ITIL TOOL DAY 2024 | Atlassian | AI in Service Management: New data separates hype vs reality
hello everybody and uh thanks for joining this session my name is Nadim matur I'm a senior Enterprise solution strategist at atasan so I have been in this role for about four years but in general in the itm industry for over 20 years helping customers in their digital and it transformation Journeys so at atlan I primarily focus on zero service management and how it can enhance service management practices uh on today's topic before we jump on any judgment on whether AI in service management is a hype or reality let's focus on a common theme between them so when it comes to AI uh literally it's a virtual assistant right who can mimic human intelligence so it's kind of a capability I would say so similarly like when you look at service management uh going to the fundamentals it's all about developing and using existing organizational uh capabilities to create value to the end users so what I feel is if they work in harmony they will certainly add up so let's explore it further before we dive in let me give you a quick rundown of our agenda today so I'm going to talk about the various influencing factors that matter in the world of service management and particularly the role of AI as a disruptor next service management is not just about break fix and customer support but rather it starts from the strategy on the need for services based on Market relev right to building operating and supporting those services so what could be the opportunity of AI in all these phases next I will go through AI powered capability particularly in jir service management that focus on providing the best experience and productivity for any role consider it user agent or administrator finally I have facts here I present the facts from thumb tag about why their teams are embracing AI features in JSM and the results that they are seeing now let us dig into why AI is the new technology disruptor for service management uh now when you look at the fundamentals of service management it basically you know it revolves around three actors service provider service consumer and services right consider a provider for any it organization that is catering to the needs of the end users that's a service provider right but if you look from the lens of Enterprise service management even the business units you know they fall under service providers you're talking about HR legal Finance all those units you know they have their own set of portfolio and the a user base that they need to support to service consumer now this is where the actual existence of service gets defined now just think it this way if there is no service consumer literally you're talking about no need of services right but at the center of all these three is what I I call it as value now this is the pive benefit seen by the end user or the customer and guess what this pursued benefit is going to change based on some external factors whether it is economic social legal environmental Etc um I'll give some examples here let's say if a country introduces some regulations and laws and if a service fails to comply with those laws then certainly you know the perception of the services is going to be dented right and uh another example is social behavior we have seen this not long time back right during pandemic like when the uh social behavior and Norms changed a lot of highflying services at that time they were like not relevant all of a sudden and then it demanded like a whole new set of you know new generation of services that are relevant to that new social behavior you know so my point is like these are some of the influencing factors but guess what the biggest influencing factor is technology and it is one of the constantly evolving landscape you know for the past two decades we have seen it moving from centralized to distributed computing virtualized to cloud computing the way services are built and provisioned has changed drastically now today every organization has their services enabled by digital technology there is no Denial on that and the recent adoption of mobile services it demands help at fingertips but now in the technology domain also there is a new player in the game and that is AI and that is going to kind of astronomically influence the entire journey of the service life cycle now let us explore further into how this new disruptor is going to impact the various aspects of the service Journey that I mentioned right and most importantly the opportunity that it is going to present now let me take you to this uh you know uh uh visual from um Gardner that it presented in the annual it Symposium so it called it as the AI opportunity radar now as you see from this it said like there are two uh domains in which the AI opportunity exists like gamechanging Ai and everyday AI now gamechanging ai is where organizations are going to use this technology you know to come up with novel approaches to Define product and services and also develop using core capabilities that we have never envisioned before right and on the everyday AI we are talking about users and knowledge workers basically adopting AI in their daily life you know to ultimately to increase the user experience as well as the productivity related to them uh now for a typical service provider that I mentioned whether it is it or business unit they are going to leverage this opportunity on the other two Dimensions external and internal now this is where you you know they will come up with a mix of right product and services and also how to enhance the user experience going forward and for internal basically we are talking about how they develop the core capabilities which is at the center of service management right so interestingly now many of the service management practices touches all of these four quadrants so let's start with that now first one is front office we all know like any organization when it rolls out service management practices service request manag event is the first one but sometimes even if you define the service catalog in its most perfect format users will still have some issues you know trying to find information and trying to get to the correct catalog item this is where natural language processing and conversational guidance assisted by AI is going to enhance this practice Incident Management we have seen it sometimes incidents happen and it's the same type of incident goes through the same activity you know same escalation routes no Lessons Learned right but this is where AI is going to suggest like if there is a similar incident is going to kind of tag those and provide the resolutions at the earliest so basically we are talking about you know improving mttr and cutting short the resolution times service desk now rolling out a 24x7 service desk is always resource intensive and also expensive what if you could just roll out like a virtual service agent who is is going to handle you know all those informative task as well as resol if if it can in in its capability and on the back office there are more practices change management now this is interesting so when you consider like developers and operations there is always a friction between them you know developers they want to get the code at the earliest into production and operations guess what what do they need stability right and in between these two is kind of a cab which is stuck and in nowadays it's like very hard to scale to the changes that happen in the complex microservices environments so this is where AI is going to come with risk-based analysis and trying to you know separate those changes into standard changes that they can just go in the fast lane and also normal changes that you you know have to go through the cab so basically going to maintain a balance between speed as well as stability that's going to be a huge win problem management now typically what I have seen is organizations tend to do reactive problem management an incident happens you go to the root cause of the incident and then you make sure that the incident doesn't happen so it's reactive right but guess what there is no shortage of data whether it is from observability Tool whether it is from past incident or even from the configuration management database now this is where uh some of the capabilities of AI are going to come and give kind of you know some leading indicators on what's going to happen down the road you know so this is going to shift the whole concept of problem management from reactive to proactive problem management software development now it touches you know entire sdlc so basically are talking about you know using AI in source code reviews pull request and anywhere in the cicd pipelines and on the right side it's pretty much you know like all those practices that lead to coming up with the novel uh product and services are also increasing the core capabilities by using the continous service improvement um process but at the core is Knowledge Management now what I have seen is in in most of the organizations this is sometimes it's kind of the neglected practice you know oh you know what our our knowledge source is so distributed and we cannot leverage this our knowledge database is not up to date so these are some of the challenges now with AI at the doorstep Knowledge Management is going to get it due share so where does all this lead to at least in the world of service support and service operation what I would say is it is in some form going to kind of Kickstart the shift left methodology you know the challenges right involved in a typical tier model frequent escalations loss of contacts delays in resolutions now ai is going to act as a catalyst for the shift left concept to get help where it is needed into the hand of the health Seekers now and what does a successful shift left going to look like we usually focus on typical service level agreements right just metrics prove to the end user that we have met the metrics but that's not the case it's going to change the landscape from typical service level agreement into a more holistic experience level agreement and the good news is J service management is well equipped with the capabilities to make this concept a reality so let's dig into some of the AI powered features in zero service management atlassian is unleashing AI for service teams in three key ways for help Seekers It's all about getting help exactly where and where you need it one seamless experience no more digging around in different systems for agents we are using AI to care take care of repetive stuff and make knowledge instantly accessible so they can focus on what really matters which is helping people helping their colleagues and for admin it's about you know just quick setup easy roll out across the business because we know like admins they have got enough on their plates so let's start with the help Seekers and the virtual service agent so we believe that you should be able to get help where you are whether you are in slack or Microsoft teams or even email or the classic Help Center or portal so virtual service agent is there to assist you in a true Omni Channel landscape so users can ask for help in any one of their favorite engagement channels whether it is slack Microsoft teams and if they wish to log into to portal and click on any notification in the email the entire conversation Trail is available in the ticket no context lost let's draw our attention to the support agents experience you know another aspect like knowledge workers care about is AI powered process automation so with the click of a button you can call the service request helper Robo by the way Robo is an atlas an intelligent powered agent and simply ask questions on the current case the agent is going to catch you up on what happened on the ticket since yesterday for example get up to speed on previous resolution attempts find subject matter experts who worked on similar request before that could assist him help identify some of the next step on the part to resolution and even draft a complete response for them based on conversation the human agent has with the robo agent so it's an undeniable fact that today we work in a distributed uh Workforce in different time zones and follow the sun is a norm today so when age and time zones change it doesn't mean that the life cycle of the ticket has to stop there but it the flow has to continue and you know without losing context that's very important now for a typical agent who work in this Global model features like this are going to be a dream Crum tool and it doesn't stop there so when we are working with end users tone of communication matters you know whether you draft a professional response or an empathetic response at last and intelligence is going to offer you assistant at every step AI powered agent assignments you know that can intelligently route issues to the right person Based on data driven recommendations so no more tickets sitting in the queue for tier one to pick up this is going to be a huge boost for response times and also on the resolution it could just speed up right when the alerts are detected you know separating signals from noise so we have seen a lot you know like sometimes it's like a false alarm and 10 false alarm is going to 11th false true alarm is going going to be totally neglected you know like so intelligent grouping basically based on alerts is going to be a huge benefit now what about admins what can AI do for them now with Atlas and intelligence you can easily start an automation rule using natural language support yes no more coding needed no more to even write a pseudo code or the flow of event just describe your problem or just describe your requirement in plain English AI is going to do do the heavy lifting for you and is going to come up with the automation rule which easily and quickly creates the automation rule for you but you can always review it so so that you remain in control of that automation Rule and with Robo you can call upon a robo agent to provide your service team with the help they need and quickly respond to a ticket significantly reducing the time to First Response yes all of this through Automation and when call upon a few agents you know to keep up the knowledge base up to up to date so so that you can analyze the existing close stickets and drafting articles that were missing for example and recommending changes to the existing knowledge articles to close any Gap in your knowledge base so that's why I mentioned Knowledge Management is not going to be a neglected practice anymore so it needs that data and again when it comes to AI the first thing that need to be fixed is the Knowledge Management with a good Knowledge Management you're going to get good insights through AI but if the Knowledge Management is outdated then AI is going to give you the wrong information right and when you are rolling out a portal to new business units like HR legal why not take AI assistance so we have ai powered help desk Builder you know with a simple click of a button it guides you through request type suggestions building the underwinding workflows and L how the portal ready to roll out literally you talking about building heless solution with a click on of a button and finally when you need those insights to make informed decisions ask your questions in plain English Ai and atashian analytics will get you started with the autogenerated SQL query so these are just a few examples of the power of AI in zero service management next I will take you through a true story from one of our numerous customers who are already seeing tangible benefits adopting AI on this front thumb tag it's a technology company that helps millions of people confidently take care of their home improvement initiatives small fixes to Major improvements so being a technology company any downtime of their service means lost revenue and more importantly tarnished customer experience so like most organizations they have lean teams and were looking for ways to do more with less so when they thought about where to get started with AI they had a simple goal optimize the existing it service management process and make life easier for their employees and end user pretty nice goal right so here is the preview to the scale of support at uh thumb tag so they have around 500 project categories ranging from home services to catering and do grooming so there are like 300,000 professionals using the platform to offer their services over 80 million projects have been initiated on Thum tag and the organization it supports around 1,000 employees worldwide so this should give you some cont context on the size of their operations on the next slide I'm going to talk about when they decided to implement Ai and why they choose to implement it with J service management so in 2023 hit by macroeconomic conditions they had challenges you know expanding the headon for teams so this forced them to explore ways to become more efficient with their existing tools uh as an existing atlassian customer it made sense to them to turn to J service management for their itm processes so so they went live with JSM in 2023 uh and an already existing knowledge base it enabled them to scale the operations much quickly so they used AI to streamline service request automate resolutions and provided 2 24x7 support by virtual service agent so this approach has enhanced their SLA compliance and increased customer satisfaction score to 4.9 on a scale of five and additionally their virtual service agent automatically resolves 15% of the tickets creating efficiencies and allowing agents to focus on optimization and Innovation such as improving their knowledge base and this is exactly what I was talking about in the shift left methodology make the work move to the closest to the user basically so they have saved around 180 hours per month by switching to JSM and automating their tasks so in the next slide I will review some performance metrics of their virtual services agent for the second half of 2024 th tag develop an initiative focused on optimizing their knowledge base and improving matches and resolutions using virtual service agent so they set a base goal of achieving around 20% resolution rate and they have reached so far 28% and they're aiming for a stretch goal of like 30% by the end of the year which they are confident that they will be achieving and on top of it they are inching further closer to a perfect cat score the next slide I'll provide behind the scenes look at the conversation of the dashboards now JSM provides an intuitive dashboard analytics related to assisted conversations as part of the process to review and optimize the results of the virtual service agent uh they examine the past conversations to identify any gaps where AI for example did not successfully resolve the issues then they create intent flows and knowledge articles to enable kind of automatic resolution of similar tickets down the road you know this reminds me of the classic continual Improvement example so it's not like uh you have the perfect AI Solution on day one it's basically you're talking about incremental Improvement based on what users are asking uh to the virtual agent itself and the return on investment speaks by itself so they have so far 60% annual cost savings and they have boosted their SLA by 90% and they are as I said automatically resolving 15% of their ticket with virtual service agent so AI in service management is a reality for sure time to get on board and if you want to learn more about AI capabilities in atlan platform please visit this website atl.com platform artificial hyen intelligence thank you and let me know if you have any questions NAD excellent very good presentation thank you very much we do have a handful of questions um just before you went on to the case study with Thumbtack you mentioned just just use Simple plain English with the AI um what about in a multilanguage environment how do you manage consistency of answers uh in the same situation across different languages yes it is totally user intuitive basically we are talking about uh you ask the question in any language of your choice even the interface is English AI is going to translate your language into English make the matches make sense of the answer and get you back the answer back in the original uh language that it was asked so basically it is like totally seamless you know like user will feel as if I'm I'm kind of you know conversing with the AI agent in the language that I'm comfortable with so it's totally supported yeah okay and it's kind of an extension to that question um how you adhere to the different legal Frameworks um around the AI in different countries um yes so basically you know like we will stick to the Frameworks and the regulations that uh that are relevant to to those countries and uh there are some limitations although I will say related to data residency and all those things uh but in general yes it complies with uh with those Frameworks as well okay yeah thank you um just how how reliable and with what effort is AI in predicting fitting and really risk reducing change models without high quality change requests input um and is it normalized and reconciled in order to predict a correctly fitting Change Model yes very good question you know so when it comes to uh uh you know performing risk analysis most of the data it it kind of resides in uh the the configuration management database that you have uh constructed so basically you're talking about how do you want your risk to be calculated so you could just Define a number of parameters right in your cmdb and tell like how how do you come up with those risk scores so that you could say like if it is a risk score of below five then I would classify them as standard changes and if it is a risk score of you know more than five more than six something like that I would just classify them as normal changes so it's all based on the existing data in your configuration management database as well as the environment of course and then it's is kind of you know what are those uh algorithmic metrics that you would Define to classify these type of changes so it's going to take all those into consideration and then spit out a score and then you could take that change from there onwards okay there there are a number of questions here and I'll ask just one more in the last minute remaining um how how is AI helping um service catalog management can it automate service request creation for the user through natural language for example software installation request okay that's a good question now there are two parts in this one one is for admins you know like how do I create my service request for example uh you know nowadays it's there is there are cases you know where it used to understand the business now it's kind of you know other way around even business people they understand it you know like they know what technology is so if you can give them the concepts of you know how you could spin up your own help desk then they need they don't need to do any coding literally they're talking about if I am a HR representative what are the typical catalog items that I need to give it to my users this is where you know they come up with those uh uh help Des Builder that I showed a while ago and ask that question and come up with those catalog items and then review them now for the agent you're talking about automating the flows so we have a powerful automation capability in JSM that's going to you know if it is related to uh you know automatic fulfillment right it just connects to the third party systems uh and you could easily fulfill some of those task which uh you don't need to involve any manual work over there so there are numerous ways on this one and again when it comes to end user you're talking about you know how they can interact with the system and get their task fulfilled by themselves so it touches all the three roles I would say so but it does help on that front yeah perfect NAD once again thank you very much for your time H and your contribution today is very informative thank you
2024-12-25 14:16
