hello and good morning once again and welcome to the AI Academy we'll be waiting just about a minute here while we get more attendees in but thank you for joining us we have a really exciting Academy for you today focusing in on analysis and Knowledge Management so I hope you'll excited to see this new feature cool so I think think we have a good quum of people so we'll get started my slides want to change one second apologies one seconds okay awesome so you guys have probably seen this quite a bit um just anything that we say here today is subject to change so please don't make any um and we might be making some forward-looking statements so please please do not make any purchasing decisions based on what we say here today based on those future looking board looking statement so um keep that in mind while we go through this material and we have some resources available for you beyond just this AI Academy so you have the Community page you've got community focus on Knowledge Management which is the topic which we'll be covering today and we also have the service now Community YouTube Hub so this video in addition to all our prior aiy are available up on that YouTube page within a playlist just called AI Academy surprisingly um so please go check that out if you haven't already have some really good stuff there so just some Basics on the AI Academy so these are created for you so we really try and Target go a little bit deeper into the products that we are releasing or have already released and give you the Practical guidance on how to use them we are recording the session and we will be posting it up on that YouTube um playlist that I mentioned just prior and you should notice that you have access to the Q&A panel within the zoom webinar so please at any time feel free to ask any questions um you have a couple of us hanging in the Rings here to answer those questions um while Ashley go through her presentation another day we'll be spending just around half just about half an hour on this to go through an overview go to an exercise and then after the exercise we actually have some suggested guidance on how to actually leverage um now syis Knowledge Management afterwards just however much time we get left we'll have make sure to have out some time for Q&A as well and then with that said I'll hand over to Ashley to show you guys what knowledge man now system Knowledge Management looks like pretty exciting awesome thank you Eliza I'm going to go ahead and just switch screen sharing here I did want to introduce myself um so I'm Ashley Snider I'm the outbound product manager for Knowledge Management uh and now assist in Knowledge Management um so we do have a Knowledge Management Academy as well which I'll go ahead and plug um so if you're looking for Knowledge Management things that aren't really gen you want to know kind of more about processes and kind of core knowledge management we have that Academy for you on the community as well uh and we'll also be talking about some gen stuff in there but mainly kind of housing all of our generative Ai and AI content within the AI Academy so thank you Eliza and Victor for having me on today um so let's go ahead and get started here um I do want to kind of talk about um kind of what you know problems and issues we were targeting for the first release um with that being said um we are releasing version one of now assisted Knowledge Management in in May so right now we are slated for the May store release so if you are looking for demos if you're looking to get more information about Now assistant Knowledge Management from your account teams uh if you're going to knowledge 24 we'll have all that information there for you but if you're not going to knowledge 24 right after knowledge 24 um we'll definitely have all this information for you online so I'm really exciting uh release and exciting time so I want to kind of talk about the problems that we started to solve for in the first release um we're really focusing on the agent experience itself um we know that agents are continuously multitasking we they don't really have the time to create knowledge articles for the cases and the incidents they resolved despite their best efforts you know they are continuously taking those chats and those calls and those incidents and cases that are coming in and a lot of the times Knowledge Management knowledge articles are secondary and kind of Fall by the wayside um we also wanted to focus on kind of the agent experience of having to start from a blank slate especially when you have like a really long and complex case or incident where do you even begin kind of summarizing the information and putting that information into a knowledge article that can be a little overwhelming at times uh and then also ultimately the lack of knowledge articles um for self-service and agents right we have our agent um issue where they can't find articles they need to troubleshoot issues so that kind of leads to longer resolution times it leads to kind of unnecessary escalations at times and then you have your employees and your customers those are kind of using your portals and your self-service Solutions um they can't find what they need either on those self-service channels just due to those knowledge gaps and those lack of knowledge articles um so this is what we're targeting for uh May and for version one these kind of problems so uh quick kind of benefit slide here you know what can you really kind of expect for the outcomes and what can you expect to see of Now assistant Knowledge Management of course we want to have that faster turnaround time for article generation we want to bring in relevant task kind of content and context into those knowledge articles so that information doesn't get missed uh we want increased self-service and deflection with the availability of knowledge articles and of course we want that faster resolution time for your agents uh we want to reduce kind of redundant troubleshooting and then also kind of reduce those unnecessary preventable escalations that are out there um to kind of get into some data um you know we get a lot of these questions already um from different customers and different account teams so I want to kind of show you the the data that we're targeting for the main release um so for version one we are gathering context from Fields such as short description description resolution or close notes and then of course our Journal field so our work notes and our additional com uh comments I will mention um if you have resolution or closed notes present when you generate the article you're going to get a better version of that article and you're going to reduce the risk of hallucinations and kind of madeup resolutions so you can generate articles that aren't um you know in the closed or resolved state but you do kind of run that risk of a a hallucination uh and for version one once you draft an article from a case or an incident or an HR case uh the the ability to update it using generative AI is isn't present um for version one that's a road map item for us so you know the more data that you can have in your cases and incidents the better quality article you're going to get out uh so in regards to the llm itself uh we are using our domain specific now llm so kind of what this means you know just like our other now assist products that are out there this is our proprietary llm it's trained and fine-tuned for our service now use cases such as generating articles from task based records um you know what this means is that we take over a lot of the development and the tuning that you would normally have to do if you go on and perform a Google search or chat GPT search you're going to get a list of things that you're going to have to do from picking an llm to doing all that tuning and testing and things like that we take care of a lot of that for you and we do it on service now code um text and anonymize customer data we also optimize the llm for our architecture so our nowat form architecture our security and our data model so a lot of that heavy lifting that way you really only have to worry about kind of the outcomes um of using our analysis products so I know we have a lot of information on the AI Academy about our llms there's a lot of information out on the community so if you want to dig deeper on that topic definitely go to the AI product Hub that Eliza mentioned earlier so let's just um before we get into the demo let's just cover um what we have coming out in version one in the May store release um so for the May store release you can generate one article from one case incident or HR case so we have that onetoone relationship an article must be anchored um or attached to a task record meaning that we don't have the capability for this release to create freestanding articles from other sources such as attachments documents or external sources for now um so just know kind of we are looking at that one toone relationship between a task based record and an article for May um we are supporting Now assistant itsm CSM and HR Service delivery for may we will expand in future releases um we are using the standard template which a lot a lot of our customers use and have been using and the KCs template so if you're a KCs program type customer you've enabled the KCs plugins and you use that template you'll continue to use it and we do we do plan on supporting more out of the box and custom templates in a future release um so article generation it does support your current um permissions and your can contribute user criteria on your knowledge bases so you don't have to worry about who can create articles we're respecting all of the user criteria that you have put in place and we're following your current um governance process um so as mentioned these are drafts so the article is not in a published date we're following your current workflows and approval processes per your knowledge bases so another thing that um you don't have to worry about it is in that draft state but I will kind of get into governance in just a little bit um other than that we did release a new three column layout for article editing and workspaces and I'll show this in the demo the classic environment form is also supported so if you haven't moved over to workspaces yet you can still use this functionality in the classic environment when you're working on incidents cases or HR cases uh and then also we do support um generation we call it in product and I'll show this in the demo which is a button or UI action in the workspace U and Analysis panel so you can do this um either or uh and I do have some caveats on kind of the knowledge bases that we'll cover in the demo here I know we have some questions rolling in let me get into the demo let me see if that answers some and then we can kind of get back to the those uh once we go into our suggested guidance so let me go ahead and share my instance here all right um so what we're going to do I want to show you kind of what the setup looks like and what you can expect to see in the setup just like our other analysis products um everything's done through the nysis admin console um so once you know you installed analysis in Knowledge Management plug-in through application manager you're ready to set up the skill um you'll go into the analysis admin console and go into analysis features U I'll show you for kind of CSM and customer workflows but all of the workflows are kind of the same it's very similar that's why we we kind of set up the analysis admin console to be very similar despite kind of what workflow that you're using so you'll see the skill here um alongside your other skills that that you're used to seeing such as case summarization or resolution notes generation and you'll see KB generation for now as we add on more knowledge skills they'll be underneath kind of that card and that layout so I have this one activated for our demo um kind of some things to tell you uh you know we do tell you what llm that we're using I know our creator analysis products um have some different llms out there but if you're if you have questions on which llm we're using we display it to you right here so setting this up the same as other analysis skills we really bring you through kind of this guided setup process where you choose a few things um review a few things and then you are good to go um so kind of version one what you can expect let's say you were setting this up for CSM and cases um we do have some guard rails here um that we put in place for version one so you can kind of see these T these fields are read only but you can see the the table that we're using you can see the input fields that we're using one thing I want to note right here default knowledge base for nap so the now assist panel it can only use one knowledge base right now so if you have agents who are using the panel this is kind of where you'll set up the knowledge base for where those articles are created if you're using it in products if you're using those buttons in a workspace or in the classic environment you're going to see a modal window and we'll see this in just a few minutes where you can actually select a knowledge base so something to kind of keep in mind when you're rolling this out and when you're thinking about how you want agents to use this if they're using the analysis panel up here that all those articles are going to go into the knowledge base that you select here uh and again it does respect those user criteria permissions as well um so availability um you can choose here what states and other criteria you want this functionality to be available on so out of the box right we're going to give you resolved enclosed cases and incidents you're going to get a lot better articles if you have those resolution notes in you can change this if you need to you can put in different states if you need to um just you know be kind of aware of you know the content that you have in your cases and incidents and kind of the outcomes that you'd expect to see but you can also add more conditions in here as well so if you have more granular things um that you want to to kind of control this on you can do that um or you can just say it's always available and this will kind of you know blow away any kind of conditions and criteria that you have so I'm going to leave it out of the box in default for today and here's is where we display um this functionality so in product like I mentioned in our workspaces with our buttons uh or now ass this panel um just like other products you can get more granular in the roles that you want to see this and I'll kind of mention this this is really kind of where you're going to control how you roll this out to agents I know we have a lot lot of concerns about now all my agents can generate so many articles you know how am I going to handle all these articles this is how you're really going to do that kind of phase and wave approach if you want to do that you would just create kind of a custom role and then apply that to a group and then onboard your agents um you know within that kind of phased approach so that's where you'll do that here um I'm just going to leave this out of the box for now and then we just kind of review and activate here we don't have too much to review since we're on version one I know some other skills get a lot more granular and have a lot more options now that you might need to review a little bit and then we're done so um I'm going to show you just a few different kind of use cases that we have um setup is pretty easy let's go ahead and go into our CSM configurable workspace um for the sake of time I'll just remain as admin but you know um you know your CSM agent would be the one who could see this um per the kind of the default yeah criteria that we have in so I have this case here it is in an open State hasn't been closed yet um don't mind this demo data here but as we can see we have a lot of activity on the case itself we have a lot of kind of back and forth on the case itself and some information and we really kind of want to capture that um within the article one thing I want to mention is if I go and and the case isn't in resolved or closed per kind of the the setup that we had that we just saw I hit create knowledge it's going to kind of take me through the flow where I can choose my knowledge base I can choose my article template and again this is all going to check your user criteria on the background and I can hit create article so this didn't give me anything this is what I expect to see this is what I want this is not using generative AI because we put those controls in place to where generative Ai and now assisted Knowledge Management isn't going to kick in until that case is in a closed or resolved state so if you're using our out of the box kind of controls this is kind of the behavior that you can expect to see and kind of that blank draft that your agents are starting from when they are um generating knowledge articles so let's go back in I'm going to just do a few things I'm going to use my analysis panel here I'm going to grab some resolution notes to make it easy for this demo uh and wrap up this case and then we'll try and create a knowledge article after that so I'm going to go ahead and just pop these in get my uh resolution code here and we'll close this case all right so I can see that my case has been closed and I'm going to do this in products so this is kind of the in product flow you just sell the now's panel flow I'll show you that on an incident but in product I'll hit create knowledge same kind of thing I can choose my knowledge base here so that differs from the panel and I can choose um the template like I mentioned it's either KS or standard so you're going to see if you select one of these it's going to default to standard on the back end one important thing to kind of note until we expand support I'm going to hit create article and I do see a question I can go ahead and answer um hrsd it will be available for hrsd in May so now I see this modal window because my case was um resolved and closed and now I can see yes I want a draft of now assist I get some information as an agent about it that you know the better my data is the better quality articles I'm going to get out so let's go ahead and draft this with now assist and now we can see that we have an article generated and it's grabbing a lot of context from the case itself giving me that kind of first draft I mean there's some things that I'll probably want to edit out um as an agent like a customer recently upgraded I'll probably want to change that kind of wording um I do get kind of my resolution and some nice bullet uh or a nice ordered list here that makes the article really kind of readable uh we're using you know correct grammar and correct spelling and I don't have to worry about typos here so this really kind of gives me that really good first draft as an agent I'm definitely going to want to take a moment to look over this ensure that everything is create uh is is correct and ensure that the resolution is correct before I put it through any kind of approval processes or publishing processes so this is kind of where that human intervention is still needed but I have a really good first start here based upon all the context that I have in the in the case so you can imagine if you have one of those really lengthy cases how it's going to provide a lot of great information for you here um so as far as um questions let me go ahead and get into some um you can change the knowledge spage afterwards again it's just like you know once we generate the article now we're kind of getting back into our knowledge management processes and flows here um so you can go into the metadata uh and change the knowledge base if you want to templates um so if you want to use the KCs template those plugins do have to be active um as far as the standard template that one's kind of of different um than the other templates in the advanced plugin uh even if you don't have the advanced plug-in turned on you can still use that standard template and we kind of default to it that's why we kind of made that decision because we couldn't expect all of our customers to have templates in advaned turn on so that's that's why you have that either that either or right that standard template versus KCs template so no you don't have to really kind of worry about activating all these templates and getting them set up until we kind of support more templates that are out of the box and more custom templates we're just going to default you to that standard template that's kind of built into Knowledge Management core itself um so with that we can go ahead I know we're getting getting low on time um same kind of flow for Incident Management uh and then also if you wanted to use the now assist um panel uh I was going to show that we're getting low on time so we we'll cover that in another webinar but you could create a prompt to generate an article it's going to give you kind of uh the article number here and a link to pop out to the article uh and put it in that knowledge space that you specified so that's what it would look like from the analysis panel uh I do want to pop out of here and just kind of cover a few things before uh we kind of wrap up and uh we will have kind of a longer session of this in the Knowledge Management Academy for May so we'll get a little bit deeper into all this but I wanted to to show you version one kind of what it looks like and what you can expect to see um so a few things before we go um you know you're going to see these slides in the PDF so I want to to kind of show you them and kind of touch on them but I won't get deep into them for time's sake but before you implement um really kind of think about the problems that you want to solve with now assist in Knowledge Management what outcomes you're expecting to see you may have already had these outcomes defined for your current Knowledge Management program uh and you'll just be kind of comparing ing them against Now using generative AI uh and kind of see you know what what those benchmarks and and what those metrics and targets look like versus um manual creation and again how will you measure success so you're probably already measuring success currently for your Knowledge Management program and just to start thinking about what kind of outcomes you want to measure how are you going to measure your generative AI articles versus your manual articles how will you kind of communicate that value um you know once you roll out nowal assis and Knowledge Management to your leadership and your stakeholders uh and kind of go ahead and start establishing some baseline so if you know you're going to purchase this and start rolling it out go ahead and get those Baseline and those metric numbers for your manual creation so you kind of have that that beforehand kind of information and then after you implement you can start seeing kind of what's happening with generative Ai and what outcomes that you're getting um so a couple of things on guidance um I won't get too much into all the wording here due to time but so knowledge quality is now tied to case quality so as you saw in the demo the better your documentation on your cases incidents and HR cases are now the better article drafts you're going to get out so something to start thinking about now is your case quality assurance process kind of what quality measures do you have in place how do you handle documentation how do you ensure quality documentation um or if you don't have these in place kind of what processes can you start putting in place that way you don't have to worry about you you've made this purchase for now assist I'm not getting quality articles and that's due to my case quality and my documentation and you're having to kind of bump up both at the same time so really good time to start thinking about your your documentation on your task um for incidents cases and HR cases good time to start training agents on documentation quality and how they can better document within um these tasks that way you can get set up and then also look into some features that can help so there's case templates incident templates HR templates that are out there um there is a a coaching module that's out there that if you're licensed for if you're now wanting to look into quality assurance and you want to start coaching we have that coaching module that can be out there and be used for this too um so again good time to contact your account team if you're thinking about this say what kind of features do you have in place to help me with this um and get ready for this um another thing is that you know AI is a great assistant but you're going to need that human verification so you're going to need someone checking those articles um you want to make sure that um he's going through any kind of review processes that you have if you're an instant published kind of shop kind of think about that now how are you going to review these articles if they're being instantly published that really would mean that you need to train your agents and make sure that they're aware that before they publish these they really need to kind of go through them and make sure that they're adhering to kind of your content standards and the kind of information that you expect to be in them before they publish them um communicating value right communicate the value of um what's in it for agents when they create these articles get them excited about it it's going to lessen um hopefully a lot of those cases that are more simple and and unnecessarily timec consuming but it's it's also going to um bolster your self-service um resolutions as well so freeing up agents to kind of do more meaningful work freeing up agents to kind of tackle those more complex cases that are out there and then feedback we have a lot of feedback mechanisms built into Knowledge Management feedback is a gift um whether it's coming from your agents your employees and your customers so make sure those feedback mechanisms um are in place for your articles um on your portals and in within your workspaces and classic and then make sure that you have kind of um process in place to respond to that feedback and I think we are going to go a little bit over I promise you I I'll keep you five minutes over um so the next one is to kind of um think about how you want to roll this out we do have that flexibility to scale like I showed you in the analysis panel so you don't have to turn this on for all your agents at once you can start doing pilot groups um you can start training your agents and your pilot groups on the the type of content that you expect to see in your articles if you haven't already and then you can kind of scale out this on your timeline with being able to give out those roles within groups so you can start working on your timeline now um if you're thinking about purchasing soon do you want to start rolling this out every three months every three weeks what kind of really makes sense to your support teams in your organization what makes sense for the training that you have in place uh and then also what you need to do regarding Lessons Learned so again you don't have to turn this on for everyone all at once you do have that flexibility to have that control in there and then finally um focusing on governance so I'm being a little tongue and cheek about the floodgates because I do hear that a lot but we do have those processes to control those floodgates but you really do want to evaluate your governance um AI is kind of a mirror to your processes and your data quality um you know it will expose any kind of issues that you have in those um so really kind of think about your governance think about how your handling duplicates right now think about how you plan to handle duplicates in the future as it becomes easier for agents to create articles uh Safe Harbor we are looking at um features to help you with duplicates we just don't have it for version one so if you are getting on version one think about the duplicate management um also focus on relevance and validity so we want to make sure that our articles are up to date um AI search really Keys a lot on freshness and other kind of relevant um fields in there and we have some documentation on the community about that so you want to make sure that these articles are up to date especially as you know we don't want generative AI creating a bunch of Articles and then they kind of just get stale um so this is really kind of where your valid two dates come into play uh and then they should really kind of drive your review processes so review processes how often do you review your articles how is that handled do you have strategy in place to review more higher value and higher traffic articles versus lower value and low or traffic articles just some things to kind of think about with that um I see we have a few questions let me cover this topic you and then I'll see if I can hit these questions and then I have a survey for y'all um so are there um as far as language I thought we had a question about Spanish so now assist does not support um any languages other than English as of the May release as of this time um so we don't support non-english cases as well so if you try to create an article from like a case that in Spanish we don't have any guard rails for you you you can definitely click that button and try to do it we can't guarantee the outcome of that article uh and that's something that we won't support with a case so just know that we are only kind of supporting English as a language right now uh we also don't have word count guard rails so some of our products do uh we're not placing those guard rails on you so we are not going to check for word count before creating an article for version one um external content not for version one but it is on the road map those custom templates we have it on the road map not for version one um knowledge blocks so knowledge blocks currently aren't supported in AI search indexing so we're not going to generate article or knowledge box for you right now we kind of have to wait for knowledge blocks to um be indexed and safe harbor that is for Yokohama and then we'll kind of put it on our road map to start doing generation for knowledge blocks so um kind of keep an eye on that I know we have a lot of information about that um coming out I do get that question a lot and then adjustment controls so we are looking at adjustment controls such as tone length rephrasing all the adjustment controls you're used to seeing in other products um Safe Harbor we're looking at this for the August release so a lot of great things coming out soon um I want to throw this survey up uh I do have some kind of questions about what you'd like to see I kind of have some ranking um of different features you'd like to see for August and so on so if you have time please fill out this survey uh and then I know we're a little bit over let me just hit these two questions I think the article templates one we covered and then going to 1kb we covered so we're good on questions so I went a little overtime I appreciate everyone uh hanging out with me and uh following along and we'll definitely have some more kind of deep Dives on this after our da release but uh thank you so much uh thank you Eliza thank you Victor um for your time today and getting this set up up and coordinated thank you so much have good rest of your day everyone thank you