all right looks like we are recording so good morning good afternoon or good evening depending on where you're calling from and Welcome to our third session of the S so modern change life cycle launch and learn um today we'll be focusing on nextg risk but I do want to make a quick announcement um before we get started uh as some of you may have seen in our last session recap email Greg is going on leave for the remainder of the summer so will no longer be our main presenter here at launch and learn that means a few things um one presentations will be held more by a committee there'll still be some familiar faces such as Isaac um and we will be covering all the same topics it just won't be led by Greg um but rest assured you are in great hands um we're gonna take care of you um that being said due to the Personnel change the schedule is going to change slightly um so the next session s so change will now take place in August which means we will not have any sessions for the month of July um consider it your uh midterm break or something of the sort um also due to that change um as you could imagine it will likely take us a little bit longer than normal to answer some of your more technical questions or your questions about the direction your or should take with um regards to change um we have been receiving your emails I I promise we are doing our best to capture everything and answer them in a timely manner but as we move through some changes on our end it it is going to take a little bit longer than normal to get back to you um we'll still have folks on these calls to answer the questions you have immediately and try to address them as best as we can um but I do ask for your grace it's it's this was a uh unexpected change so it may take us a little longer to answer some of your concerns than it has in the past um and that is all for my announcement though I do have a quick survey before we get started so I'm gonna try to pop this poll hopefully this works that's funny uh all right um I'll give you all just like a a minute and a half to to run through this poll here um and once we end uh we'll just record it on our oh you guys are so quick thank you so much I know how painful it is all the surveys and polls that we send you away so I do appreciate you filling this out a few of you have seen Terminator that's a good sign and good company want this go for about 20 more seconds hopefully get to at least 75% participation here [Laughter] there's always a kill switch all right that looks good to me um so with all that being said I'll hand it off to our presenter for today Daniel Davidson D Dan I don't know if you want to give like a quick spiel about yourself who you are maybe a a one minute intro uh and I won't take up a whole minute so I work on the outbound product management team with Isaac and I work out of Amir I've worked for service now for about 10 years and previously that I led the implementation teams out of Northern Amia and The Architects out of the UK but happy to be on the product team um yeah that's me oh I look after itm and devops for me that's that's my product okay just going to bring it the back so hi everyone just a bit of um supp a bit of housekeeping first um I'm leaving this slideing can everyone can you see my screen you just give me an odd Chris if you can see looks good y looks good to you okay so the format we're going to follow today um as we've done before just some housekeeping on uh on identification adding any company name um and some I think Chris has already covered you know how we're going to do the Q&A so are we are we saying come off mute and ask questions during the during the presentation Chris is that what we're saying no I I say throw your questions in chat if you have um a burning question and you want to speak uh just raise your hand and we'll take you off mute um just because there are a lot of you and so don't want too many voices at one time but if you do want to come off mute just raise your hand and we'll call on you if we can yeah and then the end of the session we'll have uh that as well where Q&A uh please raise your hands at the end but y so yeah and we have got a panel of experts on the call so if we get tricky questions it may take us a little bit longer to get the answer back to you so think because you've asked a question that hasn't been answered for a minute so that it isn't going to be answered we're just we're just kind of probably corralling some some answers back to you um so first of all just to set the scene um we're going to talk a little bit first of all in general about nextg change I know it's probably been covered by Greg I'm not going to spend a great deal of time on it uh up front but just set the scene for people who haven't been on the previous calls and and understand I suppose as a product team as a company where we see change going in general um and then we're going to dive in um uh with the risk intelligence stuff which is a really cool and I think quite underused feature and service now so there is a call to action on this you know we want you to be using it and and we think there's a fairly safe way that you could be using it you know to to experiment with um but we'll come to that as we're going through the call so um so hopefully we'll convince you on this call that it's something that's worthly worth at least um having a play with in a in a sub production or cloned instance maybe and then trying to move it into production Okay so we've got our Safe Harbor slide read and absorb um this is just saying that there's forward-looking statements we might make about the direction the product's going and uh as any software company and that's no guarantee that something will make it into a given release okay so this is the new schedule that we have um Chris mentioned at the start of the call this is the right schedule isn't it Chris I'm pretty sure it is this is the schedule that we have going forward for uh for our subjects and who's presenting them and when they'll happen so um again if you want to take a screenshot of this if you want to update your diary then that'll be a good thing or or perhaps you got any Daniel we've got some requests maybe if the microphone could be moved closer to you am I quiet H yeah let me is that any is that any better I think that's much better you sound a little try I don't know just keep going and see if that's the built-in microphone on my laptop as opposed to the microphone that I bought so that show you something about I'm okay cool let me know if you need me to speak up a bit right so we're going to talk about the evolution of of change and again I think Greg's covered this already but really to start off with 40 years ago we were talking about just being able to control change with an it was something that needed to happen that's what people realized and change control made it so that you had to go through a certain process to to make changes within it and really enabled some level of governance um for for for change but moving on you know we talk more about um change management and that's when processes like iil started to come to the for we started to standardize across the industry how we do change management things like normal standard emergency change were were born in that period and and C people started to adopt ways of doing it and the service that platform you know came into came into being around that time too and and we we enabled change around the business we enabled change management but moving forward now we've got a different Paradigm um and what we need to do is change enablement so the question that gets asked by a lot of customers is you know how do I Federate change whilst maintaining control how do I cope with all these different teams doing things in slightly different ways across my business whether they're part of central it maybe they're part of you know they're distributed maybe they're using different Technologies maybe they have different levels of quality assurance when you shift left into their systems that we don't want them to have to replicate in the change process how do I guarantee that they're doing that and they're being followed and how do I allow different areas of the business to work at different speeds uh within the change Paradigm so you know key to that concept is is one of the things is is automation so we need to automate um we need to automate to be able to make change flow faster so we've got a problem with velocity change we need velocity to increase high velocity changes are normally safer because we have a a better degree of accuracy about what the target system looks like if the change is designed and made closer to the change time the changes deployed um the volume of change also needs to increase because smaller changes are are more stable uh and we know what we're changing a lot more so making more small changes necessitates the increase in volume and if we're using the same process we use for big changes to make all these small changes we're going to end up with a lot of work so we need to automate those small changes and the other thing that's happening is compliance we need to be able to demonstrate to Auditors to to outside people to people looking into incidents on the back of changes we need to be able to to show them the processes were followed or that why we made certain decisions within a change um so tying together automation on these things so using uh features like Risk intelligence using features or products like devop that we have um using things like approval policy which I think we covered elsewhere we'll cover a little bit today um allow us to increase the volume increase the velocity increase our compliance whilst actually um getting through more work with possibly the same number or less people and it's really a shift for the change manager away from being you know intimately involved in every change so what they're doing is they're monitoring how change is flowing in the business they're enabling different people around the business to work in different ways and understanding how they work and stepping back into more of a second line uh role within change management and delegating or federating out the actual change process to the teams who know the change is best so they're asking them how they want to make change and then they're allowing them to make change um and then they're monitoring that what the teams say they're doing is actually done through data and also through looking back at changes and how successful they are and Central to that is is our risk intelligence for change because it's a big input into How likely we think a change is to go wrong uh is the risk intelligence of of of change and and being able to use things like machine learning and artificial intelligence as an input to that is a big bonus right it's something that we're not really used to in change management um but it's something that you have in service now if you're a Pro customer so I'm going to cover a little bit more about that next a lot of what we're talking about not everything a lot of what we're talking about today fits within our itm Pro Suite so the licensing model is standard Pro and Enterprise um in the standard model we're looking at people who are standardizing their processes right who are trying to create a cmdb who are trying to make sure that people follow the same change process who Tred to manage those processes then when we move into the pro license what we're looking to do is take what we did in the in the first phase in the kind of modernized phase the standard phase into the pro phase which is around you know how do we make things work faster Slicker how do we do more with less how do we apply technology to these things to these problems to allow change to grow faster to allow us to reduce the mttr of incidents or to understand more about our technology landscape in a more automated way and so risk intelligence is part of the itm pro site you need an itm Pro license to use it we will also talk about a couple of aspects which in in the standard license model when we do the demo and then over to the right uh we've got the optimize um licensing which is our Enterprise uh license and that's around continually looking at the processes understanding where there's blocks and gaps and and and and processes are not optimized understanding how we can work to improve those processes to make sure that we're we're constantly looking at them and how we can again use artificial intelligence machine learning to understand where the bottom legs are in our processes as well in our Workforce in where we're applying our technology in what we have in our assets so this is the Paradigm I talked about earlier in the in the call so we're seeing an increase in the number of changes being raised right devop teams raise a lot of changes um and those are being raised by machines as well so cicd pipelines are raising changes and and they can be doing you know one two 10 100 maybe more than that a day and this volume of change is something we need to handle and the way in which we it is through Automation and if we apply AI on top of that automation we get like a boost to how many changes or how many how much we can handle and so this increase in change volume allows us to increase the velocity that the business moves at and actually it ties into the Strategic objectives of it and and quite often the Strategic objectives of the business as a whole is Technologies integrated more into what businesses do into how they meet their customers into how they serve their customers into what products they provide um we find that an increase in velocity an increase in in the stability of what we're doing will increase customer satisfaction will you know allow us to operate with better margins will allow us to have faster time to Market more differentiation competitive Advantage those are the types of things that having a faster change process enables and it's why change is so much more important now than it was before before it was about stability now it's about affecting the strategies that the business has to to digitally transform and again just last high level slide before we go on risk intelligence part is the concept in service now of of service operations and this isn't a product we sell right this is this is a set of things that work together some ideas that work together and they are underpinned by our products but it's the idea of Bridging the Gap between service and operations so it's breaking down as silos between teams and allowing teams like developers to work seamlessly with teams that work in operations it's about understanding how the cmdb fits into our processes it's about thinking of it all as one thing how set Ops and and and incident interact for instance all all those kind of concepts of having one place where it's more to do with a person doing their job than a set of processes that a person needs to interact with separately so if we can make that person able to flow from say an incident into some knowledge into a request item have playbooks understand how to raise changes have them tailored towards them have experiences which are tailored towards the job they do we can make them more efficient we can make them happier in what they do less frustrated by it and we can make it itself operate more efficiently and more effectively because we broke down those silos and that in turn leads into the Strategic goals that I was talking about so okay so next gen risk okay so I think in a previous session some of you will be on it would have been on it um we would have covered change success scores so this is really looking at um how successful are changes so what's the chances of it failing and the first thing that we we look at is how successful a team are at making a change and the reason I'm talking about this now now is it Fe it feeds into the risk assessment we do so How likely A Change Is to go wrong and the impact of that change going wrong are really the two factors we look at when we're looking at change risk so understanding change success scores having that configured in your instance and part of the itm pro Suite as I mentioned before but it's really important to be able to have effective and automated risk management within change and we can't we don't just have to look at how the team does in change you know one of the um sessions that we have is about multimodal change which is a concept the industry is a opting as a whole understanding the purpose of a change and having a specific change for that purpose allows us to investigate how successful that particular change is so we can intersect how successful a particular team are at doing a change with how successful that particular type of changes and it gives us a really good indication about how successful that change is likely to be and what we're doing this is we're looking back at data that's in the system and drawing conclusions and a score based on the team and what they're doing which are other factors we want to know about risk of a change okay so when we do the demo Isa is going to show us a little bit about that as we go through but we'll move on now so the success probability is is a combination of the change success score and the model success score okay so again drawing together the different areas features whatever we want to call them in service now that are important to to risk management and how we think about evaluating risk we talked then about team success probability model success probability calcula a success probability based on both of those being more powerful than one of those on its own and what we want to be able to do is then understand the impact of the change too and actually the risk conditions of a change we should probably call them impact conditions because there are a module that's in service now we'll demo to it as part of the demo but that is a way of capturing data around risk now traditionally or more traditionally I'm going to build out this Slide the bottom we might have used a risk assessment which is a kind of Q&A a structured Q&A around the risk of change and it's okay right and it's part of the standard license and you can use it it does give us some information on on the change for sure and it records at that point in time what that person is answering which is important for audit but you know people can give answers that aren't correct on it either deliberately or because they want the change to happen a bit quicker or because they just tick the first answer on the sheet that they have in front of them and so having data driven risk conditions is more accurate you know it allows us to leverage the data that we spend time building in that modernized phase that I talked about in the licensing it allows us to leverage that and use it to build out what the risk of a what the impact of a change will be so we look at things like how critical are the business services that may be impacted by this change um and we feed that into a score for impact we also can use risk intelligence so again the focus of the call today we can use Ai and learn and machine learning predictive intelligence to predict the risk of a change based on the history of change that we have and so we will apply patterns to the data that we have in the system there's two different ways two different methodologies you just do that again we'll cover that later but that will give us a score to and what we do is we combine all these things and there's a risk Matrix we use and that will allow us to assess accurately um the risk of change and we take a paranoid view to this so whichever one is producing the highest risk score is really the one we want to go with so what we're doing is we're moving from you know maybe just using the score on ACI or or a risk assessment that we get someone to fill in accurately or not who knows what we're doing is a more data driven approach and the other thing about data driven approaches is we don't have to wait for someone to fill in the survey right that data is there we just access it and we can process it and the change can move forward in seconds or minutes rather than waiting you know a few hours or days for someone to fill in a risk assessment and because we're using structured data we be less paranoid about how the about the change flowing forward so if certain conditions are met if we're able to satisfy certain conditions around it we can then feed that into an approval policy which says actually this change is likely to succeed without any approval the guard rils have been met let's Let it Flow right and because we have a constant feedback link about how successful changes are and because we have risk intelligence turned on we're constantly looking back at the changes we have to adapt our model to be able to pull out changes that are you know looking more risky or the teams that aren't following the correct guard rails and improve their processes to allow change to flow faster so it's about the risk of that particular change but it's also about looking at the overall picture of changes that we have in the business and types of changes be able to slice and dice up by team or model or whatever we're doing in that space and be able to apply Ai and N have we I haven't been looking at the chat have we got questions coming up in the chat or anything like that so far doing all right we're doing question I haven't said anything too controversial then so all right we do have a question so let's go ahead uh let's go ahead and field that real quick here so Frank I'm gonna ask you to unmute can you hear me yes perfect so I think this is awesome you know being able to look at you know how successful a team is does this also take into account the success at the CI level so I do a good job doing my changes and you know I have a high success rate but that particular CI has a a lower success rate no matter who's doing it versus that team has a low success rate and they're trying to change a CI that has a low success rate is that brought into this as well I I'm going to give an answer and I'm going to see if any of the team correct me on that because um I know the answer is accurate but I just want to see a better way so risk intelligence is what we take into account the kind of multiple aspects of a change so we're looking at all sorts of aspects of change and machine learning is feeding in those kind of fuzzy parameters into whether A Change Is is risky or not um it's very difficult there are so many CIS out there and and accuracy of recording CI against change I'd say in practice it's quite difficult to do I think when you start looking at service level uh and those kind of logical layer especially with things like um devops which kind of happens at that application layer rather than individual CIS kind of ephemeral below then um you can start looking at the success of a of a given app service or or business app or service or service offering or whatever you want to do in that in that area but um I would say that that applies to risk intelligence more than the um than the success probability module and the short answer is we use team success or modern success to assess the probability of a change failing but not the CIS or Services anyone else is that okay go ahead Daniel we'll go ahead and uh we're going to continue answering the questions in chat right so again when we go for demo we will do that very soon we're going to talk about risk intelligence and we're going to talk about two different ways that you can um apply machine learning so first of all uh we use similarity and that's for models that we recommend below 10,000 records to change and what that does is it analyzes the text on a change and looks for similarity classification is a more powerful method it's for bigger data sets and you can set up your risk intelligence to use either one of those two things Isaac will will'll talk you through that as we go through but predicting change either uses a similarity or classification framework and we provide out the box we provide a learning model um for the classification of changes using predictive intelligence so in in here we can see we're setting the similarity and we're telling it to predict the the risk value of the change again Isaac will will talk you through this when we go through so what we're doing is we're providing a framework that you can just kind of switch on and apply and what was kind of um really kind of transformational for me is I didn't realize until quite recently that you could apply this without it actually being applied to the changes so there is a toggle in service now that allows you to turn on the machine learning and have it collect data and have it do the classification and provide as information to a change without actually kind of going live on using uh Skynet to set your changes so we you know we have we have the ability to to put it there and you can see how successful it is and maybe you could start you know if you're using change models again future talk comes to change model talk but you know maybe you can start applying to certain paradigms and and and keep a high touch engagement on it and start rolling it out it the combination of multimodal change and the other features we Us in the other nextg features approval policies devops all those kind of things tie in well with the idea of federating change and having different ways of doing things it isn't just one big monolithic normal change that everyone has to use so um again just just some information on the change so this is what happens when we're applying it so based on the calculation the risk has been set we give you a message telling you how um how predictive intelligence has interacted with that change so it's not something that kind of sits in the background um out the way it's something that's that's in front of you we we want to be able to to to show you what input that's having to the change and the last slide I'll get to before Isaac takes over and does the demo is just really the back stop for us which is where it feeds into approval policies so what you've got is Insight at the point of risk intelligence or using risk assessments or using the change success scores you've got something that tells you something about the change but you're not really taking action on it at that point because what you want to be able to do is is drive your approval of change drive the the Assurance for that change and the the ability to allow that change to flow through with this data and so change approval policies previous to this we would have something like c m or a group of people that that might look at a change and see what these values are and ask the questions or just allow the change through it's kind of a little bit unstructured what we do with change approval policies is we feed the data in that we're getting from all sorts of sources including the risk evaluation of the change maybe including individual answers to um to the risk assessments that we had in the risk conditions we had in the change uh and we can we can drive who approves it how they approve it or if anyone approves it through the change approval policies and you can build out these change approval policies to a change approval policy Builder which is quite a nice UI we have in the platform and that allows you to kind of set the conditions under which change moves and again applying it to multimodal change which we'll talk about in the later talks allows you to use different sets of approval policies in different paradigms so we're being very flexible again how do you Federate whilst maintaining control well you use automation like the risk intelligence engine use different uh risk assessments use change approval policies in the right places and put them together in a in a particular model of change that allows you to to work with that area in the business and and and have change flow fast for them or as fast as appropriate business okay Isaac do you want to take over now I can do that all right so let's go ahead and talk about our options here and the way we're going to go over the demonstration today is I'm going to go over itm standard functionality first and then we're going to go into the itsm pro functionality so to start off with um everything is included in this instance this instance is itsm pro so we're talking about when we say risk risk right we're talking about this field we're modifying this field on your change requests and there's multiple ways that we do that and calculate that this ticket it's going to show you a representation of everything so let's view these details all right so we've got um risk conditions and got my sorry about that risk conditions risk assessments all right those are going to be in your itsm standard fields and then risk intelligence so those are the the three categories we're going to be talking about today and how do you go ahead and get started and get that set up so to start with we're we're going to go with risk conditions and underneath risk conditions here I'm going to open up uh this one here at the top called insufficient lead time that's actually uh a bit misleading it really what it should say I think it um probably should say excessive Le time all right so in this risk condition what we want to have here um is setting it to a risk of high if these conditions are met so remember that field I showed you off that change ticket that's what we're setting and then uh the impact so risk conditions are are unique they also allow you to change an additional field which is going to be impact So based off of excessive lead time um and we make sure these conditions are met so the change ticket cannot be empty uh for the start date and it's uh on or before 3 days from now and it's in the new assess or authorized stage so pretty straightforward has set those conditions up here you can add more as appropriate all right so what do we want to happen if these conditions are met we want this field of risk to be set to high we want impact to either stay as it was on the ticket or we can change that as well and definitely make sure to set it as active if you were to uh want this to be used on your risk calcul ations okay so that's that's a kind of Dipping our toes here into our options right so that was risk conditions pretty straightforward we're going to talk about a little bit of a controversial topic here and that's risk assessments so risk assessments are also going to be part of your standard and some people are not fans of risk assessments uh me uh included on this call one of them would be me and the reason why is because this is a essentially we're we're asking the the individual submitting this ticket to give a self assessment and say is this going to be high risk low risk all right so for this risk assessment to even show up as an option on the change ticket and I've also set it to active so keep that in mind these conditions must be met so that needs to be an active change request one of these types in those particular States or um these conditions as well there's also an assessment designer so can go ahead and launch that and see what that looks like here so this is really nice it's uh you know looks like a nice little survey um tooling here so we'll go ahead and let that load so here's my here's my uh my survey if you will my assessment and you can change these and certainly add uh different controls to this what does that look like let's go ahead and show you so so I'm going to go ahead and open up a new change ticket new change request and something to note here too um we are going to get into multimodal uh at a later presentation Daniel talked about it but we're going to just pick normal mode normal change requests today one one example here um that we want to talk about by default some some of you may have noticed when you open up a new change request the risk is set to moderate so um that's that's something that we've we've had some conversations on um I know a lot of our customers just automatically set that risk to low so there is there is enhancement requests please add to that vote that if you open up a new change request um I think low should be that but there's certainly a UI action you can put to change that if you if you so choose to so I'm going to go ahead and just fill out some brief descriptions here I'll put an assignment group gu engineering and there's an sap service that work just fine implementation [Music] plan going for it all right so let's go ahead notice here um there's nothing in the related links as soon as I save it now I have the ability to perform that risk assessment all right this is this is available we do have we do have quite a few customers using this and um that's that's fine just understand that you can certainly put in answers that may not be accurately depicting the true risk all right all right so we'll go ahead and say it's not critical it's easy it is going to be easy to back out there is a redundancy plan and we'll go ahead and Su at that all right so the risk has been set to moderate and that was based off of the assessment that just came through so let's go ahead and verify that all right so the risk assessment came in low and the risk condition rules came in low as Daniel pointed out because I had both of those running on this change request I had risk uh conditions and Rules running on that and then I had the risk assessment the higher of the two in Risk would be selected as the the field value but in this case here they are both low so um we're going to put that calculated RIS RIS score and is moderate and that takes over right so the moderate is obviously more severe than the low let's move on but before I do I'm going to symbolically disable my risk assessments go go ahead and just save that and now we're going to move into really um kind of the centerpiece today is going to be risk intelligence and if there's anybody who has itm Pro and you haven't activated risk intelligence shame on you no I'm kidding but um there is some relatively low risk to activating risk intelligence and I'd like to show that to you today sometimes it can be a little bit um you just want to know how it's going to work before you activate it we built that into the product so let me show you how that works opening up um risk intelligence So Pro feature okay there's two different types of solution types that you can run one of them is a similarity and the other is classification so the superior of the two is really classification so classification is it's going to be a better job of building the model and it's going to have a higher level of precision so why would you ever want to use similarity well similarity is used for newer customers if you have less than 10,000 change requests that's when you use similarity if you have over 10,000 change requests and the system will actually not let you activate classification if it if it counts less than 10,000 so you have to use similarity until you get to 10,000 but then please by all means switch the classification it's just more precise now here's why there is no risk to activating risk intelligence because you can just view the value so as as a homework assignment for anybody that has itm Pro please enable these Solutions and please set it to view the risk value we won't change anything on the change ticket it won't be modified it will just show you what the calculation came out to once you feel comfortable with the results that have been coming out most likely your classification definitions then you can actually allow it to set the value on the change ticket so I'm going to repeat it again because I we really if you have itsm Pro just view the risk value see what see what it comes out with okay now I'm using similarity today because this is this is a more simple demo instance I have enabled the solution and let's take a look at what's going on behind the scenes so opening up this capability definition okay so here we go all right so step one what's relevant well obviously we're doing changes okay here is where about 80% of the decisions are made What fields do you want predictive intelligence to look through I would advise you don't go crazy when you first start this out you'll be surprised when you just use Fields like the short description and implementation plan the results that come back okay so so don't go too crazy when you're setting these up uh Daniel and I had a really good conversation one of the one of the fields that you might want to add is is going to be your close code okay so that might be something you know start with those three like short description implementation plan and perhaps add close code all right so we're not going to get into that too much though but you add the fields and then these are these are conditions so it should start to look a little familiar here you want the state of the ticket uh to be closed uh to be calculated into this training model all right and then the type is not standard because standard's probably not going to be as interesting all right and then you can train it right the update frequency this this one I just trigger on the fly so I don't have this happening every day or you could even have it happening every 15 minutes if you so choose to and um what we see here is I ran two different models and the progress came back so we do have um we do have the information that is in here some some solution parameters um I just like to um talk about this here so so behind the scenes these are the parameters that are going into it right we've got uh similarity you know um we've got clustering that's happening so behind the scenes these are these are all the the processes that we're running on those three Fields okay so let's go to one of these tickets that um has had machine that has gone through predictive intelligence and I'll show you what this looks like after after it's activated I'm just about to wrap up this is the last this is the last piece I'll be demonstrating and then we can open it up for questions all right so here's a ticket that has met the criteria to be included in that similarity definition and um I'm going to go ahead Let's uh let the internet catch up with me all right so I'm going to fire this off calculate risk right so just making sure it has all the latest and greatest no harm there and let's see what we came up with okay risk it's moderate so it's telling me that the risk and impact have remained unchanged so that's that's fine I didn't need to rerun the calculation but good to know and now uh looking at the details of the ticket this ticket did not go through risk conditions uh there was no risk assessment I deactivated that because I'm now taking over with risk intelligence and this is this is based off of those fields based off of the the machine learning that we're running and processing the risk intelligence has set this to moderate so there you go and that would be your itsm pro feature of the day uh again going to repeat myself please please please activate it please set it to just view and just just just see what it does okay all right that is my presentation let's go ahead and uh open things up a little bit has the have I created a lot of problems in the chat I wouldn't say problems but people are are definitely vocal which I love to see um there are a number of questions but does anyone want to come off mute and ask a question verbally maybe they didn't feel the chat the proper answer for Aaron I'm gonna ask you to un mute here H thanks for uh thanks for the the demo and and all your efforts on this call um and I'm curious about this um view uh only kind of implementation uh to initially kind of test it and evaluate it um could you could you kind of show and maybe I missed it um where if you said it's a view where do you view it and who who would view it how would it be viewed understood because I would hate to send conflicting messages against our current risk calculator that's a fair point so maybe I should have added a caveat to maybe perhaps do this in a test instance right but so that will show up that will show up um but we just will not be modifying the field itself ah cool yeah gotcha and it's it would be viewable to all itol users I'm gonna throw that out to my teammates um so anybody who can read the change should be able to read the the risk that's been um processed for that given change gotcha okay cool appreciate it we had a question from Matthew Brock really early on uh which I was going to answer then we moved on and I I already marked it been answered live so he said the IML pattern out the boxing predict risk based on risk of other changes not the success of similar changes or number of instance major instance outes problems caused by changes there are reason for this seems predicting risk on risk or of the changes circular logic so I think first of all we're not predicting it on on the risk SC of the changes how we were looking at the success of of those changes but um I don't know if anyone else can answer is there anything in there that could probably look at um tying together Associated inance major in and not aware of anything at the moment that looks that no it's a bit of a it's a tough one to actually get right without doing um specific solutions for specific customers so the solution that we did the solution that we that we built a couple of releases back both and the um classification was probably the best reported um Solutions we created based on the data sets we have that customers have allowed us to train on and test um very good results based on predictability the problem when it comes into doing things like um affected CIS Associated outages def those kinds of things you actually need the MLS where we train our ml Solutions doesn't allow us to quickly go I'm training on change and then I'm also going to train on this related data and that related data because then you're doing aggregations and things like that you actually need to start looking at a separate effectively what it would boil down to probably like a summary table which would relate back to the change the information you want on that plus a bunch of other things so it actually ends up being a substantially larger solution and potentially not fit for purpose for all of our customers so we sted we sted away specifically from doing that but this was wow when I don't remember when we did the these ones it was pre- Tokyo so was quite a while back but that was the that was the General the reason for us going with the way we went because we've also found that so many customers have such specific um requests and and um needs around how they want to calculate things via ml that we weren't able to provide a really good outof the Box solution that would work for most people thank you Cameron really appreciate that answer um going to open it up to David D Johnson I you raised your hand like five minutes ago uh can you hear me yeah you sound good okay my my question is is probably I got two real simple questions so do for for risk intelligence do we have to turn on predictive intelligence it is yeah it it relies on that risk our risk intelligence expects there to be predictive intelligence Ena it's a pro feature so as long as if you have it and pro as far as I'm aware you have predictive intelligence and therefore you can have risk intelligence too okay okay and and and then the second question is just what is clustering when we talk about clustering uh and I get grouping is it like different types of CIS uh uh clustering together it's just it is um one of the algorithms that we support in order to do it what I would actually genuinely just suggest is to have a look at the various ones we support similarity being one which we know that does simple txt based ones um when it comes to clustering it groups it's going to group um changes based on related information sorry apologies I used the wrong thing it's not clustering it's categorization Lee you should have corrected me on that one our clustering is us is our is our um standard change process this is categorization so we categorize these changes based on um the fields that we've selected on there and those then we predict based on that my apologies I used the wrong one correct another question for you here when um so to what extent and and Cameron what extent might users be able to um get risk intelligence to calculate the way they want it to by providing certain keywords or or modifying um the change ticket so I think that that's that's a really good question ISAC so um the benefit we have here and what we've seen with other our other attempts previous attempts at some at some cross table ml things for a good example would be incidents caused by change or cases caused by change those are two examples that failed dismally in our training ability purely because they were almost always created by two separate types of people you have incidents and cases created by agents and and that kind of persona and you have a CA a change which is often created by someone who is not the same kind of person so you end up with very different language very different references of things so what we found though is prediction based on the language that's generally used on change requests has been pretty consistent but that's where that's where the the problem generally lies is that if if it if it's really dispersed and and the ml engine struggles to identify things because terminology is too is is too dissimilar and that kind of thing that's when we we start to see things we saw things broke break down a little bit more but that's why change worked so well because even with all our data sets across all those customer instances that we were able to do it on um there was a relative consistency across that and it it responded really well both similarity and for categorization and I'm failed to mention during categor ation and the the classification there's a threshold too so if if you want to put it like you know hey we're 50% confident that this is the risk score you can say no my threshold is 75% you need so that's also something that you can do to refine the results okay um and just so you know if we didn't get if we don't get to your question uh we will be responding uh in our email too for any that are not answered live so don't worry um we will get to it get some raised hands questions ra hands yeah I want to say it was either deepo or darling um I'm gonna go with deepo first and then we'll come to you D I think deepo was waiting a little longer yeah yeah okay so the level of confidence is that is that your question Deo that was that was definitely one but I also just wanted to quickly check if you're able to U query on these outcomes by different avenues like Risk conditions risk intelligence so basically pull them up uh across all the changes and run our reports is that possible so you're more of like a performance analytics reporting type of scenario yes yes how many changes are coming up with risk intelligence which is moderate versus how many are high that kind of reporting um there is a table called change uncore risk uncore details um which recalls all the uh risk calculation results so I think you can have a look at that table um definitely can be used for qu thank you I do have another question but that's on the chat so if you want you can answer it offline too uh Darlene let's go ahead uh should be able to speak now hi yes thank you very much and deepo did get his hand raised before I did so no worries waiting there I had a question um we're currently utilizing models and was curious to see if we could have um various gr intelligent profiles based on model because we might want to apply conditions differently based on the model that um the changes that's that's not not different um ways of applying risk intell was ways of applying risk intelligence rather than Gathering risk intelligence so when you look at the the framework we've got for assessing risk um what's that called again Isaac I can't remember the name of the of the table there but you can have different definitions you can use multiple definitions and you can get them to be applied to certain models so a model can have associated with it a risk risk calculation definition I think it's called okay great thank you hey uh let see Aaron did we did we get to you I just I just have one more clarifying question um which is it's it's related to just kind of the enablement of of the um sorry enablement of of multimodal change and this risk intelligence um looking at the information that you guys have out on now um uh now create there's a there's a document that I just dropped into chat it's the service operations recommended implementation sequence and and it kind of details uh craww walk run approach to implementing parts of the tool and and I and I I I know there's a there's a huge desire to implement productive intelligence and multimodal change we really want to get there but um we're we're only in our third year of service now uh at at human and we're we're still really trying to address some of the exit criteria from the foundation crawl uh and so on um is is there good like um for some of these uh for the for the purposes of this call is there sort of a foundational crawl walk run that kind of breaks that breaks these features down into like more granular components because the last thing I think anybody wants to do is Implement something that is unmanageable or really unwieldy um because maybe you've over customized or maybe hav't implemented parts of the tool here or there is there is there some guidance that that could be had kind of on an approach to implementing these these parts of the tool so the first thing I'd say there is as Isaac pointed out you can turn on the risk intelligence features and and have them as informational so you know that that gives you a step around confidence in general the second thing is that kind of that the how big something is operationally and organizationally and how big it is technically um you can reduce the overall transformation you're going through by you know doing something technically in a few areas but for a small area of the business and whereas if you're changing the whole of how change works if you're changing your normal change flow if you've got to make make that different it's going to be a big lift right organizationally working out your op model working out who you know who you delegate to who does the change approval policies who does the risk the risk definitions that's a huge thing to do but actually what the combination of this with multimodal change allows you to do is break off a small bit of the business that maybe you know would see significant outcomes in working differently so maybe they do patching or you know maybe they're putting Wi-Fi into into offices or whatever something that kind of Falls between boundaries right now and and and and work with them to to understand how they work you know maybe a deox team um understand how they work in that area multimodal change allows you to have role based access so only they can use that type of change so it's restricted um and you can work out your operating models you can work out what the pattern looks like for them and start scaling it so you've adopted some Fe new features in the platform but you're still over here you're still using normal chains the way you did for you know 99% of the business so that's how when I talk to customers about that that's the kind of approach that I take towards adoptions start with quite a few features probably but on a very small scale of who's using it I hope that answered your question Ain thank you Daniel um we are at time everyone want to be respectful it is now 9:00 am. Pacific Standard um just want to say thank you so much again for everyone who attended for blowing up our chat we'll do our best to have a recap out to you by end of week um as always feel free to send me emails comments what have you um and we'll see you all in August so you will get a new schedule sent this program is not being cancelled but the invite will be renewed um so yeah everyone enjoy the rest of your weeks thanks everyone