okay everyone thank you for joining us today for let's boost your AI search so we're going to give you a bit of an intro to AI search um and hopefully give you some tips on on how to improve it as well so just as usual Safe Harbor notice here uh in case we do talk about anything that's coming in the future just be aware that things may change this is part of a live on servicenow interactive Series so if you would like to join other sessions J2 has posted a link into the chat you can go to the registration page and see upcoming events so a little bit of house key being we do have some time for Q a at the end but uh please feel free to use the Q a function in Zoom or I believe you can all talk so please feel free to jump in if you have any questions this presentation will be recorded and it's going to be shared on the servicenow community and after the event you'll show a feedback survey if you can provide any feedback that would be great foreign success manager for iitsm and devops responsible for the apj region so my details there you can contact me if you have any further questions so just to get an idea of where everybody's at with AI search I'm just going to run a quick poll watching that now so just give us an idea of you know whether you're actually using AI search and which parts of our search are you using it for whether you're still on Xing search or the the old servicenow search okay nine out of ten one more result 10 out of 10 perfect poll okay so it looks like 50 of you are still using Zing contextual search um okay so hopefully we'll give you some tips on how to get started with AI search today most of you are using it on the of the people that are using it using it on employee Center and virtual agent that's good to see uh not many people using it in our mobile some people using it in global search and I've got one person that doesn't know the difference okay so we'll hopefully educate you that on on that as well cool let's move on so the reason behind AI search is basically expectations have changed you know it's people are very much expecting information to be available on any device in context at the moment of need that's why Zing search wasn't quite meeting that need and we've introduced AI search so why what does AI search give us it's it's more relevant so this is um this is a search a kpi basically average click position means that they're they're really getting uh you know they're getting the the results in the top two positions so it's a bit better and more relevant results we're also seeing things like 69 click-through rates so and you it's effectively your search results are helping you to deflect cases and incidents by surfacing answers and actions uh immediately when you search and also we've continually improved AI search so now there is a much more reduced effort in setting up AI search from when it was first introduced we've been using it as such at servicenow now and here are some of the figures that we can see that we've experienced within service now so three and a half times faster our searches were taking up to 10 seconds to get us results and now it's about two and a half seconds the uh two point we're measuring uh you know relevancy increases of 2.9 times and IR search gives us the ability to you know have tuning have improved stability in Search and have uh better analytics into the users Journey as well foreign so this is just a little bit of history about search so searches come a long way since the first days of Googling so search was primitive and provided simple keyword searching but it couldn't handle little differences in languages such as you know replays matching the web run and running so things like that have improved to introduce semantic search so with semantic search uh they it's it's a bit more natural language processing so different things around how people ask for things in in natural language that that was catered for more than there was um contextual search or providing providing search results more in context this is sort of where Zing was so if you remember uh I think it's really used today but with where you have incidents and you look for similar instance that sort of contextual search is where Zing came into it and it was more also optimized more around the user's device being used you know things like location as well so that's also part of connection search then came cognitive search which sought to understand really what the user's asking and trying to accomplish instead of returning a list of websites it was uh more about providing things like you know last night's game score for their favorite team or website so trying to understand what the user is after and providing them results and now what we have is uh actionable search which is where AI search falls into it so you will see when we demo to you you use AI such thing you can provide actual results so if you search for something there's a catalog item that's relevant you can straight from the search results go into ordering their catalog item we'll show you that a little bit later so AI search is built on three pillars it's meant to be natural smart and easy so by natural what we mean is you know there's functions like autocomplete uh it starts with really detecting what the customers query intent is so what is it that they're trying to find or solve and then the user things like autocorrect so if you have spelling mistakes there it can detect and handle that it can do things like Cinema expansion I'll show you some of that later as well and you can you can introduce your own synonyms as well to be more relevant to your organization another thing is it supports multilingual searches so that's something that we didn't really have before now you can you can search multiple languages it's also intended to be smart so machine learning is continuously improving the relevancy scores based on the user's Behavior so you know if a if a result comes up and you can Mark you know the user clicks on the the fifth result the next time that score is improved right so it's continually improving those goals based on what people are actually selecting results are personalized based on the user's metadata and context so you can do things like um promote or boost results for a user based on their location based on the content we can show you some of that a bit later and you can also provide quick answers by extracting Snippets from KB articles foreign I think I also I mentioned before actionable results so we call those genius results you can have that as well finally it's it's all meant to be easy so all this is accomplished by an intuitive admin user interface there's robust search analytics available as well with dashboards to show you metrics about insights of what people are searching for the queries that they're executing whether or not they're finding answers it's a simple process to to really set up and start indexing and providing users with that Banner search experience yeah so initially AI search was more sensitive to self-service and we've now evolved that more into a productivity as well so it is now available you know within our experience um so it's available within Global search and configurable workspaces and is more set to improving productivity so from an architecture standpoint this is sort of what it looks like on the left hand side we've got servicenow data so tables Etc you can also index external content so if you have external content connectors you can you can index things like SharePoint and then service that through servicenow as well so another slide on that but we'll show you that in a second in the middle you've got the the search indexing so all those data sources get indexed and then get presented through various channels or applications like Global search now mobile portal virtual agent workspace search so you have to configure the the applications for where those search results are presented and as I said supporting all of that you've got machine learning natural language understanding and LP and analytics as well so we do have various connectors for AI search so you are able to index external content directly and bring those results into service now so what that means is you know there's no duplication in the service now to other base if you having to create content and and create knowledge articles about maybe things like Microsoft tech support articles which may already exist outside or exist in other places in your organization that uh actually just to mention that those search connectors it's it's licensed separately so please speak to your account team to check whether you're licensed for the search connectors okay so I only search for the next experience so as I said before initially AR search was only for uh the mainly for the self self-service side for the portal so we've now made it easier for you to enable AI search for the whole of the platform or for next experience so there is a store app which you can download and install if you haven't set up AI search yet so look for the AI search for next experience app and basically um you know it has all the The Styling configurations for next experience it makes it easier for you to activate that another store app that you should also pull down as well if you haven't done so already is AI search management tools and this is the app that gives you you know insights into and the analytics into people search experiences and what people are searching for and the queries are performing Etc foreign labeling so this is an innovation Labs app at the moment it's it's there to help you through shooting AI search and also helps with migration from zinc to AI search as well so if you haven't migrated yet you can you can look for this out and it will help I believe there's a guy that's set up that helps you walk through you know migrating uh zigging to her as such so a couple of things that when you're in Utah before we move into the demo so facets fast as enhanced fencing cross table support yep so there's basically some enhancements to you know the filters that are on the left hand side there uh we now support searching across multiple tables and and multi multi-value field facets as well foreign um it has next experience search uh oh okay so this is um yeah so there's some improvements to uh the next experience search as well that was introduced in uh Utah so uh and I'll I will demo some of this I'll demo the flow but basically you know adding on to the architecture that I showed you before you've got the is you first have to set up the index sources so which tables uh is is is being indexed not aware of that where is the data um you sort of filter and limit the search content so for example you might be looking at just um active knowledge base articles right so not the retired ones so that's where the search sources are you're creating that filtering and then you attach that to a profile and the profile gets used on a Search application so uh there are various places where you can enable our search and you don't have to enable the perform for all of them that's basically the um the setup I think what I'll do now is jump over to a demo so the first thing if you haven't enabled data search already is to check that is it is enabled on your instance so right now um well with Utah I believe uh and maybe even before Utah but in new instances I'll have ai search enabled by default so if but if you're a if you have an older instance that hasn't been enabled yet yeah you you need to request it and you can check the AI search status and it should tell you whether your instance is AR search ready uh so if you see if you see that it's not ready you'll need to request uh that through support if it is ready uh the the reason I I search um guided setup here so this kind of takes you through that setup process on the PowerPoint site that I showed you before so first it's about you know defining searchable content uh what are the index sources and what are the search sources there's a click into one of these uh and click configure it'll take you to all the index sources or the records representing uh that you that you need to set up so the first time that you use AI search you'll probably see this little triangle against all these index sources which means they haven't been indexed yet if you do see that all you have to do is just you know select the ones that haven't been indexed and click on index or tables let's go back um yeah and there's information about the whole setup process here within the guided setup as well for you to read okay so I won't go through the whole guided setup but as I said that's the process is there for two this is here for you uh you know to to pull together all the things that are required to configure AI search uh if you if you have that plugin for the next next experience search uh installed then uh I believe you it's it's pretty easy to get that done so most of the most of the uh sources have been defined for you and you just need to go and check them and then activate them I'm just gonna go now to uh the employee Center just to show you a demo let's have it to do that from here okay so this is employee Center uh this has had an uh search profile attached and they have such activated so if I just do a search some of the features of AI search you can see the different tables that it's searching across here so you can type across to see the results from the different areas and then you've got search facets along here for you to start filtering the results as well foreign tegrated to to um next experience so from the global search year I can type something and what this is doing is it's it's also giving you such suggestions uh and then it's also searching across various you can you can search across uh different workspaces or Global as well so I could search specifically for uh within the service operations workspace moving global um so if I can let's do this so open that it'll open in the workspace and view your results so very similar to the portal you'll see all the different you'll see all the different uh things that it's searching across the top and then filters down the left and you all could also can sort by date or most relevant on this side as well that's basically the searching um I'm going to go into a bit more detail about you know uh setting it up and some of the enhancements and improvements that you can make to it so one thing that's really useful is our search per year so search preview this is one of the utilities that comes with the uh one of those plugins I don't remember which one it was but you if you grab those two plugins you'll you'll get the you'll get the preview uh actually sorry oops and when you're when you're you know setting up and configuring your search this is where you want to be testing your search results um so for example let's search for um so if I search for uh o you'll see that it's only giving me the result for matching David Lu um so if I wanted this to also I want to create a synonym for this one so as out of office so what I can do is I can go to the search profile for employee Center this is the one that I've got tied to by employee Center I've done this before but basically within this search profile uh you've got the search sources you've got the synonyms stop words type of having genius results and results Improvement rules so synonyms is where I can go in and add a synonym I added this before but I I deactivated it it's gonna activate that so every time you make a change it'll tell you that you need to publish it uh before it takes uh before it takes some effect to publish this and then when I go back to my preview just click refresh to make sure it's active um so before when I searched uh triple o it was it wasn't really expanding that synonym it was only giving me results which matched to below but now when I search through below it's giving me it's actually expanded the synonym I can see that here right triple o has been expanded out to out of office and it's giving the reason giving me the result for out of office so this is how you can make search results more relevant for your organization I'm sure there's plenty of uh you know terminology within your organization and acronyms within your organization that you could Define into synonyms to help improve your search results so another way that you can improve your search results is result Improvement rules uh so there's three types of rules one is to promote uh uh sorry one is the Boost one is to promote and the other is to block so I've defined three rules just to show you as an example for this one I don't know what I've defined the Boost rule um I'm actually going to a second what's this thing so this is going this is basically I'm boosting knowledge results so whenever someone types laptop in the short description I'm gonna make sure that um you know the knowledge results are more prevalent um so if I search it here so this one's not active yet laptop the results that I'm getting a new pc request yeah RIT requests so a lot of this is coming from you see the top results are coming from the catalog and then the knowledge base results uh are coming later so I am just going to do something so yeah I um I had this active on on something else before I need to update this to be active on all queries is that we go back to here refresh this when I type this what I'm uh what I should be seeing is I should be seeing knowledge base articles I think it's still it's still matching and still matching um catalog item because it's it's relevancy based so if I remove new uh it's now giving me you know knowledge as the top results so this is all it's basically boosting these scores so the score the scores that you see here are being boosted by this rule what I'm saying is uh if the short description matches laptop you know boost boost the score by a thousand percent um so that's uh you can do things like that so you might be able to boost based on some some use cases that people boost based on this may be the user's location um so you can use context like that to boost search results and so that's just the really simple example another example is you know promoting promoting things so I've just got this set up so that if I type the word promo on it it's going to promote this new high orientation catalog item so no matter what I type no matter what my search is oh let me thinking okay because so do you see you see what they did there is um I wasn't seeing the results that I was expecting because oops so much bye it's actually it's on an auto correct for me so it's searching for laptop promote um so that's not what I want I want to search for laptop promo and now when I search for a laptop program it's matching the keyword that I have to help me promote new eye orientation so whatever it is on my type something else [Music] again it's done the although correct for me but if I you know if it's matchless keyword it's going to provoke this for me new higher orientation that's basically what this promo rule is doing and then the last example I've got for you is just how to block things um so I have this firewall rule change catalog item which I don't want to be made available if the user's context is China so if the user's country is China so if I search for firewall this is the this is the catalog item that um you normally get uh now if I change the location oh actually sorry should be country to China this is what I set it up with and then search that again uh it's no longer coming up that's because I changed that to China let me change it to another country search again I'm getting firewall rule so that's just the way that you can use that to to block certain things um based on different criteria in different contexts so um I hope that demo was useful it's going to jump back to the slides does anyone have any questions about what you've been showing today or any questions about the AI search there's no question in the Q a no question in the chat um just a reminder your all of you attending this webinar you can actually just unmute yourself and ask your question you all have the ability to talk I'm curious if anybody's using you know for those that are using our search are you using any of those capabilities that I was going through today um have you guys got synonyms you've got this drills uh it'd be interesting to hear what everybody is let me just try to use their hands so I think she has a question so go ahead Mimi yes uh thank you uh at the moment we haven't turned on our next experience yet so in that regards um we are not getting all these features or we have you know most of the features the one you demonstrate it they should have most of the features DNA features that you wouldn't be getting is um you wouldn't be able you wouldn't really have this stuff so the globalization and searching the workspace okay um but everything else would be would be available to you this is for the uh this is for my my ese portal search profile so when you install that that plugin you will get this for next experience you'll get this profile next experience search configuration actually let me go there let me go there oops that's the easiest way to get to it is um just to show you set up so search profiles let's find out navigate um yeah you'll get this next experience search configuration here and then all you have to do is you know make sure it's got everything that you want to search um do your usual configurations and then publish well turn it make sure it's active and then publish it so that's when you install that plugin for our next experience search you should get this search profile and most of this setup will be done for you ready sure thank you Alexia um well I'm here might as well show you some of the analytics stuff as well so I don't know if I have any data yeah I probably don't have enough data to show you some of this stuff because this is a demo instance but yeah you'll be able to start delving into you know how your indexes are performing and and what queries are going on cool any other questions comments um experiences that you'd like to share all right this is wrong here and just have a question whether AI searches so fully available for the pdis as well Now search for pdis uh I don't know the answer to that um Natalie do you know I have to check I don't want to say yeah I think yes but I have to check with the AIS all of the AI capability I'm not sure if it's been made available in in PDI can we uh sorry how can we um Alex and I will take it as an action item and we'll make sure that we answer that question uh in the um uh on the community you can check for yourself right so you can go to in your PDR just go to AI such status and check what it says on this page so if it says that it's AI search ready then you're good to go um otherwise I'm pretty sure there'll be a link to a knowledge base dialing or something there that says it's it's it is already it's available for pdis yeah we'll we'll check and we'll post the answer on the community as well thank you foreign I think the other thing to mention is you know AI search with the the establishment of uh with the introduction of generative AI um you know AI searches is a complementary piece to that so it's really important that you start moving to these new technologies so that you can take advantage of all the generative AI stuff as well when that when that gets into their platform in Vancouver other than that if we have uh no other questions or comments I think I will be able to wrap up early I will give you guys your time back and if there's anything else you'd like to see maybe in the next AI search session um please feel free to give us our feedback as well before before we wrap up um say I just posted we're having a number of sessions on to the community that are happening uh during the Asia Pacific hours um so just to note if you want to be in form of what's coming into our roadmap so I'm doing a session on the 18th of July we have the success dashboard on the 20th we are also going to do at the same time uh walk-up experience on the 27th we actually have VA and nlu early August performance analytics and devops change velocity coming up as well so we'd love to see you back in one of these sessions and so make sure you go to our community page and and sign up for these sessions and if there's anything you'd like us to bring to these events just make sure when you provide feedback to let us know would be great to hear you know what what you would like to see in these sessions uh so Rahul I see a hand up uh do you have a question or something you'd like to add um all right I think Alex has already answered my question that it's the AI search is available across workspaces yes yes it's uh yes it's only available so you're from that Global search that I showed you before you can uh once you've been enabled the next experience it will be available for workspace searches as well all right thanks uh there's another similar question about uh about agent workspace so shalesh um yeah so agent workspace so the thing about agent workspace is it's it's now I can't remember if it started deprecation process already but basically yeah so so we're a TI search we're not making it available on the itsm agent we're experience as we've announced the deprecation pass for agent workspace so all customers that are using actively itsm agent workspace we've asked them to we're asking them to start planning their transition to this service operation workspace foreign workspace if you're still on age of workspace thank you so deprecation so that we've we've just announced the deprecation so that goes over so that means uh new customers coming to service now are not going to be able to turn on the plugin um and for existing customers they have three releases to transition to service operational workspace great thanks for those questions um so if there are no other questions basically here's some some resources that you can refer to for AI search so there is an AI intelligence Community for it and from there you can you can get to the AI Academy and within the AI Academy there there's it's run by our product managers and they do it is out of the apj time zone though so it's it's mainly run in the US time zones but the the sessions are recorded and you can watch recordings about um everything AI not just AI search but a lot of things Ai and there's also a bunch of you know Community articles and things there on the community forum for you to access as well if you are just getting started there is an intro a free intro to AI search uh course on now learning um so introduction to learn how to search that one there so have a look for that one and uh you should I think it's pretty quick so it's a beginner's course it's pretty quick goes through most of the stuff that I went through today um definitely get your hands on that one if you're just getting started and then also we have a very detailed implementation guide on now create so if you go to now create and search for search for an asset search for AI search you'll be able to find a really in-depth implementation guide yeah I think it's a it's a big PDF document it's probably about oh I don't know how many pages but it goes through everything in in a little detail sorry if I don't pronounce your name correctly is the recorded session free to access uh uses um so it's a recorded session free to access users so the the session will be recorded and will be posted on the community and I think um we'll actually add the deployment guide as well there so that you have all that information easy to access in one place is that answering your question 10. all right thank you thank you okay good um any other questions cool all right if there's no other questions uh thank you very much for joining us I hope that was uh useful and uh like I said if you know please give us some feedback if there's other things that you'd like to see in AI search or or anything else and for those of you who are still here um we posted the additional sessions so if you want to see what we're doing not just in itsm and part of our Vancouver in Washington releases I'll be delivering the roadmap on the 18th so make sure you sign up for this one love to have you there again so thank you so much for your participation in this webinar today thank you Alex for this wonderful session on AI search and I think it's a wrap thank you thank you thank you thank you