(bright music) (uplifting music) - Hello, everyone, good evening, good afternoon, good morning to everyone around the world. Really excited to share our story today, Now on Now, how we are transforming self-service for both customer and employee workflows. Here is just a brief safe harbor notice of some forward-looking statements of what you'll see through the course of today's discussion and presentation. And my name is Rob Muro. I'm the senior director of technical product management here at ServiceNow, part of our digital customer experience organization. And I'm all about the customer. I'm all about customer workflows and exactly how we can drive value into our business units, leveraging our technology. I'm really excited to share today our experience with regards to AI and generative AI and I'd like to introduce and welcome my partner today, Ali. - Hi, everyone. My name is Ali Fatemi. I'm with the digital technology experience team here at ServiceNow where I currently oversee the employee self-service program. I have over 20 years of experience in IT and customer support where I've helped drive efficiency and improvements in employee and customer experience through the application of innovative technology solutions. - For today's discussion, we're gonna talk about how we're putting AI to work across ServiceNow. What's our vision for transforming experiences with AI throughout the ServiceNow ecosystem? And specifically how we're leveraging AI in customer and employee support to drive business value? Also we will close with what we've learned to date with our experience with generative AI. Our business objectives have not changed here at ServiceNow. We're still all about growth, growing the customer base, and creating the most exciting and productive employee space here for both our employees as well as our customers. And one thing that we have is, as our tool that is being used horizontally is AI. AI is one of the ways to help us get to achieving these business values and these business outcomes. And what's our vision? Our vision is AI is everywhere and it's weaving seamlessly into our work and our interconnected workflows. It's gonna transform our experiences to help not only our employees, but also our customers make better and faster decisions. And what you will see is our mission which stays true, which is to accelerate customer value. Our job number one, which is to drive customer and business value through connected digital experiences and workflows. And that extends all the way from the employee to the customer. - In the support world when we think about shifting left, that traditionally has meant moving work from higher, more expensive levels of support to lower level and ultimately to self-service, which is level zero. When we think about support experiences, self-services, empowering employees and customers to find information and resolve issues on their own. And for most support organization, that continues to be a key focus area. And self-service helps support organizations scale and deliver better experiences and helps employees and customers get to answers and resolutions faster. But with AI, GenAI, we can do even better. We can aim higher. We're redefining what shifting left really means. While self-service is still a much better experience compared to engaging with an agent to resolve an issue, it still requires an employee or a customer to spend some time finding information, resolving issues on their own. In an ideal world, issues should never happen. And when they do, we should be able to resolve them proactively before they impact our users. So the new shift left is really about shifting work from the right-hand, right side of this diagram to the left side, which is more preventive, proactive and predictive. I'll give you a couple of examples. One, we're deploying an agent or agent client collector to all of our employees' laptops. What it does, it monitors it laptop for its performance and it looks for things that may become an issue down the road, and it proactively notifies the employee of eminent issues or it proactively takes action on behalf of the user. So, for example, it could determine or detect that a process has taken a long time to execute or process consuming a lot of CPU power. It could proactively shut it down or it could ask the employee whether they would want ACC to shut it down. Another example of when thinking about predictive is when somebody books international travel today, they go to our traveling software and they book their travel and it usually doesn't occur to them that they may also need an international data plan for when they're at their destinations. And sometimes they land and they realize their phone is not working or it occurs to them later on they have to go and submit a separate request. So being predictive in this case means that as someone is booking their travel, based on their behaviors and patterns in the past, we know that every time they booked international travel, they also requested a national data plan, so we can offer it to them before they even express it. That's what it means to be predictive and that's what can take the experience just to the next level. AI has been embedded in our platform for over six years. For the past six years we've been implementing and piloting more and more use cases using AI within everyday work across all of our departments. We found that when used effectively, AI can significantly improve three core aspects of our business. The first one is experience. Now experience with AI and GenAI, experiences have become much more natural and conversational, which is a vast improvement in how employees and customer experience support and self-service. But also using AI, we can analyze massive amounts of data and based on that data, tailor experience to our employees and customers' preferences. So in other words, we can hyper-personalize their experiences. Speed is another aspect. Since we've integrated AI into our employees and customer workflows, what's happening behind the scenes simply happens much faster. And we're gonna see a demo in a minute of examples of where some of the manual work that happens, used to happen in the past, now happen with AI and GenAI and that results in much faster timed resolutions and other key KPIs. And lastly, decisions. AI can process, analyze large volumes of data, identifying patterns and trends that may not be apparent to us. This intelligence helps our decision makers make more informed choices to positively impact our employee and customer experiences. - And when we're talking about decision making, there are a number of AI capabilities that customers and employees can both leverage, and these will help drive business OKR such as increases in NPS and CSAT, increase in cost avoidance, as well as the lowering of churn rates. As we look at some of our solutions that we leverage today, one of the key areas of operational alignment specifically is in our customer support dashboard. And one of the operational alignment areas between both employee and customer is exactly how we measure self-service. So through a number of iterations, our definitions are aligned, specifically how we're measuring self-service exactly when customers are leveraging our automations as well as our content. This is very key in terms of helping drive our decision making and to see some of our additional decision making within the platform. Here we'll now go through some of the innovations with a brief demo. - We're gonna see how AI can drive decision making in this example. And we're gonna be really looking at process optimization. And process optimization enables us to identify and remediate process inefficiencies, gaps and bottlenecks. We're gonna use the example of Sam, who's the support operations manager. And he uses these dashboards to measure and track his KPIs, including mean time to resolution. Now he notices that MTTR, which is one of his key KPIs, is trending in the wrong direction. So with one click, he can go to the process mining workspace to understand the bottlenecks and look for improvement opportunities. The workspace provides opportunities, categorizes them for him. That way he can see what steps are being repeated or incidents that are being bounced around. One of the steps that he decides to focus on is awaiting caller info, 'cause he notices that that's one of the bottlenecks. So he decides to focus on that and it double clicks on the subset of incidents to understand more of the bottlenecks. He then uses the breakdown on the left-hand side to understand these incidents that are going in and out of the reading caller info. He tries to understand which assignment groups are mostly impacted by these transitions and then he can also look at priority and category of incidents to see if there's any correlation or if there's any impact. He finally looks at channel and for channel, he notices that there is, there seems to be some correlation with higher MTTR. The majority of tickets are coming in through self-servicing portal, which is great, but he focuses on email. Now he wants to analyze that are coming in through email and he wants to analyze awaiting caller info. The reason that he wants to do that is he wants to understand whether possibly, for example, a virtual agent conversation could help improve the experience. Now with incidents coming in through email, he notices the average entity MTTR is two weeks and he decides to, this is where he's gonna focus his effort on. Now he drills down in the 1,200 records and then he also uses clustering. He notices that there's a cluster of incidents when users are trying to update their email address on their profile. And he decides that this is a great self-service option to provide to users. So he ultimately thinks that analysis-powered virtual agent topic would be the right solution. So he creates an automation request and routes it to the implementation team for execution. - Now when we're talking about AI and customer and employee support, we're talking about speed. It is a known data point that within 48 hours after a case or incident has been submitted, typically the ESAT or CSAT for that case will go down. Just as in our everyday life, we're all looking to having our ability to self-serve and we want that ability to self-serve and we want it to be done fast. One of the areas that we really see the help facilitate that need for speed is routing and categorization, and specifically how we can better fine tune the routing and categorization of our cases and incidents into the right assignment teams and the assignment groups based off their skills. This is where we've actually seen, through the utilization of AI and ML, how we're able to increase that proficiency all the way up to 85% and beyond. - Now we're gonna see a demo of that, how AI is improving the speed of what happens behind the scenes. We're gonna look at five different use cases. The first one is chat summarization. This is where a summary of the virtual agent chat is presented to the agent that takes over the ticket. Incident summarization, which happens mostly in the case of handover and for a engineer agent to come up to speed. Suggestions by AI based on historical data and recording of resolution notes, so after a ticket has been resolved, summarizing the notes into resolution notes, that's one of the use cases. And finally capturing the learnings from resolution of a ticket into a knowledge article. It starts with Ashok who's an employee who's interacting with virtual agent. He has an issue with his laptop. And when virtual agent not able to help him, he has to be connected to a live agent, which is Catherine in this case. And Catherine decides that she needs help from higher level of support, Melvin and eventually Alan, who's the on-call engineer, becomes involved. So you can see Ashok here he's gone to My Bot or a virtual agent. He typed in his issue and got a response from a virtual agent. He got analysis response, which is a snippet of a knowledge article. But after trying what's been presented here, he realizes that this did not help him, so he feels like he should chat with a live agent. So he connects with a live agent. And now we're gonna see the view of Catherine, who's the live agent. She first accepts the incoming ticket and then she can see a summary. She asked to receive summary of the chat that the, everything that transpired between virtual agent and Ashok. He looks through it, she looks through it, and then she realizes that this requires the skills of an engineer, someone who has more knowledge and more expertise in this particular type of issue. So she decides to elevate this ticket to higher levels of engineering, so she creates an incident off of this ticket and assigns it to Melvin, who's our level three engineer. So as you can see, there's so much has happened throughout the resolution of this ticket and Melvin wants to come up to speed. So she clicks on the summary. So the incident now is becoming summarized for Melvin. Now he comes up to speed a lot faster than he would normally. So he decides now to see if there is an AI-suggested solution that would work for this employee. And as it happens, the suggested solution is very relevant. So Melvin decides to share this resolution with the employee. He posts a comment and then, but he's coming to the end of his shift, so he's gonna reassign the ticket to the on-call engineer, Alan. And as part of that handoff, the incident, again, is summarized for Alan, so that way Alan can come up to speed much faster. Now when Alan sees the ticket, one of the things that he notices. He notices that the employee has accepted the resolution. The employee feels that that resolution helped them. So he changed the state of the ticket to resolved. And as you can see, resolution details are being generated automatically and put in the resolution notes. So the last thing that he's going to do, he wants to, he feels that what transpired with this and how this ticket was handled could benefit other agents and employees, so he decides to create a knowledge article from this. And the knowledge article, a draft of the knowledge article is created, he clicks on it and you can see the body of the knowledge article. And he can now make changes to it before editing it. And finally, this knowledge article becomes available, not just to other agents who might encounter a similar issue with other employees, but also to employees to drive self-service. - And that's a great experience to see the power of being able to shift left through the creation of a case or an incident, how that could be transformed into knowledge information that can be fed back in to the AI services. And within that, we really get into the concept of AI search. And the importance, as you see here, of content, the content as well as the automations that are being leveraged to help drive an improved experience for our customers and for our employees. And these all translates into behavior metrics, specifically how we can measure and grow self-service as well as increase case deflection. Also what is the search conversation success rates as our users are attempting to self-serve through the utilization of Now Assist for AI search. And then lastly, the user feedback and how that telemetry information can create more intelligent LLMs for our customers and employees through better content recommendations to once again drive increased self-service. - Now here we're gonna see an example of where generative AI and AI are helping improve our employee In this example, you're looking at an issue reporting form. So again, Ashok comes here and types in his issue, which is his laptop is overheating. Once he presses Continue, now GenAI takes over and you see, you will soon see a snippet of an article that deals with laptop overheating. What we're finding is a lot of times our employees are finding this snippet very helpful and it resolves their issue, so they basically do not need to open a ticket anymore. You can see the source of the knowledge article is also displayed there, as well as other knowledge articles that might be relevant related to this issue. So all of that is presented to Ashok as resources, and again, we found a great deal of success. When this doesn't help him, he can still click on Report an issue and here where you see AI in action again. Now based on the description of the issue, which is laptop overheating, we predict what business service this maps to. And in this case, it maps to laptop endpoints Mac. What this does, the business service is mapped to a specific assignment group. So that ticket goes directly to the right assignment group, which then cuts the time of MTTR and gets ticket to resolution much faster. Now as you saw in the examples and the demos, generative AI is changing the service landscape. It can do so much. It can help automate repetitive tasks, enhance the overall self-service experience. It reduces the skill gap between agents, it guides decision making, it assists with high-value essence and much, much more. Ultimately, what we expect to achieve and what we've already seen is improvements in deflection, agent productivity, customers and employee sentiments are improving, because they're getting a better experience with AI and GenAI. Agent retention agents are more satisfied and fulfilled with their job because they work on high-value, highly complex, more challenging types of issues. And we're gaining all of these benefits while at the same time, decreasing manager escalation and costs. - Now what you'll see are some of the value metrics of what we've already started to see with regards to the business value, more specifically those benefits that we're generating through the utilization of generative AI. Now when we're talking specifically about self-service, as we mentioned, we've already seen a 10% boost to our case and incident deflection rates, and that's twice the deflection rates with the utilization of Now Assist and what's being serviced through our Now Assist for search. Now as you also see with regards to our key business driver, which is productivity and how we have already seen, for customer employee experiences, over 21,000 hours saved per year with generative AI search and 30 seconds for our customers to self-serve with Now Assist. Now once again, when we're talking about productivity, it's all about our agents and we are already seeing up to 30 minutes saved per day per agent with Now Assist and a 15% productivity improvement per case and incident. And that's just in such a short amount of time. Now we're also seeing the ability to eliminate tedious and repetitive tasks. Over 30 minutes time saved per KB article that's generated. And that just boosts back into our ability to shift left and drive self-service as well as 15 minutes saved per case per incident with each handoff. Great results in such a short amount of time with our ability to drive value back into our business units and more importantly to our customers and employees. And what are some of the things we've learned over just the past six to seven months? First and foremost, it's all about the data, feed it. Have generative AI trained on your own production data, whether it's from an employee perspective, as well as from a customer perspective. How we can ensure that when we're talking about the utilization of generative AI and completeness and hallucination rates, it's all based off of our data. Also how we can begin to feed it with more additional data, whether that's content types, automations, to produce more results to drive better self-service. Focus on your top use cases. Start at the top, make sure that you're measuring, because you'll be asked exactly what were your results prior to the utilization of generative AI and how those results have improved through the utilization of generative AI. Also what are possible opportunities for additional prop tuning for more and exact results through the utilization of generative AI as well as traditional AI and ML. And user experience is critical. Continue to test those flows, whether it's from an employee perspective or a customer perspective. It's not just the accuracy of the workflow, it's the delight of the experience as well. So please continue to test. And lastly, it's a journey, so embrace the pace, granted it's a journey that we're all being asked to execute at a five-minute mile, but that accelerated pace is what's gonna drive even faster business value. And so what are the three reasons we're excited about generative AI? One, it's in the Now platform and it's a game-changer for ServiceNow, and it's happening so fast, so embrace the paced. Even more so, we can deploy generative AI without technical or organizational complexity. We can move fast. It doesn't take a tremendous amount of product management or engineering resources to get these use cases out into your environments and driving business value. And as you see, we've realized over 10 million plus in annualized cost takeout savings and productivity savings through our implementation If you have any questions, please feel free to reach out. Really excited that you've taken the time to listen to our session and really wish you the best on your journey with both AI, ML and generative AI on the Now platform. Please scan the QR code on the screen to view other sessions presented at Knowledge and learn more about how we're using ServiceNow technology across the organization to run our own business more effectively, thank you.