Name: Reimagining ServiceNow CMDB with Agentic AI Workflows
Uploaded: 2025-11-10T17:31:39.007Z
Duration: 1 h 4 min 36 s
Description: Reimagining ServiceNow CMDB with Agentic AI Workflows

Transcript for "Reimagining ServiceNow CMDB with Agentic AI Workflows": Hello, everyone, and welcome to our third Now Ben webinar. Thank you very much for signing up and being with us today. If you don't know me, I'm Saima, the senior campaign manager at Now Ben, and I will be your host for today's session. So some very quick housekeeping before we get started. This session is being recorded, and it will be shared in our follow-up email. And you are in listener only mode, so you may have noticed that you only have the chat and the q and a tab. Please do use them respectively to interact with us throughout the session, and there will be time for q and a at the end, but please do share your questions throughout. So just to give you some more background on now, Ben, the team and I have been apologies. I just realized I haven't been sharing my screen. Let me very quickly pull that up for you now. Lovely. So, yeah, just to give you a little bit of background on Now Ben, the team and I have been monitoring the tech ecosystem for a while to see where the hotspots are, and ServiceNow is an ecosystem which we have been watching for a very long time. We've seen a lot of people actually that have switched from the Salesforce ecosystem to the ServiceNow ecosystem. So we're really excited to enter into this new space given all of the innovation that's happening on the platform as well as the growth, obviously. We're really keen to support you with great content that essentially tells you what's going on in the world of ServiceNow. So we want to cover at a higher level what acquisitions ServiceNow are making, the direction of the business, and also dive into some of the more technical aspects as well. And we love to bring in thought leaders from the ServiceNow community to be a part of that conversation. They'll be sharing best practices and tips on how you can ultimately advance your career on the ServiceNow platform. And, obviously, this webinar forms a part of that picture and is a little taste of something new that we hope to continue doing at Now Ben. So we really do appreciate you showing up and spending your time with us today. We have been launched for a few months now, actually, and we have a newsletter where you can sign up to receive all of the ServiceNow news direct to your inbox. So if you're curious to hear more about what's going on in the ecosystem and hear from other thought leaders, then do sign up via the link in the chat right now. But, of course, so you're all here to hear more about CMDB. So I would love to invite to the stage our two speakers for today's session, Abhijit Yadav and Hardeet Singh. Hi, Abby. Hi, Hardeet. Love you both for joining us. It would be great if you could, yeah, kinda tell us a little bit about yourselves and, your experience. Abby, if we start with you. Sure. Thanks, Haima. Hello, everyone. Abhijit Yadav here. Abhijit has been popularly known over over a decade in in ServiceNow space. I started my career in ServiceNow pretty much when it was ServiceNow Berlin release in the market. So you can imagine it's pretty much up to, end of the alphabet. That's that's where we are at the moment. So it's been around fourteen years I'm in ServiceNow space. Started my career pretty much as a as everybody else being developed for an ideas platform. Gradually, learned quite a lot in the platform, be it integrations, be it item, be it FSO, FSM, and other areas. And, now I'm through with the platform. It's it's lovely. Now I'm pretty much on the high level designing and solutions where I architect, the the end user infrastructures more. How should be they're working on the platform? How the interaction should be happening, how what are the best practices, and that's where I'm directing my own firm, Alfa Root Limited in The UK at the moment. So, yeah, it's been a great journey throughout the platform. Lots of learning. Amazing. Thank you. And, Hardeep? Hello, everyone. It's great to be back, and thanks for inviting me again, Sama. For those who don't know me, my name is Hardit Singh, and I have been working in IT for around fourteen and a half years and eleven plus years of experience in ServiceNow. And like Abi said, we are very old in ServiceNow, and I think I started my journey with Helsinki release. And I've worked across multiple modules of ServiceNow. Currently, I'm working primarily on agentic AI, CMDB, and custom solutions. I'm working as a technical architect and trying my hands on presales as well. And, mostly, what I have seen in CMDB right now, nowadays, people are facing issues because we have millions of records, and we will be discussing today how now assist is helping them out, how we are dealing with the duplicate CIs, how we are dealing with creation of CIs, and so on. Thank you. Amazing. Thank you. Perfect. So I'll just very quickly go into, what it is that we'll be, covering today. So we will be covering understanding CMDB, and we'll also be moving into discovery basics and configuration strategies right the way through to Now Assist for CMDB. And we're gonna spend the first couple of minutes, the next ten or so minutes, I would say, discussing today's topic a little bit more broadly before going into, the presentation. So I'm gonna stop sharing at this point. And, yeah, to kick us off, it would be great if we could understand a little bit more about how you have both worked with ServiceNow CMDB and Now Assist in your recent projects. So, Abhi, if we start off with you. Sure. Thanks, Sama, for the question. So, CMDB has been heart of the projects that I've been working on for for for ages now. Be it any IT SIM project, be it any other project, I have seen the most critical part that organizations are considered about is their CMDB because that has been the heart of the system. Every ticket that you create is centric around that CMDB. Every for example, if an incident imagine you if you don't have the right CI data there, it's just like you're creating an incident and you don't know which CI you're tackling with. So that is where the major block. So you call up your service desk. You tell them, hey. I have some issues. And they don't know which CI to tag that issue to. So that's where the heart of the system comes in a picture that you need to have to have the right, correct, and complete CMDB. When I say complete, again, getting a 100% CMDB is typically a myth. There are few things that I'll be talking about as well when we go up to the time, but having 100% accurate CMDB, is it possible? Very hard. Very hard to reach up to that level. I wouldn't say impossible because there are scenarios where for some time, it can be up to that person, but there are reasons why we cannot be on 100% of the time, plus the completeness. You need to have a complete CMDB. We cannot have the missing data there in the CMDB. You can have few missing tickets at point, but missing the CMDB completely is is definitely not not acceptable. Those are being the critical parts in the system, just not only for the ticketing purpose, but only but for the reporting purposes as well. So that's where the question comes in. How many CS do I have? What is the cost of my infrastructure? Am I paying much more than I'm supposed to? Imagine I'm I have 100 Windows servers, but I'm paying for one twenty. That's the whole idea of having the right same DB. You need to not only mark the correct CIs on the on on on put put those on the table, but also need to make sure that you are not overly paying those with respect to the, actual requirements, what you have. So, yes, it it it has to be people's depending on the people operating that, it must be good enough for agents to rectify the issues correctly. It should be cost effective for the organization to keep it line, keep it maintained. So there's there are multiple aspects to it, and we are gonna cover some of those today. So it has. been very, very critical part of my routine projects. Perfect. Thank you. Okay. Yeah. And we're looking forward to. diving deeper into that. Hardi, anything to add there? Or Right. So like Abhijit said, we have millions of records in CMDB and CMDB completeness is a myth. But right now, we have agentic AI. We have generative AI coming to help us out. And I think down the line, there will be many automations where you can actually go in and check how many duplicate records are there, how many, missing CIs are there right now. So I think in the future, I see many AI agents coming up, many generative AI features coming up. And currently, we have many already there, which we will be talking about, like how to detect a duplicate CI and all. But, yes, down the line, I think we are there, and ServiceNow is working a lot to, create those agents and have the those automations. In my current project, what I have seen is, couple of issues, actually. The first one is the duplication. So if there are, like, with if we take Abby's example, there are 100 Windows servers, but it's showing up 200 Windows server. Right? So how to identify those duplicates? Is there an automation? And secondly, how to fix them? So all this can be automated. And maybe if you don't even want to automate, agentic AI can help you identify those and rectify those issues as well. So I think we will discuss we where we are today and what's the future, in ServiceNow. Great. Thank you. Really, really looking forward to diving deeper into this. Avi, I know that we're going to be touching on discovery, as well. So are there any prerequisites for for discovery? Yes. There are there are prerequisites in in in discovery. When I talk about discovery, that we should be not the first things first comes into picture is, do we need discovery and why? There are multiple ways to populate your same DB. Discovery is one of the method of it. Don't directly jump into discovery. It is required, but figure out the need of it. That's the first thing. Second, there are the there are some technical fulfillments as as well. So talking about installing a mid server is there, then having your firewall connection changes, that that needs to be implemented organization wide. When I talk about credential, there are scenarios where you'll need admin credentials of some of the devices. Agree those initially with, with your technical teams that, okay, they are going to provide those credentials. If not, what is going to happen? So there are prerequisites, Swish. So these are there are some these are the technical ones. And as I said, there's some process level prerequisites as well to get those preapprovals for having admin credentials in place. Raise those changes before for the infrastructure that you need to, discover in in your project. So there are some guidelines, which will help you to save time. So, yes, these are the frequent prerequisite prerequisites there for discovery. Yes. Amazing. Thank you. And appreciate that, I don't wanna take up too much of, the webinar time today. So just one last question, Harvey. What would you say are the different analysis skills and AI, agents available out of the box? Right. So ServiceNow, is currently building AI agents. I think every month, they are launching, new AI agents every month. But currently, I would say there are major two skills which I see. The first one is, CI summarization where you can summarize your CI. And, secondly, there is one deduplication skill, which helps you finding out the duplicates as well as rectifying them. And then I think we have many AI agents as well. So if you want to create your CIs, if you want to, sort your service graph connectors and so on, So there are these AI agents currently available in ServiceNow which are helping us in fixing data of CMDB. And like I said, there are thousands of c AI agents with ServiceNow is currently working on, so we are expecting lot of AI agents coming up in future releases as well as they are releasing every month. So anything. Yeah. Amazing. Thank you. So, yeah, those are all the questions that I think we have time for right now. As I mentioned, in my intro, we will be, leaving some space and time for q and a at the end of today's session. So if you do have any questions, feel free to pop them into our q and a tab. Otherwise, I will hand over to you, Abhi, to kick start the next part of today's session. Sure. Thank you so much, Aima. So let me start, sharing my screen, and please confirm if you can see that. Can you see my screen? yes. Awesome. So let me start with something very, very basic. I know, there's few audience over here who are understanding, the basics, and there are there's a few audience who wants to jump into some of the additional details of well, basically, they're here for more than the basics. We'll find some time for that as well to cover as much as deep as as we could. So we obviously know what CI these are some of the screenshots coming in from ServiceNow site. How do you address those configuration item? There's been few discussions on that CI. If I get into too technical and if I want to be very specific about it, believe me, it's just one record in one table. So you have a hierarchy of these tables, which is in your CMDB, configuration management database. When I talk about CI, it's particularly highlighting about your infrastructure elements. You could have your 100 Windows servers, as I said, so that Windows Server, that becomes a CI class. At times, it is known as CI type as well. So if you don't don't don't get confused between these terms. Some people call it a CI class. At times, it is referred as CI type. There are different ways to call it, but, yes, that's how it is. So Linux servers, all of these coming under one common class server. All of these are coming under hardware, which is the parent loss of it. So that's how these things are organized in terms of CMDB. That's where we are at the moment. So CMDB, it's just a set of the tables which tells you, okay. This is how my table is organized. Again, how can I add my new CI class there? Multiple ways. You can go just by your existing way of, getting into tables, create a new table in ServiceNow, where your parent table would be some other ServiceNow, CI class. Or you go to CI class manager and then add another CI class there. You know what the hardest part here in this in this case? It's not creating a table. Creating a table may may be five minutes, ten minutes job. You may have your the moment you mark your parent table there, it just grabs all the field straight away. The hardest part is where do I need to put my new CI there. That deciding that is the hardest part. So make sure you do it right away. Make sure you understand, the consequences of it. So for example, I cannot put an there's a different class of servers there which talks about the application servers, Tomcat, another, other servers. You cannot put your physical server there. This is just one of the examples that don't put your CI in the wrong wrong hierarchy, which eventually tells you when I talk about wrong hierarchy leads up to getting some wrong or not required field on your CI. And that's where your same to be health monitoring comes in the picture. How would you monitor your health? You can monitor it either using your class view, service view, or any other views. When I say health, it talks about your completeness, your correctness. These are two critical parameters that that I need over here. When I say completeness, all the mandatory fields should be populated on the CI. When I say correctness, it just checks whether all of those fields having the right values. Some of the fields expecting a text, and then they are entering numbers. Some of the fields are entering number. There are different checks you need to apply on these fields, and that's where it it it goes in. The next thing, ways to populate CMDB CMDB. Of course, I'll be showing some of the quick demonstrations as well how will this look into your platform. But let me just give you a good idea of, of of what CMDB is and how do you have that data in place, understand the basics of it, understand the nature of it. There are multiple ways to populate your CMDB. These are some of the popular ones that that that I'm putting in over here. The first one, obviously, is manual entry. Simplest one. You don't need any, hard things to to do that. You have your Excel or you have your documents received, and you keep entering those one by one by one. Again, who would do that? But in some cases, there are few CAs which needs to be entered manually, especially when you don't have your ServiceNow Discovery access to some of the DMGs, which are secure zone. So that's when the manual entry. Eventually, the imports comes in a picture. Other than having hundreds of the CS coming in manual entry, you prefer imports. You use your CSV files, your Excel files, even your some of the time, the integrations. Do you want to get the data through some of the rest APIs? That's also part of the imports. Do you want to get your data through file, which are FTP? So file transport protocol something. Secure protocol. Then that's also that also comes under your imports. Have you ever imported through some JDBC connections? That also comes in the imports method as well. Again, this is moderate to important. All of these two methods, manual and imports, the clear challenges here are data accuracy and the relationships. Why I put it that way? Imagine entering manually and rather than putting, you know, 512 MB of RAM, you put five one and miss misentering two at the end. So that becomes a 51 MB of the RAM to that CI. So that is the reason manual and imports are not so accurate. The third method, obviously, which is service graph connectors. Again, my personal opinion here is service graph connectors is definitely not discovery. People confuse that with Discovery and Service Graph connector, but Service Graph connector, in my view, is completely different process than Discovery. Service Graph connectors is nothing but your import, which connects with your database of the CI somewhere else. It could be on the cloud. It could be other platform. For example, your SCCM. It's it's another platform, so your service graph connected for it. It imports the data from your SCCM to your ServiceNow tables. So I don't consider that as a discovery by by myself, but there are few ways to convince it to to your customers. It's it's common at times. They just understand. Okay. Discovery means getting your ServiceNow same to data in. The context changes, and that's that's where this pitch changed to to talk to our customers about it. But service wrapped connectors are imports or integrations for me in in in in other words. Obviously, the use case here for service graph connectors is you should have that trusted data source. It's it depends how much you trust your data at SCCM level. For example, when I was talking about it, let's take the same example here. If the data in the source is not correct, the data that the graph brings in will be the same from there. So make sure the data and the source is correct in that case. That is where the field mapping and credentials, the common issues over here. Imagine there's some credential changes on the, on the on the source side of the data. Eventually, it will be hampered on the destination side as well, which is in ServiceNow. It clearly shows you its credentials error in the in the in the in the logs. So that's where it is. It is fairly simple to implement. Again, you must have been facing some challenges there, and some of you might be surprised why Abhijit is calling that as simple to implement. Why it is simple? Because there is standard set of errors around service card. You will have either fields missing. You will have the credentials missing. Some values are going here and there, so you need to correct the field maps. That's the reason, whatever I've implemented so far, I think I feel service I feel service graph connectors are a simple way of getting it. Discovery, I love this because I see this as a pure part of your CMDB CI getting the data in. When I talk about discovery, it is fairly complex to implement. When I say that, ServiceNow is doing most of the job for you already when it when the data comes to ServiceNow through mid servers or through the ECC queues. But when as a fairly complex, the complex part here is starts with the starts with your client connect client communications. You need to convince them that you need admin credentials, first part. Second, you talk to their, firewall and network team, and you tell them, okay. You need to open some of the ports to communicate to Internet. You need to install mid server. So that that is way more complex than actually implementing discovery. Complex part comes in a picture in the actual implementation when you need to write your own patterns or customize the existing patterns, whatever you have in ServiceNow. But the common challenges here are credentials and the firewalls, as I mentioned earlier, which are your, initial level challenges that you need to think of. Let's understand some of the basics that that we have over here. First questions first questions first, why it is needed to get the clear visibility of the data? That would that would be my very simple answer. I believe sticking to the basics of it. Discovery digs down into your CIs. There is something which is actually connecting from mid server to you. It goes to your actual Windows server, Linux server, whatever the device is. Could be a switch. Could be a router. It actually goes there, runs some commands, extracts the data, and then feeds that real information into ServiceNow. That's why it is needed. That's why I believe it's the crucial part to get the correct information of the CI. It could be just one field for us as a technical person who want to come on the platform, but that field, its value means a lot to the actual guys working on that, on that actual infrastructure. There are may major services depending on that one particular CI, and that's why it's needed to have the correct and the live information of it. How ServiceNow actually collects that information then if I need to have that information updated? It has its four phases. There are some LinkedIn posts that I that I that I'm pretty much active on talking about these phases before. You'll find that post already. However, for the larger audience, I just thought to go over this as a part of the discussions over here. The first phase, old scam. Exactly what happens here in the diagram if you see, the job for discovery is initiated from ServiceNow. It when you run your discovery, its respective job is placed into ECC queue, which is our external communication channel queue. ServiceNow just creates a job there. That's it. What the what mid server does then, it picks up that job from the ECC queue and understands, okay. These are the ports and the IP addresses I need to ping on. So it first figures out the IP addresses. Okay. And let's go on and scan the different ports on these devices. So mid server fires probes 1234, five in this case because I have five devices over here. Quiets a probe. Gets the information. Okay. This particular device has responded on this particular probe. Eventually means it decides what type of CI it is. When I say CI class, that's what the classification means, second phase terms in. It depends on the what ECCQ has gathered in the first phase in the port scan. So if one particular machine or IP addresses has responded with port, you know, one six one. That means it's a network device. There are some standard port groups given by ServiceNow. You can change those, obviously, but it is rarest of the rare. I have seen different device company and using any different ports for this because that's organized. It's no way to change. So that doesn't happen very frequently. Never ever in my case. So classification. Decide in the class of your CI, which is the next thing. Are you a Windows server? Are you a Linux server? Are you a printer? Are you a router? Or what are you? That's when you would get into the classification phase. The the third and important thing is your identification, which means, tell me, are you a new CI or existing CI? Which eventually means, should I update the CI if you're already there in my tables, or should I create a new CI if you're not there? That's when the identification comes in a picture, and that's when your ServiceNow patterns are triggered in your identification section and exploration. Now that I know, okay, this is, the this is a new CI. Let's create a new one. And in order to create a new CI, I need some of the more fields to be populated in my tables, and that's when I'll be firing some more commands and getting more in in information about you. That's called as exploration in in in in in a nutshell, basically. How I remember these phases? PCIE. That's my way of remembering it in a in a in a quick quick view. Let's jump into the session, with with with respect to how this looks in ServiceNow instance, and then probably we can take it forward from there. Let's understand how the schedules look like, and then we'll be getting into the implementation strategies. Again, this is PDI. So if you see some slowness, that's definitely on on my heads. It's it's PDI. We all know that at times, it it runs slow. Hopefully, it'll not be today. So what I did for my own convenience, I have just marked some of the critical fields as as mandatory. So let me show you, with the CI class manager. This is how it looks like. I just opened some of the, tabs already. So this is your Windows Server class. Again, you must be thinking why Abhi is checking all the Windows Server examples because it's easier for everybody in the in the in the in the webinar to to understand that. We can go as complex as possible maybe in the q and a session. So in the basic information, this is the hierarchy where your Windows servers are. So it's for example, if I say, Windows server at the moment over here So where my Windows Server sits? This is the hierarchy. This is whole seem to be of it. So total, I have 2,784 items, out of which I have eight Windows Servers over here. Eight is the number. The information, you can change the display name over here. Server name, you cannot description basic. My critical field here is the managed by group. By default, imagine if I'm creating a Windows server, who is managing that group? So when you create an incident, if it's a Windows server class, you can make this to be assigned to this particular group. Applied all the members of unless see, this information is given already here, but you we'll just need to read that and and map it correctly. Attributes, nothing but the fields. I won't be getting into much details of it. We all know if we want more fields, we can create new fields from from here. Identification rule. This is critical. Why do we need this? How do I identify my CI uniquely? That's where my, PCIe phase identification comes in a picture over here, and I talk about it. So where did that go? Here. Here I am. In general, in most of the CIs, it is the serial number. How I identify my, hardware class in particular. It's pretty much exactly like so there would be multiple people with the same name, Abhijit Yadav. However, what are my individual identities? My passport number would be different. My my driving license number would be unique. So those act as my unique identifier. If you see this frequently known as criterion attributes, this term is nothing but your unique identification of those particular CIs. And this you will see this term used in patterns, basically. They call it as credit attributes in the pattern, so that's how you can link those altogether. So it's it's good to have that in mind. Make sure that you have the right identification route. Be very careful while changing this identification rules. Try not to change it. Try stick to as close to out of the box. Avoid changing it. Reconciliation, another something very, very critical over here. This talks about how determine which discovery source are allowed to populate the attributes. So if you have let's say your discovery is getting up the data as your RAM is 512 MB, whereas SCCM comes in and then tells, you know, the RAM is actually one gig. Which platform do you want to update that one particular field? Whether you trust discovery more or the the data coming from SCCM is marketed for you. So that's where your reconciliation rules comes in a picture, which data source you trust more for one particular information in that particular field. So that's where your reconciliation comes in a picture. So this talks about your data refresh rules, again, pretty much which is the more important source, and this is for the particular field that you need to apply to. Few fields, let's let's move on to actually getting the data in. How do we do that? So you have discovery schedules. You can create tune that you can go ahead. Again, let me talk about some of the things on the slide that that we're talking about, some strategies over here in in discovery schedules. Something difficult to understand is how you name those schedule. I particular prefer the location, whether it's all CIs or one particular CI type. You can select from here what do you want to scan in the network. I prefer configuration item as my generic one, but this is this is the list of IP address on net network discovery to understand what we already have. How do you want to, use the mid servers? You can use one particular mid server or you can use a specific cluster. I wish we could get into the details of it, but, clusters are much more, efficient when when when to use those wisely. Maximum time, people frequently avoid this, how long you want to run this, for. And I put it as maximum of five or six hours depending on the infrastructure that I'm scanning on. This is where you have something important which talks about discovery range set. How many ranges you want to include here. In this case, just because it's the I'm scanning all the CIs in London for my for my organization. That's where I'm putting a range for network devices where a range of IP addresses include is it ten dot zero dot two dot twenty to thirty, and there's another range which talks about my Windows servers or or so. Again, you can automatically create those those ranges using, you know, quick ranges, and you can add different ways to get those ranges. It's quick. It's easy. Something important about the ranges is if your client tells you, okay. I need to avoid some of the IP addresses. That's where you have a global IP exclusion. This particular range will tell you how would you want to exclude some of those those ranges to be scanned. So yeah. Talking about the strategies now is either you can go by locations. When I say, okay. Let me scan all the device types in my one schedule or in for that one particular location, or you can divide those as per your, CI types. So when I imagine I create a different schedule for Windows servers if I have too many, only for London. So I would call that as London Windows discovery. And I would call this in this case, it's London all CI. So it could be the location dependent. It could be your CI type dependent. Try to keep it as low as possible. One schedule should not run forever. That's why you have this maximum run time given over here. How frequently do you want to run that is over here, but critically important fields here are after discovery. So you you want to sequence those? That is also possible. So I would run this after discovery and run after another schedule that you have in the mind. I I don't have any at the moment, but this is where you can choose from. So, yeah, that's that's a quick one on how do we actually run, run this. And, I'll be waiting for q and a and hand it over to Hardit. Let me stop sharing it over to Hardit over to you for the next one. Thank you, Obi. And, it was an amazing overview of discovery, and, definitely, I will remember that PCIe thing. And I'll just go. ahead and share my screen now. Right. So let's talk about Now Assist. So those who of you don't know what is Now Assist, I'll just give you a brief. Now Assist is the generative AI feature of ServiceNow where you can actually go and generate some data out of a record. Maybe if you want to summarize a record, if you want to generate some fields, data, and all. So that is how Now Assist is configured in ServiceNow. And then we have agentic workflows as well in ServiceNow, like for ITSM, for CRM, CSM, and other modules. And for CMDB, also, we have many, AI agents currently available in ServiceNow. How we can use them? How can we go and test them? I'll discuss about that as well. Because in your personal developer instances right now, this Now Assist and agentic workflows are not present. So now assist for CMDB. Right now, we will be talking primarily in detail about two skills which are there, CI summarization and manage duplicate CIs. ServiceNow has many other, skills which are there and many other AI agents, but we'll be particularly talking about these two because these are available in our now learning instances. Okay? So first of all, you have to go and activate these skills in your ServiceNow instances to start using them. The first one is CI summarization, and the second one is duplicate CIs. I will go ahead and show you in a second how to activate these as well. And let's talk about the first one, the CI summarization. So this is a skill where you can summarize, your CI. So if someone is coming on to a CI record like Abi said, it's just a record. But if somebody comes on to that record and wants to look at the whole picture of that record. Like, who's the owner? What's the class? Discovery source? Are there any related incidents, changes, or any vulnerabilities as well if they are configured on your system? And if there are any, they are actually displayed in a summary on the top. Let's go to ServiceNow and, first of all, see how can we actually activate these features or these skills in ServiceNow. So you can go to all and just type now assist here, and go to the now assist admin console. So I'll just click on this. And before I do that, if you would have noticed the u r URL here, it's a now learning in sense. So like I told you, it's not available in your PDI, so you have to rely on your customer service instances to learn them. Or you can go to now learning, and then actually opt for this particular course, which is called analysis for CMDB and deduplication templates overview. Procure one ServiceNow instance, and then start using that. Let's keep it as a secret. Okay. After that, you just go in actually and, first of all, here you can see a now assist panel on the top. So this is where you will actually chat with ServiceNow, Now LLM. And this is, by default, not activated in your ServiceNow system. So you will have to click on now assist experiences and just turn it from here. So you have to just turn on your Now assist panel from this particular button. So it will, be actually turned on at that time. Secondly, now if you have to activate the CMDB skills, so you go to analysis feature. So we first went to analysis admin homepage, and then we have now clicked on now assist features. Right now, if you see these, skills are not available, these skills were, discussed in my previous podcast with now then. If you want to go and check out those, please go, and I think, Saima will provide you the URLs as well. Right now, we will be discussing about CMDB. And here, if you see, there are two skills which are available and both are active. Out of the box, when you when you take this instance, these will be deactivated. So you will have to click off sorry. Click on view details and act activate these one by one. So like we talked about CI summarization, it's basically for summarizing your CI. And, unfortunately, in this ServiceNow learning instance, the skill is not working as expected. So if I go into one of the Windows server and go to one of the servers here, this generative AI feature or the analysis feature shows up like this. And, if you click on the summarize button, it should summarize the whole CI record for you and if there are any incidents, who is the owner, and everything. But, unfortunately, it's not working because it might be a now learning instance thing. But I have taken a screenshot for you. If you look at the screenshot, this is how your summary will look like. So the CI is an AIX server, so it will tell you the class of that server. When was it last discovered? Who is owning the server? So it is owned by capacity management. And then are there any incidents or changes related to this? So it says there are three changes related to this, and all of them are three moderate priority. So whenever you are look c I or any other part CI, so you can handy for anyone who wants to take a quick overview of your these CIs. Now let's talk about the second, c, feature, which is manage duplicate CIs. So, Abby, said, like, there are many issues, you know, for discovering the data. So one of the issues is with dealing with the duplicate CIs. So, for example, if your I IRE rules are not set properly and somebody creates a CI manually, and there is then this person who, creates, through discovery and then through automation, through import sets and all. So there there might be duplicate CIs in your CMDB data, which is quite common. And to deal with these is really important because if the same data is available, on this table, it's really hard to, build the relationships. It's really hard to identify on which particular server or on which particular router the issue is. So it's becomes really important to deal with these duplicate CIs and fix them. Now ServiceNow actually has, deduplication templates here if you see. So these are templates where you will define how to tackle those duplicate servers or duplicate routers. So we will actually, I will go and show you an overview of these templates as well. But through these templates, this now assist will actually fix those duplicates. Before we go there, I want to just, take you through some basic terminologies, and I promise after this, we'll go and see the demo. The first one is deduplicate library. So it's basically a library where you actually, create a class maybe. And under that class, you will have, how you are fixing that class through templates. So for example, I created a library only for servers. Then you can go ahead and create templates for Windows servers. Then you can create another template for Linux servers. Then you can create another for maybe generic servers, which are not classified. Right? And then there would be some duplicate tasks. So these tasks are interestingly created by health dashboard, which I will show you as well. And these duplicate tasks, we will assign them to that template. So if you see this arrow is going up. So we have to assign these duplicate task to the template, and then we will resolve those. Let's go in ServiceNow and look at all this setup, and then we will go and see the demo. So right now, when you have activated the second feature as well, manage duplicate CIs, and you have activated the now assist panel, so everything would be set up. So here, if I go into my favorites, actually, I've, bookmarked this Windows servers, and I have actually, pulled in a field called duplicate off. And here you would see these are the duplicate servers. Now how did ServiceNow identify these duplicates? So you have to actually go to health dashboard. Under class view, you would see here the correctness. Right? So it's it says there are, percentage of correct CEIs is six percent. But now how did we get this number, the correctness? So you go to CMDB health settings. I'm sorry. You, go to management. And then here you would see deduplication dashboard. So here, all of those duplicate dashboards and the templates are here. But here, if we go in those settings where we had gone, I click on schedule jobs. Here, when you run this particular job, the CMDB held dashboard correctness score calculation. So these are the jobs which are already written by ServiceNow. You have to go and just execute them, or you can schedule them. And after that, all this data is automatically updated on this CMDB health dashboard. So I'll go to that here. So after running each job for completeness, compliance, or correctness, these would be populated. So these are the three c a three c's of your CMDB. Now once you have found out all the duplicates which are there so, right now here, you would see there are, 18 Windows servers which are duplicate. And if I go in and look at those templates so for looking at those templates, what we need to do is, again, we click on this management. We go to deduplication, and here we have created a template. Now, looking at the time, I will not go through how to create a template, but they are really easy to create. You have to just give, like, which class do you want to create template for, and what would be the action, and what do you want to do with those duplicate servers? For example, that job identified all the duplicate CIs. Now how to tackle those? Which one to keep? So if there are, for example, two servers which are identit identical or maybe you can say same, which one to delete? So you can go and define in this template you want to keep the oldest one and delete the other one, or maybe you want to archive the other one. So there are these options in this template. Now let's go to now assist panel here. And here you would see these two options create, KB article and then manage duplicate CIs. So if you are actually working in a ServiceNow instance at your clients, there could be many more options of agentic workflows coming up. Right now, we had activated only one manage duplicate CIs, and that's the reason you see only this option here. Now when I click on this manage duplicate CIs, here, it would say I have found nine tasks for Windows Server. And for computer class, I have found five tasks. So it says it has found 14 unresolved deduplication tasks. So these duplicate tasks were created by that job which we had run. And now, we this agentic workflow is asking you which one to fix. So we have created this template for only Windows, not for computer. So we will be going ahead and fixing our Windows Server. And one more thing I wanted to keep it open, actually. So we have an option here. After we selected Windows Server, now it's asking me, do you want to go and see those Windows Server tasks? Let's go ahead and see them. So here you would see all the duplicate tasks for Windows Server. Now you have the other option to create a new template, but, like I said, we already have created a template. So it will ask me. You can go ahead and look at existing templates or view more options. So we will actually go and look at existing templates, and it found one which we had created. So this is a custom template which I have created. And in this template, what I have done is the records which are older, we will keep them. And the newer records, we will go ahead and delete them. Now when I click on this, it says, okay. Here is a link to review those template or you want to run template for all tasks. So for all these tasks, if we want to run this template, we can click on this, and the remediation will start. So now it has said that it has started remediating that. So if you want to go and look at that so you can actually go into home. I will just, unpin this for a second. I'll go to CMD workspace, health settings. I think I missed that option. Sorry. Management, duplication dashboard, and running. So you see here, after we went to that deduplication, dashboard here, we are seeing there is a job running to remove those nine servers which are duplicate, and it's still running. It might take a minute or so. But if I refresh if I go to complete it, it's still not complete. So once it completes, all of the data which is duplicate should be deleted from the system. So I will just give it a minute. I know we are running out of time. Still running. It says everything is ready. And during demos, it's really slow. But, yeah, it worked. Okay. I'll go to complete it now. Still not here. I think it's just completing everything. We'll go to Windows servers. I will refresh this. Right now, there are 29 records. So there are now 20. So duplicates have been removed. So I think this will update maybe in a second or so. But right now, if you would have seen all the duplicate servers were removed from the system using this agentic workflow. So if you go ahead and create templates, for your system, all your, service desk people or the people who are responsible for CMDB, they can just go and remediate those themselves. What improvement, actually, I am looking for here is, like, ServiceNow, should build an agentic workflow which should create its own templates suggesting us. So maybe down the line, that's one more thing which I would like to see in ServiceNow. And if you think there are more things, with ServiceNow should do, please type in the comments, and maybe I will take it to them or just go ahead and put them in the idea portal as well. One last slide which I wanted to show you here is if I go here so there are few AI agents which are currently available in in service now. There are more to this. So this is just, the tip of the ice book. There are many other AI agents working behind the scenes. So one is CMD, principal class manager. So this AI agent helps you identify which class should be kept as principal or which class should not be created as or should not be a principal class. So this AI agent suggests you that. Similarly, we have data manager to manage your data well, data ownership manager, health metrics, life cycle manager, and CMDB search. So the last one, actually, I love the last one where you can actually give you in your natural language and say, I want to find the CI. And if it's not there, it will ask you, do you want to go ahead and create that CI or not? So using those AI agents, you can actually go and create your CIs just through natural language. And I have seen many use cases for OT devices, operational technology devices, and you can go ahead and create those CIs in your ServiceNow CMDB using this AI agents. I think I'm done with this, and I will let Saima take over. And please let me know in the comments if you have any questions. Amazing. Thank you, Hardi. Thank you, Abhijit. We did have a couple of questions come through. The first one is how does it handle partial duplicates where only some attributes match? So duplicates are identified through the, IRE rules. So when you are running that out of the box job, it goes through your IRE rules, which Abby had talked about, and it will find out those duplicates for you automatically. Great. Thank you. And then someone Rocky here is asking, can you show us how the template looks? Oh, okay. Yeah. I if we can I think we've have time little bit, for yeah, we've got a little bit of. time. Okay. Yeah. Let me go ahead and, go there. So we I will first, if you remember, we went to this, library. First of all, we went to management, and we were in CMDB workspace. We went to management. We went to deduplication dashboard, and then, we went to view template libraries. All the library. So I like I told you, library is on the top and under it's just creating a name of the library. Nothing else. It's just a container. And under this, we will create we can create multiple templates. So I clicked on this create template button to create a new one, and then I had given this name. And then for which particular class you want to create that template. And after that, if you want to put any conditions in there, I did not put any. And then like I told you before, running that agentic workflow, which CI should be create considered as main CI. So I had, mentioned here the oldest created. So there is an option of the latest also. So you can go ahead and, choose that or maybe the latest discovered one. So multiple options here as well. And then you have an option to merge the attributes as well. So I just selected one attribute or one field to merge was manufacturer. Then you have the like, you can again choose from the latest record or the oldest record. Multiple options. Then the relationships. So if there are any relations like other related softwares, other related CIs, so do what do you want to do, with those, like, merge all relationships and do not merge any related items. So depends again on your, requirement. And lastly, if you want to delete that so I had, actually selected delete to, delete the duplicate CI. And when you create this template, you have to publish this template. And after publishing, it would be appearing here in your agentic workflow as well. I hope that was helpful. Fantastic. Thank you. We just have two more questions, and, then I think we'll wrap up the session. The first one is, how do AI agents differ from traditional CMDB reconciliation and identification rules? Abby, you want to take this up, or should I? That's fine. I I pretty much know about the traditional methods of of doing it. Okay. The good part here is what I see as as a differentiator here is, other than writing your scheduled jobs before where these jobs were running at one particular time for deduplication tasks and everything, these AI agents can do it on the fly. The moment your new CIs are created, the agent tech AI can pick this up a moment. So imagine. your scheduled job or your traditional ways for deduplication task is running around 5PM in the evening. But it's when you get to know about the duplicate CIs, imagine there is a ticket created for that CI already, Intent associated with that already. That's that's where it creates a problem. However, this agent take AI, you can trigger it pretty much on the fly and saying that, okay. There is a duplicate CI. Take this action on it immediately. Remove this one associated with this particular CI. Delete the duplicate. tasks immediately. Assign it immediately to the ticket. Replace it with your oldest CI. That is the key difference. I find it. But, yes, job, it pretty much does the same thing. Gets your serial number compare comparison with what you have. And that's how it identifies duplicate anyway. And rather than running it on one particular time, it just does it on the fly immediately. That, I believe, is the key changer. But, yes, Hardeep, Right. would you want to add anything to it? Yeah. Right. So I think, the one other difference would be your AI agents will keep on learning from your data. They will keep on learning from your architecture, how you have configured your CIs. And according to that, it will keep suggesting you new things as well. So if you look at traditional things, they are, like, predefined, and they will happen in a sequence like a in a flow designer. But these AI agents will learn, and then they will suggest you more things. I think that would be a game changer. Another part, just on the same question, would like to add here is about understanding the trends that your traditional ways of finding it won't understand. Imagine you are getting frequently duplicate CIs coming in from your one particular data center. Something that your traditional way of finding it a duplicate will not be able to understand that. Okay? It just tells you these are duplicates coming in, but agentic AI will tell you that your duplicate major of the duplicates, CIs, are coming frequently from your x y z location or x y z CI class. That will help that will give you a much more idea. Okay. This is where I need to look at. So it gives you a better pinpointed direction about where to look for for the issues and to avoid those. So, basically, it's it identifies a trend as well, which your traditional duplication task cannot think. It it cannot do that. So it basically eliminates one of the jobs for you. Amazing. Thank you. So we are overrunning by, quite a bit, and I don't wanna keep, everyone for too long. I know there were a couple of questions that we didn't get around to. So, afterwards, feel free to, connect with, Abhijit and Hadi and, yeah, carry on the conversation, afterwards. But, yeah, thank you so much for your time, everyone. We will be sending out the recording in the next couple of days. It usually takes up to two days, so you should receive that in the email that you use to sign up for this session. But, yeah, thank you all for, taking the time to join us. And, Abhijit and Hardeep, thank you once again for, leading on the session as well. Thank you, everyone, and thanks, Ima, for inviting. No worries. you so much, a. lot. It's just lovely meeting you all. Nice questions. Nice interactions at the time. Thank you so much. Thank you. Bye bye. Bye. Bye bye. Thanks. Bye.