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. [Operator Instructions] The ON24 room allows you to adjust and resize all panels that appear on your screen. Simply click and drag the top of each window to adjust. And if you like, you can also maximize panels or view in full screen by clicking on the top of the window. Closed captioning can be enabled or disabled by clicking on the icon at the bottom of the media player. [Operator Instructions] We'll take as many questions as we have time for. And if we happen to run out of time, we'll circle back with you after this session with an answer. We also invite you to check out the items available to you in the Related Content panel at the bottom of your screen. We appreciate your real-time feedback during the webinar using Reactions found in the bottom-right corner of the toolbar. In addition to submitting your questions via the Q&A window, chat directly with one another and presenters in real time using Attendee Chat. If you experience any technical difficulties during today's event, please enter a question into the Q&A panel, and we'll be happy to help resolve any issues you may be experiencing. One more thing, your feedback is extremely valuable to us. So please fill out the survey by clicking the icon that looks like a clipboard below. The survey will appear at the end of the webinar. Finally, thanks again for joining us. We hope you enjoy today's webinar. Now let's get started.

Hello, and welcome to ServiceNow's webinar. And today, we're going to show you how you can fast-track the delivery of your applications on the ServiceNow platform with low-code powered by generative AI. I know this is a topic that many of you are interested in. I don't think any of us can turn around or read an article online or anything right now without talking about generative AI. So today, we're going to kind of flush it out and get into a lot more detail that's going to help you with your strategy. So with that, I need to just kind of flash our safe harbor notice here. We're going to talk about some stuff that may be coming out in the platform in the future. ServiceNow is a publicly traded company here. So please do not buy any stock based on anything Dale or I say here today. With that, let's jump on into it. So again, I'm Gregg Aldana, and I lead ServiceNow's Global Creator Workflow Solution Consulting Organization for ServiceNow. I've been with ServiceNow for about close to 8 years, and I've been in the application development space for well over 25 years. Very few technologies that I have not either personally developed software on or managed large teams to. But really, over the past 10 years, I started out as a ServiceNow customer in the U.S. federal government, working for the FDIC and developing workflows during the last global financial crisis. And over the past 7.5 years with ServiceNow, I think I'm in a very great position because I get to work with some of the most talented people at ServiceNow. And I meet with about 200-plus customers, CIOs and business leaders in every country and every industry a year. And today, we're really going to share with you some of that insight. Now I want to allow my co-presenter today, Dale Stubblefield. Dale, do you want to introduce yourself?

Sure. Thanks, Gregg. Everyone, I'm Dale Stubblefield. I'm a Senior Advisory Solution Consultant here at ServiceNow, focused on Creator Workflows, which is a lot to do with taking the power of the platform and building custom apps, integrations and intelligent automations. I've been in technology for about 20 years and going on half of that in the ServiceNow ecosystem, starting as a customer, then a partner and an implementer. And then finally, I joined Gregg on the Creator Workflow side of the house here at ServiceNow. And nowadays, you're right, Gregg, you can't trip over the news that hearing another article about GenAI. So I'm looking forward to the presentation today.

Absolutely. Well -- and hopefully, we can separate out some of the myths from some -- from the actual reality and show you some of the actionable things you're going to be able to start using today and kind of really hope you lay a foundation for the future. So let's jump right into it -- to this here. I think, first of all, before we kind of get into some of the details, I think the one thing that we can all agree on is that every company today has to do several things when you're on the ServiceNow platform. No matter what product set you own, whether you're using our app engine, whether you're using our prepackaged workflows for HR service delivery or customer service management or obviously, our core product of IT service management, every company out there has to build scripts to implement very specific business logic into the way that these products actually operate. Second, we notice that our customers, every single one of them, regardless of the product you own, have to generate specific business workflows, processes such as [ tasks ] and approval of how they get assigned as well as how business processes, like tasks and approvals, get routed to when and to whom and to how. The other thing that we're noticing, all of our customers have to build out business playbooks that really map out how specific business processes should operate, like investigations or onboarding new employees or onboarding new customers or vendors. Even incident escalation procedures have to be mapped out in a very detailed playbook. Next, we noticed that all of our customers have to maintain and update catalogs in different business departments that contain all of the assets and all of the services that each department can provide and offer to other parts of the business as well as external parts of their business. And then finally, bringing all of these items together, scripts and flows and playbooks and catalogs and all the various UIs inside of apps. And so today, we're going to show you how you can actually drive all of these capabilities in an accelerated manner with generative AI. So I think every company out there that we speak to, every CIO that I speak to, and I'll probably speak to about 3 or 4 a week, has to come up and is expected to have an AI strategy. And ServiceNow can really help you be a first mover in this area. I think AI is poised to really unleash that next wave of productivity in the business landscape right now. I mean we saw this early on with low-code and different types of tools and cloud-based software really driving productivity. AI is really going to unleash that next wave. And so generative AI is really just the latest manifestation of this across the entire development life cycle. Now according to McKinsey, generative AI's impact on productivity could add trillions of dollars of value to the global economy. So this is something that every CIO and every company has to take very, very seriously. And this is going to impact -- the impact of generative AI and productivity is going to be felt across every industry and across every department in every role. This isn't just for IT. This isn't just for technology companies. This is for every industry, manufacturing, financial services, public sector, health care. And every department is going to be taking the advantage of this, whether it's finance, whether it's IT, whether it's procurement, whether it's supply chain in every role. And so getting an AI strategy right means that you can unleash these productivity improvements and really fundamentally transform the way that your employees and your customers interact with everyone and experience your brand. And really, ultimately, the criticality of this, is this is going to drive the agility that you have as a business to drive specific business outcomes and really stay ahead of your competitors. But with everything that we're hearing and we're seeing out there, it can be really hard to separate out the hype from the reality. And most of all, it's really hard to know who to trust and who to listen to. It seems like every time you turn around, there's another software company or another consulting group trying to sell you on something or their paradigm of their area. So in today's presentation, we want to offer you our perspective as a company on why marrying AI with the workflow automation platform, like ServiceNow, is a powerful combination that's really going to uniquely help you to drive all of your business transformation goals. So if you think about it, generative AI by itself doesn't really do much. I mean it can -- it's not really going to help you transform much in and of itself. I think everyone has used generative AI in some form or another. You've gone to a ChatGPT prompt, you've typed in some questions, and you've gotten back some answers. But to really transform the enterprise, you need an intelligent platform, like ServiceNow. So for everyone who's used generative AI, look, it's really adept at organizing unstructured data, taking articles and Internet pages and code samples and chat conversations and then transforming it into relevant new information. But you need to take action on the results of generative AI. So how do you do that in an enterprise context? Well, the answer is through workflow automation. And using an intelligent platform like ServiceNow, we can help you take the compelling output of generative AI and connect it with people and processes and systems in your organization to take action and to structure very specific workflows that meet all types of business imperatives, such as compliance, precision and predictability. So the opportunity that we're going to discuss today is really to narrow the power of unstructured data that can generate with -- that generative AI can organize for you with the structured capabilities of ServiceNow's workflow engines and harness it as an intelligent platform that can help you as a company and as a customer take action on all this generative AI content that your organization is producing. So large language learning models, it seems to be the talk of the town. Every time you turn around, you hear another company talking about their LLMs, and this is the LLM you should use. Well, at ServiceNow, we kind of have a really unique approach to this. We really have a dual strategy for delivering the best GenAI capabilities for the task at hand. So LLMs are not new to us. As a company, we've been using them for years. But the ServiceNow platform really provides you with a high degree of flexibility in this area. So there are definitely general-purpose large language models out there, OpenAI, Google Cloud, Microsoft Azure. And really, some of the key attributes of these types of LLMs is really when you are looking for broad intention, you need large -- very large pools of data sets so that you can respond to any topic, whether -- you can type it in any language. And you're not doing something that's very domain-specific. On the other hand, when you're trying to achieve something in a certain area where you need very first-hand, very targeted knowledge, domain-specific large language models that have been trained on very, very subsets of siloed data can be much more accurate, provide you with information that's much more relevant. It's optimized for your data. It's really anchored in the best practices of a particular area. And you can actually be safe that the data is retained within ServiceNow. So for example, we have developed large language models in concert with NVIDIA and Hugging Face around coding and producing workflows because those are things that are very domain-specific. If you want to actually produce code snippets and produce workflows on ServiceNow, that's a very domain-specific area where you're going to want to have the best practices and you want to train it on our coding best practices. For a general-purpose model, you might get -- it's trained on all the code that's ever been written out there. So you might get some very bad code examples to follow out there. But a general-purpose large language learning model might be much more relevant when you're trying to just build an app from scratch or you just want to look at what is the best practices around a specific business process, like doing a fraud investigation process. There, you want to look outside a specific domain and really see what the best practices are out there in the totality. With ServiceNow, you don't have to make a choice between these. You can take advantage, and you will be able to take advantage of both language learning models for the applicable situation. So at ServiceNow, we're really focusing on empowering 2 groups of users with our generative AI approach. First, the Creators. And really, this is the area we're going to focus on today in today's webinar. So these are the people in your organization that are building scripts and building workflows and playbooks and catalogs and apps. And generative AI will enhance both their productivity to build these apps in a faster manner, but also using our large language models, building better quality and higher-performing apps with better business processes and workflows with less errors. And you're going to be able to fast-track the delivery of your apps using our generative AI technology. But once those apps have been built and couple that with the prepackaged workflows that are on the ServiceNow platform, you'll be able to use this and these apps to boost the worker productivity and help your workers make better business decisions. So let's dig into how we can make creators more productive and build better quality apps. The way that you can do this on the ServiceNow platform is with Now Assist for Creator. This is how we're going to enable creators and developers within your organization to build and figure applications faster and make it much easier for all of the developers within your organization. The first area that -- and the first way that Now Assist actually achieves this is with something called code generation or affectionately known as text-to-code. I think everyone's heard of that. And for those aren't -- who aren't familiar with that, that's where we take natural language, just describing the business logic or the requirements of what you're trying to achieve. And then the ServiceNow platform with Now Assist for Creator will automatically generate the specific scripts and the actual syntax and the code in ServiceNow automatically. This capability also helps you autocomplete the scripts that you are creating with a code-complete feature, which we're going to demonstrate today. And we're seeing this, in most cases, help accelerate workflow implementation and the maintenance of scripts by up to 30% in some instances. So you may be asking yourself, "Well, Gregg, what's an example of this? What are the types of things that I can do with text-to-code that it will help me do it faster?" I think a great example is where you maybe have to maintain a catalog and produce a new catalog item. And then you want to assign that catalog item to a specific group based on a delivery plan test that has been assigned to someone. Say that you're trying to count the list of attachments that are attached to a specific incident record or maybe you want to change the incident priority or change it to one critical or create a new change request to escalate something if it hasn't been resolved in a certain amount of time. There are a myriad of things that developers and people that have to write scripts have to constantly remember: the syntax to loop through a set of variables until you're done whether you're trying to raise warning messages when you don't get the right type of data that you're looking for. Say you're querying large amounts of data from a specific table and you want to limit the number of records that are returned when you're querying these or set the limits or you want to avoid infinite loops that might take down a server. Normally, developers or creators would have to go to a code library or maybe go to the help and look up the syntax or go to a Team's form or an online community to look this up. With text-to-code, they can do this in a matter of seconds, and they can be much more productive. And they won't have to worry about creating the right code or remembering the right syntax. And so it's a much cleaner way to actually develop scripts and business logic within your organization. The second area and the second way that Now Assist for Creator actually helps is with something called flow generation. And so really, this is very similar to text-to-code, where we're taking in natural language and just based on whatever the business and the user requirements are that you type out, it will generate a comprehensive workflow and workflow -- and Flow Designer based on what you had described initially and will also make recommendations for the next steps that are typically seen in very similar workflows. So this is important because it's really lowering the barrier of entry for people in becoming workflow designers and being able to create new workflows. And you can generate these workflows with fewer errors simply by just using natural language. So you might be asking yourself, what is a good example of this? Well, say, you've got to create a business workflow to create some new users and then assign them to specific groups based on their profile. Say that you want to create a business workflow to approve time-off requests. And you could just type in natural language every day at 3 a.m. I want you to approve any time-off request that came in that's less than 7 days because that's the company policy. But if it's longer than 7 days, I want you to escalate it to a manager for approval. And say, can you raise a Team's event and send an e-mail to both the requestor and manager of letting them know about this? You could do that in a matter of seconds with text-to-flow. But you can also use this when you're building workflows within any of our products. So say you -- in HR service delivery, you want to have every employee enter their goals and -- their HR goals by a certain date and then route them to managers for approval. Say you own CSM and you want to query social media posts for key phrases and then create an incident when you detect certain words about unhappy customers with a product. You can automatically create these workflows extending our products in a matter of seconds with flow generation. Now I think we want to see this in action. So I'm going to turn things over to Mr. Dale Stubblefield, who's actually going to take us through a demonstration so you can actually see what these functions look like on the ServiceNow platform. Dale? Dale, are you there? Or are you on mute, my friend? I think we're having some technical difficulties. If you could bear with us for a second here. I think we still have Dale on mute here. Just bear with us. We seem to be having some technical problems with our demonstration here. It wouldn't be a live webinar if we weren't having any technical problems here. I think Dale might have had a crash on his computer because I don't see him here. Let me see if there's any questions that anybody has so far in the chat. Can we go ahead and see if we have any questions here? Let me see. We have a question here. Let's see the first one here. Does ServiceNow have a library that is available to the community of practice of high-quality natural language examples here? We do have some prompting examples that are coming in online that we're going to be posting online so that we are going to be showing best practices on how to write better prompts. And so there's definitely going to be some training that we're going to be launching in the coming months, in the coming days to actually show this. So you -- definitely, stay tuned for online, and we can post some links to some of that material in the Resources section as well. Can generative AI update flows besides creating them? I think right now, it's limited to just creating net new ones. But I think in the future, we're going to be able to go in and update specific flows with flow recommendations. Dale, are you back?

I'm back, Gregg. Thank you.

Okay. All right. So we had to -- we took a few questions there while you were down there. But I think everyone's really anxious. They want to see this in action.

Yes. Yes, absolutely. It's definitely one of my favorite features of the platform to show off. So let me set up a little story, if that's okay, Gregg, to help go with the demo. It will make it a little bit of fun. So let's think about a company called Cloud Dimensions. If you've ever been in ServiceNow training, you might have heard of this company. But let's say, for example, they've been a long-time customer of ServiceNow. They have a lot of the products everyone knows, like ITSM, CSM, HR, a lot of others. And they actually have a custom version of incident management that they've built using App Engine. And let's say that they need to add a new e-mail field to the table, the developer says, as usual, "absolutely." So let me share my desktop and show you a little bit about what this looks like in a real-world scenario. So I'm just going to share this one. And you should see the App Engine Management Center for Eric, our developer. So when Eric first logs in, he sees the App Engine Management Center because he's a developer. But he's going to go to App Engine Studio to work on this task. You can see his Cloud Dimensions incident management application. It's just an extension of the ITSM application he built with App Engine. And he's going to go to this incident table where the admin asked him to add that e-mail field. And he's actually just going to go ahead and add it real quick using the table builder here in App Engine. And he's just going to add it as a simple string field, and then we're going to add some validation. So pretty easy. He added a new column, no big deal. So he's going to check out the policies on that table. And today he's going to write a business rule for this because it will help show off Now Assist code generation. He opens up the Business Rules and he goes to make a new one. And this is a normal task for a lot of developers, validating input, writing code, writing business rules. We call this validate e-mail syntax for his business rule and make it advanced, so he can code it. So before we insert or update this Cloud Dimensions incident record, we want to validate that e-mail. And here's where we get to the first Now Assist part. You can see the -- I call it the sparkle icon, but it's our icon for Now Assist, and you can see this one is powered by Now Assist for code generation. So he just writes, validate e-mail using reg ex. I'm going to take my hands off the keyboard for a second while I explain this. Reg ex is regular expressions. This is really long way developer -- okay, it already generated, really long way developers have of validating things are correct on input. Now you noticed that a little swirly came up and generated code. I'm going to hit Enter on my keyboard to accept it. So if you're a developer, you're already reading this, but we're going to get the current value of the e-mail. We're going to validate it with this really long regular expression. And if it's not correct, give the user an error message. Pretty good. Now what this does is it saves Eric all the time of writing really this line 5 here. The other ones, you could probably write, but writing this off the top of your head can be challenging. You probably would have had to have done quite a bit of research to come up with that. But with Now Assist code generation, he didn't have to. He put in a comment, and it generated automatically. Now to help with governance, you can see these purple lines over here show that the line was generated by Now Assist for code generation. So if Eric goes in and changes something, you see the purple goes away. Now that's become a human-edited line, and we can see what was generated by AI and what was not. Now we call that type of code generation text-to-code, but we actually have 2 more. We have what we also call code completion. So let's say Eric does start writing this code out, and he writes -- I'm going to copy and paste it to save us a little time. But he writes the e-mail, he grabs it, he puts the regular expression in there. And he kind of stops for a second, and you could see it was going to start generating before I even move my mouse. But it will then do the swirly and then start suggesting things. So if he sits there, I believe it's 2.5 seconds, it will then start suggesting a completion. He can still hit Enter on his keyboard and accept it. But wait, there's more. There is one final way we call Sure Shot, which means let's absolutely make sure it knows what we're trying to do. That's actually a combination of a comment. The developer starts writing from code, and it still does the same thing. It auto completes it and puts the rest of the code in there. So pretty cool as far as code generation. This is aware of what kind of box you're in. So whether you're in a Business Rule versus a Script Include versus a Client Script or whatever in the platform, this will know what kind of function and syntax it should be writing. You can -- if it doesn't know, it will put a suggestion here and say, "Please clarify or add more information." And just -- this isn't related, but I thought I'd throw this out there. For obscure things like, say, glide aggregate, it will go in, do all the syntax for you and just create it just like that. So that's really how code generation works. And it works even here in the context of App Engine or anywhere else on the platform. So Eric was able to easily write that code, validate the e-mail and then that would make sure when a user submits that e-mail, it works correctly. I'm not going to save this code because this is just an example. But the other thing that we really want to check out is the flow generation. So if we go back to the platform, in the new Washington D.C. release, we have what we call Workflow Studio. So we've taken Flow Designer, process automation designer, integration hub and put them into one interface, we call Workflow Studio. And so within here, Eric is going to create a new flow. So he'll go to a flow. And now he has the option to build from scratch or get started quickly with build with Now Assist. So this is our flow generation. And so let's give it a few examples we could try here. Maybe we want to create a flow on a schedule where we have a lot of problem tickets come in. Our team has a hard time remembering to assign them. So I'm going to copy and paste the prompt to save us some time. But we want the flow to run every day at midnight, find all the new ITSM problem records. And if they're not assigned, assign them to a group, move the state and notify that group. So I'll click build with Now Assist, take my hands off the keyboard. This is sending it to our ServiceNow LLM, trained on by ServiceNow, on ServiceNow, on our ServiceNow hardware, the one you saw Gregg show a slide with we did in conjunction with NVIDIA. And it sends it back, and now we've generated that flow. So that saves Eric quite a bit of time. We've set up the full scaffolding. So we can see there every day at midnight, we're going to look up some records. For each one in it, we're going to do some things. We're going to update them. And it even knows that we wanted to do a Team's message. One of the things I like to do is just add some syntax so I could easily add a comment to the code to say, "Oh, I find all the created for each problem found." Maybe I'll add another comment. So we can quickly put in any kind of prompt there and generate a full flow in Flow Designer. I'll do another couple of examples real quick just to show you what that could look like. So maybe let's do a different one around SLAs. Maybe you've used SLAs in your ITSM or CSM Life. We could say, let's create an SLA. And I'll tell you what this is doing, what it generates, but wait for 50% of the SLA to be complete, send a notification. Wait for 75% of the SLA to be complete, send another notification. And it -- okay, before I could finish reading the prompt, it's already generated it. Now there's a little bit more to fill out here in the current version. In our next release, we plan to actually have this filled out with the data pills or some of you paying attention at home. And finally, I'll put in one more final one that's really complicated just to show you, these can be really long flow prompts. Let's do an unauthorized demo change request. So this one is kind of long. But basically, whenever a change request is created or updated, where it's an unauthorized demo, do the following in parallel. So that's what I told it to do, apply a change approval policy. If the approvals are approved or skipped, update the change request has approved; if not, update it as rejected, evaluate change model. If rejected, send an e-mail. So you could see, I put a pretty long prompt in there. I didn't even finish reading the whole thing, and it built out the entire flow for me with the exact same logic that I had put into the original prompt. So this is really great if you have citizen developers, no-coders that want to put a prompt in and have someone help them finish it out or even me as a seasoned professional developer, this saves me quite a bit of time of dragging and dropping in from the flow out, which I find to be fun, but we can actually focus more on just the business logic of that. So Gregg, I hope that was a good overview of what we are doing these days with Now Assist for Creator.

That was great, Dale. And I always love seeing this in action. It takes me back to my coding days here. And even when I jump into an instance now and dabble in things, it's always pretty extraordinary. I think when you step back and you look at this, it really shows how for -- our generative AI functionality in this area, Now Assist for Creator, is really a game changer. And we're starting to see customers one of the fastest adopted new products that we put out on the ServiceNow platform in years. We've never seen customers adopt this so quickly in all these different industries. Sometimes you might see trends like -- and I see some questions in the Q&A that maybe a particular industry will be slow to adopt it. But even some of the primary users of creator workflows globally, which is public sector, governments and regulated industries, such as banks, have been adopting this quicker than any other industry, which is kind of ironic. And I think it's because they're really using creator workflows to build a lot of these custom applications and custom business workflows. And so kind of in summary, the reason that I think Now Assist is being adopted so quickly is that it's showing that you can provide better efficiency because it is an embedded experience. It's not like you have to take code from ChatGPT and then bring it on over. This -- our text-to-code is embedded within a ServiceNow instance. So any time -- when you have licensed Now Assist for Creator, every single text pump in your instance now is activated with text-to-code. Whether you own CSM, whether you're using App Engine, whether you're using ITSM, whether you're using HRSD, it's now embedded. So you don't have to go. And it doesn't require extra prompting and extra information to feed into the prompt. It knows full well the context of what you're trying to do, trying to create a new business role or do something in that area. I think the other area that's really making this a game changer for a lot of ServiceNow customers is the quality. As you pointed out, our text-to-code and text-to-flow are using LLMs that have been fine-tuned on modern ServiceNow code by the best in the business, our developers. Working with NVIDIA and Hugging Face, the StarCoder model has been trained and that we have taken it and we have specified it to focus on best practices in ServiceNow so they can get clean, performant and maintainable code. If you contrast that and you go out to ChatGPT or you're using any general model to generate ServiceNow code, well, there's a lot of bad code out there. And you might be getting recommendations based on someone's first ServiceNow project. So you have to realize that you're really getting the best of breed because this is focused on our best practices so that you can write the absolute best code within ServiceNow. And then I think another reason that government and a lot of public sector customers and a lot of banks in the ServiceNow customer ecosystem have been very quick to adopt it is that it's the security and the safety of this. And so there's no legal risk that you're using someone else's code. This is within a ServiceNow instance. It's not going externally. It's not making any external calls to generate code or generate workflows. So you're using code recommendations and using flow recommendations that have been trained on data that you really have explicit permission to use as a ServiceNow licensed owner. So it's not legally ambiguous whether or not I can use the output of text-to-code or text-to-flow. But it doesn't stop there. We showed you today text-to-code and text-to-flow, but we are constantly innovating. And in fact, next month, in conjunction with our Knowledge '24 event, we're going to be releasing the next version of Now Assist for Creator, where we're going to be adding to flow generation and code generation with playbook generation and app generation. And in the road map, we're going to have bot generation. We're going to have catalog item generation here. And so as you can imagine, just imagine this coming in from scratch and just typing in and creating an app simply from a Now Assist prompt. Soon, you won't have to imagine it because it's coming later this year. And that's really what's making ServiceNow's offering in this area very, very, very unique. And so there are so many opportunities ahead for you to leverage this within your organization and both empowering your creators to enhance the developer productivity and build better apps, to enhance your knowledge workers that are using these applications so they can make better decisions. Not only are you getting faster development, but you're lowering the barrier of entry for low-coders and no-coders to really expand that pool of available developers to help you [indiscernible] your backlog of automation request, and you're simplifying the deployment. And in some cases, you're really -- you're taking that -- you're enhancing the self-service of employees and intelligently generating content. So are you ready to learn more beyond what you saw today? Well, then head on over and you can look in the Resources section here on this webinar, where you can click on the link for the Unleash Your Developer Superpowers. It's in the Related Content window below. And if you really want to get hands on and you really want to see firsthand what's coming and learn a lot more, you've got to go and register for our Knowledge '24 conference. It's going to be May 7 through May 9 in Las Vegas, the fabulous Las Vegas, Nevada. Dale and myself will be there. We'll be running labs. We'll be running presentation. So register today. But if you like what you saw today and you'd like to see other webinars in this area, check out this link here in our On-Demand Webinar series here, where we've got a lot of other topics. And we'll be launching some webinars in the future with future generative AI technology within the Creator Workflows area. So with that, we'd like to see if there's any questions and take any questions that anybody might have here. Dale, we got any good questions there in the chat you're seeing?

Let's see what questions we have here. I'll take this one right here. I see this one here. What do you consider to be the #1 2024 App Dev trend with -- in terms of generative AI? I will tell you, one of the biggest trends I see in the shift with low-code and code generation and generative AI is governance. I think the fear that I see a lot of CIOs and CTOs have in both not just public sector and regulated markets like banks, but even in manufacturing, in the entertainment industry, in different customer segments, is that low-code lowered the barrier of entry so that people could automate things faster. And now with generative AI, you're lowering the barrier of entry again, and you're speeding up the pace at which you can actually automate and produce new code and new low-code applications. But at the same time, you can also accelerate the amount of technical debt and waste and maybe useless stuff that you take on as an organization. So having a good governance program in place is absolutely critical. And that's why I've seen most customers kind of shifting a lot of their low-code development over to ServiceNow to address this area, where a lot of other low-code platforms help you build stuff fast, but they just help you create more technical debt. And so I see -- this is a trend that I see taking on and people starting to use generative AI to help gear their governance so that not only can you register your low-code applications and the text-to-code, text-to-flow things as assets in your CMDB, but you can now start to see the usage and see the trends and what is the business value I'm getting out of these automations. And guess what, if nobody is using it anymore, start to use generative AI to proactively retire these applications so that you don't wind up with a pile of debt 10 years from now generated by GenAI. So that's where I see the platform like ServiceNow really kind of marrying low-code development, generative AI capabilities, workflow and integrations, along with application governance.

Thanks, Gregg. I'm back. I had a little hiccup again there. That's was the Internet. We had another question. I have a great answer for, if that's okay...

Yes, go for it.

It's one that I actually hear probably the most these days from customers. There's a couple of the same flavor of the same question here. And it is around licensing, which is, do I need special licensing for this? Short answer, yes, you do. But it's a little easier to understand than some of our other products. And it is the first time you're buying a license for a sub-prod or non-production environment. The Now Assist for Creator license is for the developers in the dev environment. It doesn't matter if they're building ITSM, CSM or -- it doesn't matter what they're building. The Now Assist for Creator license, it lets them use any of our generative AI features for developers. So we saw code generation and flow generation. And like you said, Gregg, we'll have other features coming out very soon at a conference called Knowledge that's coming up. You might have heard of it. But yes, it is a separate license, but it's just for the developers.

Excellent. I see a question there coming into the queue. Does Now Assist for Creator work the same way for on-prem customers? And I think this is an important question, especially in regulated markets and public sector, we see a lot of customers with on-prem installations. So the short answer for now is, yes. By using contained closed LLMs that are domain-specific within an instance, it would work the same. I think moving forward, we're going to start to integrate features into the ServiceNow platform that require the use of external LLMs. Maybe text-to-process or text-to-app, you might need a bigger pool. You might need to go to more general-purpose LLMs. So for those features, they would probably work a little bit differently. But for the current features that we showed you today, text-to-code, text-to-flow or Now Assist for Creator, some of our earliest adopters are public sector customers that are using them in an on-prem environment.

Yes. Absolutely. I'd just to add to that, Gregg. One feature we have in addition today that's maybe for our more advanced customers is we do have a generative AI spoke with Now Assist for Creator. So if you're one of our more advanced customers that's already building your own models related to other things in your company, we do have a spoke. You can easily plug your flows straight into those models that you've already built as well. We've got another question here, Gregg. Is the prompt engineering similar to LLMs in a sense that the way the prompt is formulated might produce different results and the created flow? Would it be an iterative process in case we need to recreate it? Sort of, yes. That's kind of the nature of generative AI is it's almost a little bit more human-like than a lot of other technologies I've used in my career, in that you're going to get similar results with the same prompt every time, sometimes the same results. But yes, you might need to tweak the prompt just a little. If you remember, when we came out with -- where you can go into your search bar or your filter navigator and add a query, you kind of have to learn how to talk to it at first. But compared to where we were to where we're at now, it's vastly improved. So now you can use mostly natural language and get really consistent results.

I see a lot of questions about one topic, and I'm just going to kind of summarize the answers around security and not having data go outside of your instance for -- obviously, for legal reasons and getting contaminated with bad code. So by using a closed and domain-specific LLM, that data is never leaving your instance. And our models are not being retrained on your data, so you don't have to worry about somebody writing bad coding and contaminating the recommendations. They will only -- those LLMs will only get updated with new releases by ServiceNow. So you can be much more peace of mind that anything that you type in there is not going outside. It's not going to a public model. We will never use your information to train our models. That is not part of how this works here. So I think there's definitely that peace of mind. So I just want to make sure that was clear because I see like 5 or 6 questions around that same issue here.

Yes. Just to add to that, Gregg, when you do, say, generate code, our model knows all about the ServiceNow data models. It doesn't know about your custom tables just to add to that. If you give it the name of a custom table and say, in this custom table, I want to do a thing with this field, that is this kind of field, it knows what you're saying and it will give you a response. But that type of custom information does not go back into our training data, which kind of leads to one of the other questions is, how do we train it and make it better? That's a really good point. When I was in the demo, and you saw it produced the text and it was in gray italics, that is considered a suggestion. So when I hit the Enter key and I create real code out of it, that's me accepting the suggestion. And that's what goes back into our model to train it. Now we don't have private information about your company that goes into our training data. And we do allow a very easy opt-out feature. You just -- your security team wants you to just uncheck that box, it's very easy to go uncheck. But that's how you train it. You don't really have to do much. It's if you accept those results, we say, okay, that was a good one times thousands and thousands of check. That's what goes into training it.

I see there's a lot of public sector customers, U.S. federal customers on the webinar. And I see a question there is, whether or not that Now Assist for Creator has been officially approved for -- inside of the FedRAMP-approved ServiceNow data centers. I know that's coming. It hasn't been approved yet. It will be available very, very soon. I believe by the end of Q2, beginning of Q3 is when it's targeted. But very, very soon here, within the next few months, it is slated. I know that is in process. Our engineers are working with federal regulators on a daily basis to make sure we deliver on that. So very, very soon.

Yes. We don't want that to be a Christmas present. We want it to be a July 4th present.

Exactly. Let's see what else we got in here. So could Now Assist and Creator Workflows to be used to assist with the configuration development of the Strategic Portfolio Management, SPM module? Yes. So 2 things I'll point out in answering this question. One, when you license and you're active and you have active licenses for Now Assist for Creator, text-to-flow and text-to-code and the subsequent functions we'll be adding to this portfolio get activated on your instance for every single product you own, whether it's SPM, CSM, HRSD, ITSM, ITOM, it doesn't matter. You now have the capabilities to use text-to-code and text-to-flow to generate scripts, generate workflows within all of those products. So it's -- I mean, in some ways, it's Now Assist for the platform. You really -- Now Assist for Creator allows you to add generative AI to your entire instance in your entire portfolio. But I will share with you some creative ways that I've seen some customers start to use generative AI with SPM is by -- we actually have something where you can actually hook in external controllers and external LLMs to generate as part of the logic. And I've seen some customers to start to use that generative AI external controller within SPM to generate custom user stores very quickly. And I thought that was an extraordinarily creative way of using generative AI with SPM specifically.

Nice. So Gregg, I think we're getting a message. We have time for one last question. I think I can handle this one a little bit, but I definitely want to hear your take. It's about what the learning path is for an average IT ServiceNow person to take advantage of these kinds of features. And does it require additional licenses? We covered that one. So when we say an average IT ServiceNow person, it is important to just continue to reiterate: it's for the people that have access to your dev environment to develop things. So if you're an average IT person and you do have access to dev, then you would have access to these features. And the learning curve is pretty small. Obviously, we have a lot of Now learning modules. But in generative AI, we kind of put a prompt in, and you see what results you get back. The one thing I'll add, though, is if you have never coded in your life before, Now Assist code generation is going to produce code. And you may not exactly understand what it's produced, and you can test it and see how it will behave. But it does help to have a little bit of a coding background on that regard. But the code that it produces and the whole point of coding in the ServiceNow platform is it's low-code and it's easy to understand. So that learning curve is pretty small. So the average IT person should be able to tackle it without any issue.

Yes. I -- from the customers that I've spoken to, and I've obviously talked to a few customers that are implementing this with App Engine, with Field Service Management, with a bunch of our other products, some of the early feedback that I've gotten is this actually -- it decreases the amount of time it takes to onboard new people to start using Flow Designer and start writing scripts on the ServiceNow platform. So it was kind of one of the unintended consequences of it, not just take experienced people and make them more productive and write better quality flows and code, but actually bringing new people into the flow of doing this and really lowering that bar of entry and reducing the amount of time it takes to onboard new people. So well, listen, hope you guys have enjoyed today's webinar. We've had a great time doing this. If you want to reach out to your ServiceNow rep or if you want to go over to servicenow.com and get some more information. But we're really excited, and we hope to see you all at Knowledge '24 next month in Las Vegas. Thank you, everyone. Have a great day.

Thanks.