is going to work for everybody, every enterprise, but we need to come up with a clear strategy. We're going to use large language models, not only that we offer, but as an enterprise internally, we are using multiple large language models, and they are providing based on the use cases out there because we don't feel like there will be one general purpose large language, which can address your needs out there, at least not for us, it's doing. And internally, we are also, from a product point of view developing very domain-specific large language models targeted to certain use cases. Some of you are doing this where we are continuously learning, deploying it quickly. Iterating that was a whole idea. So initially, we came up with 90, 9-0 use cases. We created 4 x 4 metrics in terms of which is going to give value versus the effort and then how do we increase. We have close to about 40 generative AI use cases, which we have already built in and deployed internally. And we need to keep on investing in training, not only for our technical folks to go and build these use cases but overall in ServiceNow to educate that how AI -- we are leveraging, how do you look at it in terms of the output, how it's going to do address that concerns employees have that? So from an approach point of view, we came up with this approach. And then we wanted to address this key thing, which I know everybody has this question. And obviously, our employees were also there. Like what -- how are we going to make sure that we are responsibly using AI and then how we are going to win the trust of our employees and also of our customers. We want to make it human centric, means you will decide when you use it. So it's not going to be just we're going to make actions internally make it available, and the employees are not aware of it. We're going to make that -- make it transparent and what we are doing out there. We're going to make -- we have our own AI governance, policies and governance. We have put together internally. So we have also gone through an audit and we feel pretty good about it where we are with security, privacy, legal or the compliance. And then we want to make sure that we are -- if we are giving training on any of the models, which we are testing and reflecting the diversity of our customers or our employees internally from that point of view. So again, these are the principles we are working on and to make sure that we are responsible AI factors are played in. Now when it comes to -- as I mentioned that we are looking at as a multiple GenAI providers out there. If you look at, there are hyperscalers like Azure and Google and Amazon, which are providing you those [indiscernible] large language model, we are also using that because if you look in the middle of that, that is the key. Everybody is concerned about our questioning. We look at it for ourselves because the Now platform domain-specific. So that's more specific to us internal. And then we are looking, as I said, these other SaaS providers. And so in your case, ServiceNow is one of them. But if you look at other SaaS providers, what kind of LLM or large language models they are providing is part in the platform because they have the technology, they understand the data model. They are best suited for us to do that. But what we have done is what we call internally an AI control tower, which is looking at all the models in terms of making sure that we are passing all the reviews, how they are performing, all the activity in terms of the model, be it performance and the accuracy point of view, we are tracking it and managing that through our AI control tower. It's something new, which we have gone up for our internal use. And obviously, these models are getting consumed in any of the experiences we are internally building through any of the UI stuff, whether it is through our portals, to our search and through other workspaces out there. And obviously, lot of data. Again, we are a very difficult enterprise, like I think so many of you, we use a lot of third-party applications to run our business out there. So this is our like high level, this is not like all the details behind. There are a lot of more details behind it. And if you are interested, we can definitely catch up on that one. But when it comes to our platform because if you -- ServiceNow platform itself, this has a lot of capability in that. So our portfolio is built on our ServiceNow platform, whether they are technology workflow, I think so most of you might be using those ITSM and asset management and IT operations management type of products out there. We have the employee workflow, customer workflow, the Creator workflow with AppEngine, and now we have been on the journey to expand our products into finance and supply chain. And then we have industry-specific workflows in the platform out there. But that is all built on one ServiceNow platform, which has one data model, one architecture, running in cloud and AI is embedded in the platform. It doesn't sit outside that. So if you look at it, the AI capability that I talked about a few minutes back, what is in the platform, whether it is search, whether it is Virtual Agent conversation, and there are like 7 different AI capabilities, and we are adding more and more platforms. So this helps me internally to expand the use of our platform, not only with the portfolio, which is available, but also with the technology, which is available to me on the platform to go build more use cases using AppEngine out there. And with generative AI, here are the set of products, which are obviously available, which we are shipping as a product, and we are using internally, but we are not limited to this. We have the GenAI controller, which you can use it to connect to any large language model. And that's what we see ourselves to be transitioning into it. When you take large -- the ServiceNow platform, you take the large language model, which is our Now LLM and then with the Now Assist capability, these see ourselves like now getting this ServiceNow platform getting into be able to provide you this AI platform for your business transformation. And this is one of the platforms, not going to be the only platform but you need to figure out how you're going to leverage that end scale. Now when it comes to large language model. I know this is a key component in terms of generative AI conversations. We see, and this is also internally we use not only our large language models and now like LLM but also we are using something like Azure OpenAI, a private instance of that for our use cases, which are with the data which doesn't reside in our platform. But the wide matters is because the domain-specific models, which we are using internally and we are also making to our customers, it just helps to make it more accurate. The Now LLM sits next to your instance so that your data is not leaving the data center, a lot of -- all of us have that concern that we are sending data outside, ServiceNow know resides there. So it was very easy and fast to get our privacy, legal and our compliance group to sign off on our use of our platform with large language model next to it rather than sending it out there. And then obviously, when it's close by, then all the other use cases, it can keep supporting them. And from Now Assist point of view, as I said, we have a bunch of products, which we have built, which we are using internally. Generally, it is helping us from a summarization point of view. We see like conversational exchanges are very key out there, creating content and then convert text to any kind of generation in terms of code or anything like that, and I'll share some of the use cases, which have been done. What we see that Now Assist, which is already driven has increased our productivity, it is changing our experiences and then making things faster for us, and we'll see that in a short period of time. So when we started our journey in middle of last year, first use case, we deployed like around October time frame, which was fully in production. And since in the first like about 120 days, we already had seen the benefits and these are tangible benefits equivalent to about $10 million worth of benefits annualized. We see that, which is enhancing the productivity equivalent to 50 full-time employees. Whether it is on the employee side, driving more self-service as the employees are getting knowledge articles summarized out there. So they are using that. We have already seen our -- about 14% increase in our incident case restriction means we have less number of incidents are created by our employees because of the results they are seeing when they are asking a question. Same thing on our customer support, like, say, yourself, all of you if you has -- who are customers on ServiceNow, and you are familiar with our support portal, when you come to our support side, if you're searching anything, you will probably see a generative AI answer there or you're going to open a case related to your instance, then you will probably get more related to that, you will get generative AI response as you are trying to open the case based on what you're describing the problem. And we have already seen 10% boost in the deflection rate out there, and you can translate that in terms of what kind of savings, and then also on the developer side, we have already seen productivity increase in terms from a developer point of view. And this was just starting points for us. And this is -- again, this is like back in February when we did this calculation. But since then, another 4, 5 months have gone by, and we have already seen more and more use cases getting deployed, and each use cases, obviously, we are driving the adoption. We are driving the usage of it, the value of it. As we stand today, you can see by different personas within ServiceNow, what kind of use cases we have deployed. So if you look at whether we have like our IT support agents or our customer support agents, which are helping you or our support cases, which are coming from employees, or you as a customer or internal employees or our product managers or our developers in IT, these are all the use cases, which are live today. And as I told you about these set of use cases just around the platform. There are some other use cases, which are outside our platform. But what I wanted to share with you that we were able to rapidly build and deploy using different large language model, obviously, our own where it came from a product with our generative AI Controller, we were able to connect to Azure OpenAI, and we were able to deploy these use cases. We have more coming, and I'll share with you a little bit more about what are the next set of use cases, not only from the product which we are deploying internally being Customer Zero, but also what we are looking to internally building the teams are building out there for our [indiscernible] use out there. So really excited about it. And a lot of hard work has gone in for the last one year in this particular space internally in the ServiceNow for our own internal use, working closely with our product team, and you can already see that we were very quickly able to pivot from just exploring and learning to actually -- and evaluating but to actually deploy those use cases because that's where you start seeing whether it is going to give you the value and the ROI, which we feel like for us, we are already on that journey in that regard. So I'm going to share now some couple of like demos out there. So it will give you -- I know you saw a bunch of slides. Generally, when I meet with customers, I do 90% of the time, I just show them real stuff. But today, I'm going to show you some recorded video. So first one is, as I talked about, one of the #1 use cases where how it is helping with self-service where employees are asking for questions and getting answers using Now Assist. So it was an extension of search. So as people -- as we all have seen, [ Zanca ] the Bing of worlds and the Googles of the world, when you ask a question, now when you get the responses not only you get the link but you also get generative AI response that very similar to that, we have made it available both as a product and internally. So in this particular area, you will see that as a user asking a question about [indiscernible], how do you schedule an upgrade to their instance instead of going and picking a knowledge article, which already is available to you. That's a source and that's the reason that we talked about building the trust with our users that we always provide the source of the answer. So they get demo step by step, how do I schedule my instance and pretty quickly, they can move on from that. So that's just one of the use cases of many which I talked about, which is deployed internally. The next one, this one is when we talked about deflection when an employee is going to be trying to open a case or like a ticket in IT and the user will get a response back before they actually submit the incident so that they can look at the response and then they can decide whether -- they still want to open the ticket or they did get the answers so that, that's again a big, as I said, the boost in terms of our deflection rate. This is a recorded video with some voiceover. So please, for a minute, it will run through that. [Presentation]

Okay. So what you saw, again, there was a question related to how do [indiscernible] a question about the MAC, give an the answer, they can provide us a feedback. So again, what we have embedded everywhere where we are providing them generative AI-related answers or responses, it's already, I would call it watermark, so we will know that it's AI response. They can provide us a feedback. And those are the critical data points for us as we are monitoring what kind of responses users are seeing and what kind of feedback they are getting. And I think I mentioned to you, like we are -- we are interested in the key part for us is 5 key data points we are trying to track here for every use cases because use cases are not like just for the sake of building it. We want to see the adoption of the use cases and are people using it, who they're intended for, like whether they are agents or anybody. And not only they're using the adoption, but what is the actual usage, how many times they have used. Like, say, for an example, how many times is something that was summarized or something was there? Like are they actually using not only using it but also adopting it and then they get their feedback sentiment, are they liking the responses or the overall experience they have made. And then the value. Value is critical because -- in the end, it's all about like is there a productivity gain from an end-user point of view. So that is also we are measuring in terms of -- from a metrics point of view. And the fifth point is the -- what we call model accuracy on the exemptions race like did I -- when I gave you a response, you gave me the feedback, did like it or not, accepted or not, things like that. So those are the 5 key data points for all our use cases we are tracking because that's how we can justify all the investments we are making internally for ourselves out there. And the other use case now, we talked about mainly from how we are self-service side of it. Now when you look at the -- like our IT agents, customer support agents or HR support agents out there. As you know, ServiceNow platform, it's all about service management and fulfillers are helping our end users. So how we are empowering that agents now through summarizations, through generating resolution notes or generating, they can generate knowledge articles, they can do alert summarization. So this is a very simple use case, a small video here you'll see about summarization. This will very quickly go by. So I just want you to realize that, that these are not very long videos out there. So you may -- if you blink, you may miss something in that regard. So here is a very quick video about how an agent can summarize kind of all the work notes related to that particular incident out there. So as I said, if you blink, this video will get over very quickly, and that's how fast it works too. So that's the beauty. And then once they're summarizing it, they can now generate resolution notes, as I said, about resolution notes. Again, a small video here to get through once they have summarized it, they can quickly watch that and go to the next one out there. And as I said, when they summarize it, they can see not only what were the issue, which was reported, but what were the actions based on? Because if the agent is recording all the actions that have taken through the work notes, now it's available in a summary so that anybody who's picking up that incident, if it's getting reassigned can see what actions already have been taken rather than sometimes an agent [indiscernible] the user that, hey, try to do this, and the users say, Okay, I've already done this. So why are you asking me? So instead of going through the work notes and trying to figure out what was done or not, now they can very quickly do that. So again, I gave you a very small snippet of what an experience are for the agent out there. And then the other persona, as I said, I shared with you that we had like 8 persons and I'm going through some of the demo and we can definitely have a follow-up. You can reach out to your account rep and say, hey, we want your Now and Now team to share with us all the use cases you have built and we have all set of demos available about how we are using internally. This area, I'm very much excited because I run a developer experience team within our Emerging Technology Group for all developers inside ServiceNow, which are building on ServiceNow platform. We have worked in BD, we have about 400 developers out there. And from a developer point of view, one of the product-related use cases, which is available is to generate workflows out there. So now very simply, you can write in text, which is generative AI world is called prompt, you're writing a prompt. In this particular case, I want to create a flow that runs every day at whatever time it is, and it looks at a newly created incidents and then you -- like how do you get it assigned to the level 1 [indiscernible] group. Again, in a very simple English and our belief is that the English will become or your native language will become your instead of Java or Python will become one of the most prominent languages to develop code in. So now with building with Now Assist, again, this -- you start with that. And it will generate you a shell of your workflow and then you go and complete the workflow out there. So these are some of the use cases for -- from a demo point of view, which you can definitely experience yourselves or put it in your [indiscernible] out there. Now with that, I'm going to now share with you that we have already -- as I said, we are very focused on ROI, what is Now Assist is giving us as such at ServiceNow, whether it is for IT, HR, customer or for employees. But what you see on this slide is not only the use cases we are already live but which are already -- some of them are -- as I said, the ones which are highlighted in green is using our generative AI Controller, and these are some customer use cases, which we have built for our internal use. And then there are certain ones, which are -- we are already in the process of testing it, and they are going to be coming soon, and we will be deploying it internally. So we are absolutely excited in terms of driving this agenda internally for our ServiceNow employees. And obviously, as we do it, it shows up in a product at a certain point of time. And from where we stand with, if you look at it, we are very excited about generative AI, we are all right when we see outside world but for an enterprise internally, we feel it's a game changer for ServiceNow, and it is really happening fast. You can see the use cases, which I just shared with you how that is moving quickly out there. Generative AI, we feel that I'm sure you all have also experienced that if you are evaluating, you can, in fact, have seen people creating their own bots using ChatGPT but in an enterprise world, it doesn't really work because of data privacy, legal, compliance issue. People still ignore it, which is good, but if you're putting a risk there, your business at risk by doing that. But we see that we can quickly deploy GenAI with not much significant technical capabilities or complexities you need, at least specifically from a platform point of view. Now when you go to -- that is with our ServiceNow platform with our own large language model and Now Assist -- within like, I would say, I have heard customers doing it within like a week's time, especially some of these basic use cases like summarization, generating resolution notes or generating knowledge articles, they are pretty, pretty easy to deploy it if you have not customized your environment a lot with very easy, you can do that in the ServiceNow platform. And for us, as I mentioned, we have already realized a good tangible benefit and which is improving the productivity of our employees, and we just are getting started. We feel like we are just getting started here, and we have a great opportunity in front of us to make an impact for our employees and for our customers. And then going back to our key objectives, are we able to grow our customers, and are we able to make our employees productive? And yes and yes to both of them. And we see that AI is going to play a critical role in that. And hopefully, you have seen some of those examples here, some of the use cases we have deployed. I wish I had the time to show you all the demos of all the 30 plus use cases we have out there but we can definitely chat about it at some point of time. But at this point of time, I want to now not only thank you, but open up for questions out there and try to answer. We have about less than like 12 minutes to get through questions. So we wanted to give that time. We know there will be a lot of that. So let's -- as we are trying to put AI to work for everybody out there within ServiceNow, hopefully, you'll join us in that journey. And with that, Mike, let's look at what kind of questions we have got from our audience today.

Absolutely. And there are quite a few. So we'll try to get through as many of these as we can. There's a question one of our attendees is asking how -- or can we integrate OpenAI or Now LLM outside of ServiceNow content? So can we get that content from outside?

So if I understand the question properly, that Now LLM, it is kind of -- I would say, it's more like a closed box because we are managing that, and we are abstracting a lot of that thing [indiscernible]. But if you want to connect from ServiceNow instance to like an Azure OpenAI, which we have done, as I saw some of the use cases. I've talked about it. We have connected it to Azure OpenAI. We are using the RAG architecture, if you are familiar with the retrieval argument generation aspect of it, and that's what we have used it to that. So you can connect to any large language model, at least a couple of them I know for sure. And what my understanding from a product point of view, sometime in the second half, which is not very far away of this year, you will be able to bring your own large language model because we know that all the data you need to get the answers or make the part of your generation aspect of it doesn't reside in ServiceNow. So we want to make sure it is open and be able to use that for sure.

Okay. Great. When we were talking about developers, are we referring to developers on the Now Platform or any coders out there using technologies like Java and platform for analysis?

When I was talking about today, I was talking more in the context about ServiceNow developers. We have internally within ServiceNow IT building on our platform. That was the mainly -- that is what we're driving like text to code generation or you saw the work flow generation we are doing. We are doing story generation and things like that, which can be technically are not generating code, but we have code explanation, we have code refactoring. So things like that, which is very specific to our things. But this can be -- we can apply -- start applying this kind of things what we have done on other technology stack pretty easily. But today, our goal is more -- we're more focused on our platform.

Right. Okay. Here's one. Did we train the Now Assist on all versions of the product documentation with KVs, community questions, answers where are we training our Now Assist?

So if you look at it, so when we talk about large language model and training the large language model, obviously, we are using the data we have access to. And if we have customers who have opt in and we are using those data sets to train our models. Now for the answers you probably saw on our support side, the case support where it uses -- it is we're bringing data from our product documentation today to get your answers in probably in the second half of the year, we are creating this what we CX an experience, which we are building. If you have not heard about that. I'm happy to share more on that, that one customer experience. That's again, our experience for our customers like you all. When you come to our support side, you go to our learning side, you go to our community, whether you go to our developer portal, all that will be all brought together. And then the answers when you ask a question will be generated based on the content, which is available there. So that's pretty something coming for you all the use of when you are kind of engaging the ServiceNow whether as I said, for support or community, that's something experience we are building for ourselves, for our customers for that reason.

Okay. Let's move to kind of Virtual Agent, MS Teams type of question. So there's 2 questions here. We can keep them kind of separate. Do you see Now Assist replacing Virtual Agent? Or is it augmentation? What are your views on that?

Yes. So Now Assist -- see, Now Assist is -- there are parts of, which is embedded, as I showed you, like in the search you saw Now Assist that is giving a generative AI response. You saw Now Assist in terms of our agent where they can now summarize or generate resolution. So our thinking what Now Assist is, as I said, it's an AI-powered assistant users wherever they are in the platform, and we are looking at it, making more broader and broader available across wherever our employees are on the platform. Now Virtual Agent is -- there's Now Assist within Virtual Agent, which we have already made available like how we have for search, then you see generative AI response. So Now Assist in Virtual Agent. That is the next generation of Virtual Agent. It's not like replacing Virtual Agent but it's going to make it much better. It's going to make it much more intelligent, it is going to make it more conversational because when we start using generative AI, the intent matching will get better. So instead of NLU we are going to use AI. So there's a lot of technology behind it. And by the way, that is already released to customers who are using generative AI and we have already deployed internally, one of the places where Now Assist is part of Virtual Agent. And one of the key areas where we have seen a huge benefit is, if you are familiar with Virtual Agent, you create topic, conversations and like to build topic, I know we have done, like we have 300 conversations. We have built internally over the last 5 years into our journey. But now you don't have to -- like there were a lot of catalogs, which we kind of replicated these topics, like, hey, I want to order a software. Now a software, you need 4 questions, the user to answer. And that also as a catalog, and that is also a topic in Virtual Agent. Now what generative AI has given us the capability that we can take that catalog and make it conversational. So that's pretty good, I would say. And that takes away the need to create topics in your VA. So that is a huge lift. So Now Assist is not going to replace VA but it's just going to make it -- it has already made it better and it's going to keep on making it better and better.

I would agree. As a user, I've seen it, I witnessed it. I've experienced it at ServiceNow and it's great. When you see that, you're like, oh, all right, let's check this out, let's see what's going on here. So as a user, I really like it. There was another question that was semi-related is, Now Assist integrated with MS teams? How do you see that functioning with one another?

So Virtual Agent is connected to Microsoft Teams that has been there for a period of time. So it's been there for a while, where you are a virtual agent, which generally people see it on the portal, it's also available in Teams, and we use it internally. In fact, our usage has gone up significantly out there. So that's one part. And the second part of that, integrating teams, what we have done is, we have also -- that's a way to proactively engage users because that is where you can send some notifications to the users through the Virtual Agent on Teams. Now Teams and Now Assist, there is -- when we talk about Now Assist like in ServiceNow's world, there's a Copilot in Teams. And I know there's a lot of like all these different copilots and assist and bots are going to -- are coming up. So we have acknowledged 24, we already announced and we demonstrated the integration between Copilot and Now Assist. So when you have both of them, how do you -- how does it go? So that has already been demonstrated at Knowledge 24. I think so from a product point of view, it is going to be made available sometime in August, September time frame, just don't quote me, think about Safe Harbor but any time early second half of the year, we'll be shipping that from a customer point of view. Internally, we are looking forward to it because we also have Copilot, Microsoft Copilot. We have internally activated for our employees, and I'm definitely looking forward for the integration with Now Assist so that if you are a user who was using Copilot, it will get responses using Now Assist from the ServiceNow platform. So you don't have to go and figure out, which one to use out there. And we want to make it easy. And I'm sure our product team is making it easy for our customers. We want to make it easy for our internal employees to get it.

We have a lot of questions, and I know we're not going to have time to get to all of them right now. So I'm going through and kind of looking at ones that I think will be useful for everyone. So we have a lot of people that are exploring, right? So as people are looking to implement GenAI, it just you guys are asking so many questions. Where did it go? Well, I just talked about how -- what do we need to do to prepare our data for GenAI? Is there anything that we need to do as customers to help prepare our data as we go along this journey?

Yes. Because se generative AI and we know it's about constructive data. So first of all, there are 2 sides of generative AI. One is it takes the data and summarizes for you. So for that, it needs good data, it will obviously give you good summarization out there. So generally speaking, we are talking about knowledge article in ServiceNow today, so if you have good knowledge article, then your summarization will be good. And we're -- we found it really, really interesting outputs we are getting when we are summarizing it. And then that you can already see the numbers from a deflection point of view or a search success rate and everything has gone up because now employees are getting more and more engage because of getting the results out there. So data is an important part of you, especially when you are summarizing something. You don't want to -- If your knowledge article is not good, then obviously, summary is not going to be good. And generative AI doesn't really can do much about it when you don't have good data out there. Now when it comes to from a generation point of view, then it is very helpful. So that is where I strongly have been suggesting to some of the customers who will start to improve them, they really want to take one step ahead. I say start with summarization. Because for summarization, you just need work notes. And if you're an agent may not be doing a good job of like operating work notes to begin with but you can train that thing, hey, let's go forward -- going forward, every incident you are working through, generate your work notes, enter that and then summarization will show up better, and then you have resolution notes will show up better out of that. And so when your cases or incidents are resolved, so it will very quickly. So you can do case-by-case basis literally in that particular case and you can get started. So you will have to invest time to go clean up some of the things which have happened historically out there. And once you are trying to resolve their cases, you can generate knowledge article from there. And as the knowledge article gets generated from a good set of work notes and resolution notes, it generates good knowledge article, and that will feedback into, for example, back into your search and deflection out area. Now that's one side of the coin. The other side of the coin is the developer side of it. Developer side, mainly it's about generation, text-to-code generation. So if you have developers in your organization or yourselves, if you are building codes in our ServiceNow platform, it does a pretty good job in terms of generating ServiceNow code out there, because there's no input, we have trained the models. And that's the beauty of using our platform with our large language model because it is very domain-specific. We have trained it for generating ServiceNow code out there. So that's the benefit you see.

Okay. We're going to probably wrap it up with one last question here, and I'm going to combine. Some people are asking about how they can learn more through training. So -- and maybe also kind of combined with that, how important do you see prompt engineering training? Is that something that in prompt engineering or just training for GenAI, what are your thoughts on that? Like what's available to us? And what do you think?

Yes. So there is a lot of content available for our learning team, which is on Now Learning, our training -- learning and training development team for our customers have built and published those trainings on Now Learning [indiscernible] on Now Learning and look it up and take those trainings, especially around Now Assist and everything. Prompt engineering related on our platform, we have not opened up yet to anybody to prompt engineering platform. That, my understanding is also coming soon pretty much where the users will be able to do a little bit of prompt engineering. And our whole idea is that if you are a platform developer, you should be able to build generative AI solutions using your platform skills out there. So that's the whole thought process around it. But there is a lot of training available on prompt engineering. Otherwise, if you are interested, like my team is also doing it outside our platform, and there is a lot of training available on LinkedIn Learning, Udemy, YouTube, and all those places out there. But as I said, specific to our platform, there's a good amount of training available on Now Learning, and I know, Mike, you said there are a lot of questions that I was also looking through the questions, I was trying to answer some of them here. So again, I know I see that like close to more than 50 questions in there, which are visible, so we'll do -- I apologize, we cannot answer all of them right now. So what we'll do, we'll reach out to you with specific answers if there's anything we need to follow up, absolutely. One key aspect is that if you're ServiceNow customer and if you're interested about what we talked about today from a Now and Now point of view, what are the use cases? What we are doing and exploring? What is the architecture, everything internally into ServiceNow? Please connect with your account team, ask them, hey, you just watched this webinar and on this webinar, you're interested to have a follow-up conversation with the Now and Now team, you can use my name, you can say [indiscernible] talk to Rajeev and then we'll figure out a way to. And I do think that's outside my job #1 to deploy these things with my team but I also speak with our customers in one-on-one cases if needed or somebody might from my team, or within ServiceNow, we'll be more than happy to connect with you and try to do a deep-dive into the use cases, if they're exploring or a platform or noises or anything related to it and we'll be happy to share the knowledge. And hopefully, we'll also learn from you based on your questions and interest you have that. And in addition to this, there are a lot of other on-demand webinars, which are available to you. So please see the URL on the screen you see there. Please copy that URL and then you can not only this webinar will be available as is getting recorded, but there will be other webinars also in that. And with that, from me and Mike, thank you very much for being with us for the last one hour to where hopefully, this was something interesting where you learned a little bit about our internal journey in ServiceNow, how we are using Now Assist and using generative AI to drive the product DB for our employees. So thank you very much.