now assist for CSM so gener to AI privileged enough to be able to present parts of this slide a few weeks ago to a group in Orlando at a snug service now is doubling down on Genai so there's a lot coming a lot that I haven't even seen some stuff you know the Safe Harbor statement so there's a lot on the horizon for AI so we should all uh we should all get familiar with it service now has been investing in AI I won't speak to all of these I don't know all the details to all these Acquisitions they've made along the way but as you can see it's pretty linear as I alluded to they're they're not slowing down so they're continuing to invest continuing to buy into generative AI because let's be real it it really is the future it's just kind of the tip of the sword in terms of efficiency and all the good stuff that comes with generative AI generative AI is different than t intelligence generative AI is not predictive intelligence it's not a lot of the things that we're used to in the system there are other parts in pieces predictive intelligence specifically requires a boatload of data 50,000 records to go and churn through to try to understand detect algorithms patterns and then be aware of those to kind of tell you what's what's coming generative AI is not that okay but it is a partner to it so if you use the two in concept you can have a very very streamlined and intelligent platform that can really really really help you avoid big catastrophes reduce your meantime resolution lots of things so very very powerful if you're able to use them in concert how does it improve the experience cost is is probably the one that that most people p up when they he when I spoke in uh Orlando a few weeks ago I I made sure to bring home that the person the anal sitting at a desk with hands on keys that is your most expensive resource customer service specifically we put a lot of stop gaps between the enduser and that agent so knowledge base contextual search virtual agent all of the things that we try to put as as a wall essentially uh to keep you from getting to our most expensive resource generative AI is another wall that they're trying to put up which is a good thing right so we come in and we Implement for customers our our end goal is return on investment we're buying this tool they're paying us to implement it they want to see uh their costs go down or their productivity go way up reflective to that cost right but ultimately the goal is to to get more back on the dollar spent so now ass say for CSM they got a bunch of things are rolling out this was a September 23 release these are the most recent slides I'm able to get put together I'll show some slides some screenshots of the platform as we move through this deck I don't have access to an instance we now assist on it they're not letting us turn it on in our vendor instances and I've not been fortunate enough to land on a project with one yet either the screenshots I was able to grab are from a lab instance for some enablement that I took a few few months ago I guess at this point so we'll roll down the list now assistant AI search it can be used uh in the portal for virtual agent things like that what it'll do as you type you know the contextual search will try to surface knowledge-based articles for you sometimes those knowledge-based articles are written in technically which means that unless you kind of understand what's going on in that knowledge-based article in terms of its content understand the the language that it's been given to you it can be confusing which is counterproductive so now assistant AI search will kind of surface those knowledge-based articles and it will kind of turn through the content that's in there and reward it simplify it uh summarize it and in a lot of times it's it's it's kind of dumbed down for lack of a better term it's it's boiled down to to something that is much simpler to understand I know myself as a as a technical person sometimes I struggle not to use the vernacular that I do with my peers and again that's true in a lot of knowledge based articles so unless you have a team that is really dedicated and really good at going through those knowledge Bas articles and making sure that you know even people AR super technical can consume them AI search can help with that chat summarization if you do land up on a website or on a on a on a service portal and you're conversing back and forth with a virtual agent and you get handed off to a live agent the agent has the ability to summarize that chat with a with a quick action so they can get a summary of all the context all the things all the knowledge articles all the things we tried all the topics you consumed bulleted when that chat is handed over similarly case summarization we'll do the same sort of thing there's a requirement of about 200 words in customer comments and work notes there's a quick action in in the in that activity section of the workspace where you can submit all that data to generative AI and it'll summarize it in a bulleted fashion which is really really nice and I've got screenshots of both of those the now assist panel is the top right a lot of us use it they'll copy paste an incident number or a case number it's Global search what that will do it's contextually aware of the case that you're looking at or the record that you're looking at on your screen and it'll summarize it you can tell it to change the tone right things like that you can summarize the case you can summarize where resolution notes lots of things and then the now assist admin console is where you go to turn it all on where you go to configure it right now it's pretty basic but again we'll cover that a little more depth here in a minute key personas your end users your customers your agents your admin and kind of talk about the experience that all three of those folks will have so how do I turn it on what do I do how do I configure the skills how do I you know control access to it things like that it's really pretty simple there's a a very clean looking UI UI Builder type experience you can see on the left side uh there's technology customer employer Creator and platform those are the different uh silos that now assist Works in right so technolog is going to be for itom itm customer is going to be uh CSM employers hrsd so now assist and the generative AI functions that they're rolling out are available platform wide case summarization this is what uh these are what the configuration records look like as of Vancouver uh some of the things are locked down uh the case table is locked down um yes case types will work as they extend that table the input Fields you can see description short description work notes and additional comments that's it for now we can't uh we can't modify that you can see digging in a little further so once we've defined the inputs and the table that we're going to use for this we're going to go ahead and uh determine what features we want to use so in product is the case assist and the and the case summarization the work notes and then the now assist panel as well right so you can use one or both of them and you can Define what roles have access to them there is uh we'll touch lightly on this later on but of course there's a cost to all this right because these are all calls to a generative AI service once you've done that you can go through and uh again kind of turn it all on figure out exactly when you want things to to be triggered and and when you want them to be automated how you want it to be formatted that property property bulleted list property at the bottom I always recommend turning that on the a little bit that I've played with it uh if you don't turn it on it just ends up as a text blob bloed list applies some formatting and makes a little easier to read the admin panel so you can see here we've got the different plugins that can get in this particular instance you can see service portfolio management itom hrsd and then creator for scoped apps of course just a little bit of an overview there the account one of the concerns one of the questions that came up in the uh enablement and then again in in The Snug was okay what are you doing uh what are you doing with my data where's that going right are you are you sending this to chat GPT how are you making sure that my Pi or my intellectual property that I own as an organization is Is Not Just Landing out in the nether where evildoers can get a hold of it and do bad things with it so service now uses an an inhouse natural language Library so it's not sent out into the nether it's kept inhouse it goes through a number of screens before it's um collected and used to train their model removing pii removing intellectual property uh basically sterilizing it as as far as they can before it's added uh to the model to use for training of course to make it more uh more productive more responsive because it's kept inhouse it's not sent out to uh chat GPT the response time is much faster I don't know if you guys have have used chat GPT before but you know depending on the time of day it can take a take a little bit for it to to return some results you can imagine what that experience would be like to a customer interacting with your virtual agent or an agent looking for a case summarization or a resolution suggestion right if it took several seconds it doesn't seem like a long time but when you've got lots to do it is and it adds up so keeping it in house increases the efficiency and reduces that round trip from submission to response portal search so in in the portal and again I apologize for not having a live uh instance to display this to you uh portal search when it's when it's enabled there a customer or um internal user can go and use the the search knowledge search or just the search on the landing page I'm sure you all have seen it to submit what it is they're looking for again now assist is going to consume that data it's going to suggest catalog item or a knowledgebase article when it suggests a knowledge based article it's going to summarize it for you and again boil it down make it a little more easy to understand remove some of the technical jargon out of it and virtual agent will do the same thing so again when you go into virtual agent it says how can I help you and you say you know can't log into my email it's going to surface relative knowledge based articles and again kind of boil those down you can see here uh this icon this this little sparkly icon anytime now assist is used you you'll see that icon it'll be evident that it was it was used and it's being um referenced on the screen this is the the little glittery icon I was talking about you will see that here it is I know this is a little smaller a little tougher to see here but again everywhere analysis is used whether it's a resolution or case summary you you will see that so here you can see this is an example of George Warren who is a contact went to the knowledge base or went to the portal and and submitted a value saying that you know my my internet is intermittent slow and you can see this is extracted from the troubleshoot your internet explore these tools right so it surfaces that top article and then it starts to summarize it for you you could of course view the article as well it'll take you right into it same thing here so if he added the same search you know my internet slow or I can't connect same deal same functionality between the two interfaces how does it work from the agent perspective how how does this tool Empower me to do what I need to do more quickly more efficiently so agent workspace here you can see we've got a chat here on the Le hand side this agent would have utilized the the chat summary when it came over so that was in the now assist it can automatically uh once it's handed over to an agent summarize that for you okay and again here's the here's the AI chat summary little bit of an ey chart see a little glitter icon there it is private so of course the customer doesn't see it but you can see you know it goes through everything above here and and and summarizes it for you what articles you've viewed what topics you've been to basically everything you've done before you get handed off to the agent little easier to see here quick action in the chat SL summarize we'll get you that summary again so if uh after the handoff you want to resmar things again it'll consume everything that's happened previously in the chat it'll bullet that out for you this is the the bulleted format without it just imagine no bullets right be a little hard to kind of pick things apart little easier to to read with the formatting pce summarization uh so in the middle in the activity section again it's there's about 200 words that are required for it to really be useful anything less than that doesn't really give the the language model the information it needs to to summarize things so it it will work but it's not very useful so again there will be in this section there's a there's a summarize click summarize there's the little sparkly icon again we can share it directly to the work notes so put that summary in there which is nice because if you ever doing a post case review where you need some um milestones in resolution you can you can always go back and review those those of course can be added to uh your case action summary copy paste them out of there put them in your case action summary there's a couple of quick link buttons down here at the bottom think I have a better image of it I do cool so now assist case summary is here little button real easy to use click it it's going to take all that data send it off about 200 words you can thumbs up thumbs down it remember we're training a model so this input is very valuable to the curation of useful responses okay and again you can drop this right into the work notes here you can see that it maintains the format it's really nice uh copy it go over to a case action summary paste it in there or if it's not quite right not quite what you need there's also a you know refresh so it'll send that off and try to summarize it better resolution notes again especially if we've got a case that's kind of been around for a while or it's complex or you know I've just got a lot to do resolution notes work similarly when you go to propose uh your resolution you have the opportunity to to use generative AI to do that again little Sparkle here and then it summarizes is again the the case in its entirety give you a second just kind of scan through that normal suspects that add resolution notes to comments if you want to do that check box is there so so again we're we're not Reinventing the wheel so in three clicks we can have now assist construct our resolution notes construct our our proposed solution you're going to add it to the comments and send it Rec clicks so some of you type really quick some of you don't you can imagine typing a paragraph or two resolution notes even if you copy paste that into the comments or you use the the check box at the bottom three clicks is a lot quicker and 300 characters so you can see where we start to pick up time right so from the the onset where the the customer interacting with nowy and we're trying our best to surface things trying to make them um easier to understand to the summary to the agent once they get handed off in chat to the summary of the the the details of the case so far the case summary and then proposing the solution right you can see where all of a sudden we're picking up minutes minutes minutes each step of the way that across you know however many cases you touch in a day you just you just bought yourself the ability to touch you know 20 more cases today so you can see that across a service desk across a year that starts to snowball and turn into a lot of savings now assist panel in the top right um I know I have a better image on the next screen of of all of the functions that are displayed here so from now assist remember it's contextually aware to the record that's on your screen so it knows I'm looking at case 2003 so if I go and I want to summarize the chat or use generate some resolution notes SU Miz a record it's going to give me a response I have the ability to tell it look you sound grumpy so let's change the sentiment a little bit so you're interacting with a generative AI so you have that functionality so that same case we went in and I hit case summarization resolution notes and summarize records and you can see uh these are the different models that it presented me the different bits of data and again if you did want to change the center or the tone if you wanted to make it longer or shorter those are all options terms and usage so we talked about how we can save a customer money let's talk about what it costs them right so there are a couple of parts and pieces of this I won't speak directly to cost there are skills actions and assists so I'll let let you kind of glance through there skills are going to be all the the features that I've told you about so far in terms of the resolution notes the summary the chat summary an action is what happens when you take that data and you send that off to now assist okay and the and then the assist is really reflective of the complexity of the action so you can see in the examples below a case summarization is one assist if you ask some something um in the now assist panel that's five assist with less than five interactions so the the summarization of the record the chat summarization right so if for less than five interactions with that or change the sentiment make it longer then it is five assists anything more than that jumps up to 10 service now has built the ability there are spokes available to call chat GPT so you can send your data out that of course is uh very complex uh and requires 50 assists so you buy those in buckets just like anything else I don't know the cost around them hopefully we get um some information on that or maybe we can hit up some of the sales folks I see on the call maybe they have some idea and they can they can help me augment that data into this this slide deck how do I know what I'm spending and it's like they thought of everything so there's the assist admin page that's going to go through um and and month after month tell you what you're using tell you what agents are doing tell you what your the the button that's that's getting mashed on the most um track that over time this nice performance analytics dashboard can help you kind of understand your usage unfortunately it's in a rears so this has already passed the Bill's already been sent but at least you have visibility into you know if somebody's really overusing it or maybe you're not using it enough things like that so there is there is an out ofox way to to kind of get your head around what it is that's going on in your instance relative analysis so a few things you have to have in order to use it here pretty soon the uh the Vancouver patch to or higher will be a moot point because everybody be at least there if you do decide to use the open AI parts and pieces of course you'll need a subscription there chat GPT things like that and now subscription a lot of these are pretty self-explanatory right got to be on the right release got to have access to the uh the right resources right now only English is supported sometime I heard sometime over the summer they're looking to spin up some additional language they didn't narrow that time frame down anymore for us and they didn't indicate which languages would be next on the docket so I don't have any more to share with you there unfortunately enablement lots of it these are all urls of course this deck will be available and go to now learning and search now assist or generative AI uh if you need help finding them hit me up uh Tyler Hogue is kind of spearheading all of our efforts for generative AI I'm he's a he's a good resource so if you have questions free to ask him too oh [Music]