Skip to main content Search

AtlassianAIThe Service Management Show

The practical use of AI in service management

In this episode of The Service Management Show, Eficode UK's ITSM lead Gary Blower is joined by Atlassian's Simon Bradford to discuss the practical use of artificial intelligence (AI) in service management. They discuss AI's role in automating processes like ticket classification, chatbots, and predictive analytics, and the distinctions between rule-based automation, machine learning, and generative AI. Gary and Simon cover the challenges surrounding people, processes, and tools and recommend starting with pilot programs—integrating AI with existing systems, and maintaining a human-in-the-loop approach to ensure a balance between automation and human expertise. They also emphasize continuous learning, data security, and personalization to enhance user experience and share insights into future AI applications.

The practical use of AI in service management
Transcript

Hello and welcome to the service management show from the team here at Epic code and in this episode we're going to be talking about practical uses really of artificial intelligence in service management I'm joined today by Simon or I'll let introduce himself he you Gary good to see you again I'm Simon Bradford I look after the the itm or the JSM team solution Engineers inia for Lan yeah okay cool yes indeed um and I forgot to introduce myself I do this every show such an amateur and I'm uh I'm obviously I'm the IT service manager and practice lead for f UK and host of this podcast right so um yeah s just and I just having a really good chat actually just from here looking forward so um um all we want to do today is really sort of talk a little bit about um artificial intelligence in service Management in a more practical way in the way in which it can be used in a way it's going to really enhance your business and not necessarily the you know the chatbot or the scary stories that you might read in the Press um but start with so you want to just briefly talk about you know your history and your kind of exposure to things like artificial intelligence yeah sure go I mean it's an interesting conversation actually because I mean when you when you first messaged me Simon can we do a can we do a podcast and I'm like yeah what should we talk about should we talk about service management itle you know yeah let's talk about AI I'm like o what can I say about Ai and you kind of realize that it's it's part of your every day and you and you and you assume it's going to work for you but you don't really think how is it working for me so I was thinking back and some of the things I was thinking about is how we do virtual assistant and and chat at the moment and how that looks and I think back to my my experience I've worked in in priel for a number of years I used to work for a big Telco company and um we had um um internal support we used a chatbot the telefony was turned off because it was easier to use the chat bot and the only way you could get any any good answers and information out of it by was really scripting what you wanted to put in there to say what you wanted and then the the feedback we had was was di a lot of people saying that this isn't not working for us this is really not helping and it I think back then it gave it gave chat Bots and chat a a bad message and people were very very reluctant to to use them I mean I think I think forward now I mean there's there's a lot have changed the the the way I mean we look at virtual agent we're using now yeah I was talking to a customer who works from a utility company they' just implemented it and um he was well first of all what he did is he used Lan intelligence to look at the requests he was getting coming through already so first of all you you can do you can do a triage on the request you've got to see what category they are what type they are I think that's the biggest issue you've got a vast amount of data and you don't know what this dat is telling you yeah so he was able to help him see what the the biggest hitters in terms of tickets were were and if you think about service manager one of the key words that we've been saying for a while is is ticket deflection or or shift left so how can you get as much information that you're the resolution you've got close to the point of source and how can you help people help themselves and but by doing it in a way that you can assist them and one of the things that he was doing is he's like you know Simon we were getting people asking us to spin up AWS instances because they wanted to test some of the applications and I'm like oh that's interesting and I and I didn't know where he was going with this ation and he said look watch this and he showed me yeah he said look okay he goes can you pray a an AWS instance please for me um for for this this application and he's typing this into in into the slack Channel yeah and I'm watching what comes back and it comes back in seconds it comes back and goes I see you want to create an AWS instance okay um what account is this built to what application is this going to be before um who wants to manage it and it's asking these questions so he's already working out where the where the questions are going yeah and um and he said we'll update you when we've created instance for you and he shows me this and it shows him the update where it's coming through and in 30 minutes 30 Z minutes it produces a link to this AWS instances here's your instance that's been created yet and I was thinking back to this I used to work for a hosting company and we used to provide um Al synthesis for our customers that we used to set up it would take us days and that's using the portal approach and this here we've done it in like 30 30 minutes yeah so from a productivity point of view is great and you've not even hit a service desk agent so it's all it's all fully automated yeah yeah yeah and I think I think this kind of goes to the the heart of the matter really and you know we're talking before that I was saying that I what we call AI artificial intelligence actually been around for 50 odd years it's not a New Concept at all and all that's really changed is the compute power has increased and so you can do a lot more with it and I remember back in my student days quite a long time ago now we were writer we were using a programming Lang called prologue which is one of the like the first sort of mainstream AI languages if you want to call it that and all that really happens today is those those prologue transactions prologue basically you would create lots of relationships within it and then based on input that you put the AI would be able to work out the relationships and then give you an answer and that's just like the early stages of you know neural networks and machine learning all all they did when they they took those basic principles and they sort of expanded them out made them huge and added lots of compute power and they okay they put a decision engine in and what kind of stuffs but that's all they they did they went for kind of basic automation through to neural networks which are another form of automation but better S patent recognition then through to what we have today with generative AI which is just massive data sets and a lot more compute But ultimately it's still the same thing it's just automation I think you and you were saying that you know there was a there was a time people were saying oh it's going to end the world and you know it's going to make everyone redundant and but it's just a natural progression I think of providing better and more flexible Automation and I think your example is a good example of that no thanks go I think it's interesting because I know if you think about the the the areas that you and I work Gary it's about we a lot of people we talk to I wouldn't say they don't get technology but they don't appreciate how how technology can help them and I think one of the key parts of our job is how do you make technology accessible to people and like you said as soon as you start talking about AI on my word you're going to end the world I went for a um I went for a blood test this morning and I went to see my GP this morning as well so the first one was my blood test and the the nurse and I have a really good rapport she she's in her what late 50s really looking at changing her career but doesn't really know how to and what she wants to do is she wants to become an inspector for the air quality Commission yeah it's been an aspiration of hers and I always ask her how are you getting on you're still here I thought you had moved on by now and got and followed your career perect say she's oh no I can't I'm too scared I'm so I don't know how to do the exams I don't know how to how what what's required of me and there's a point to what I'm saying yeah so I'm like have you Googled it and and you can just see the fear in her eyes she's an excent nurse practitioner she knows her job she knows what she's doing she can she knows more about medicine than I ever will know yeah but when it comes to technology she's like she closes down and I'm like have you heard of chat GPT so I showed her chat PT GPT and I just typed in there what do I need to do to be a carequality examiner and it gave her a list of stuff and she looked at it and she could relate to that straight away CU she knew what it was she's like is it that easy I said it's not as much making it easy but it's making it accessible and empowering you to do that yeah so then I went on to my next appointment with my doctor and I got talking to him and he says Ah Simon last time we saw each other was this time last year I'm like yeah I know doctor coming for my annual checkup and it's like you told me that AI is going to take over the world where are we now and I'm like and I gave him some examples about how I use AI which I'll C off in a minute about I use it in in my job but I said for example with you year I've just walked in to see you see hundreds of patients already imagine if if AI was working for you you could pull up my notes you could see all my consultant notes it could pull out anything in there from the consultant letter that is important for you to see straight away and you could highlight that onto my record that's not taking your job that's helping you do your job so that all the time that you spend doing this admin work which takes you away from your patient enables you to speak to your patient more and understand them and look for the things that you wouldn't necessarily spot like how are they feeling what's their body language what's their demeanor whereas the Ari is helping you point out the stuff that that you would probably miss and I said for example what we do we use AI when we when we talk on on Zoom chats and I go through these AI transcripts afterward cuz there's always something I miss it's like having a second pair of ears that's listening you go oh by the way the customer said this and I missed on that and I can follow that up now and do that so this is where I think the the point I'm trying to make and and you get this as well Gary is how do we make our technology using AI more accessible to us so it enables us to do our jobs better and get on with the stuff that we want to do rather than the the admin and the mundane jobs that we could potentially automate and get er to handle yeah exactly I mean all three examples you gave there it is possible for someone to go away and find out that information in a kind of analog way you know that the first lady you spoke about she could go to a local library she could make some phone calls she could find out that information for herself so all the door chat GPT did was do the search for her read the information consolidate it down into a readable format and present it to her really quickly so it's that it's that accelerator it's that um it's that as we say sort of superpowered automation really which is what it's about rather than trying to replace what people do cuz ultimately the human is making the ultimate decision and I think when it comes to service management I think that's a key tip I I've been to a couple of sort of service management conferences where there's been like AI discussions and you know the general feeling as you I think you alluded to earlier with like bad chat Bots is that the chatbot doesn't feel like it's there for the customer the chatbot just feels like it's there as a barrier uh and that's often a poorly implemented chatot yeah rather than the example you gave of AWS where the chatbot is not really a barrier there it's an enabler you know someone asked a question and the chap kind of detects that this is a common request and it has it knows what to do with that common request and says well I can do that for you as opposed to sort of say know go away I can't answer that computer says no well it's interesting you say that because you probably heard more and more now we're talking about instead of service level agreements we're talking about experience level oh it's my favorite subject I thought it would be and this is where I think we start talking about what is your experience what is your experience like how has this been enhanced I mean I had a we use zoom virtual company um since covid all working remotely Works incredibly well I had some challenges with with zoom so Zoom was implementing updates very very quickly because it was keeping Pace I had some challenges with some of the things I was trying to do with zoom and I ended up checking to zoom technical support directly and they said look we need to install something on your machine to enable us to track the logs on what you're doing and I'm like yeah I feel very well but I mean it's a company machine I need I need to get at least some sort of approval that's going to happen so I went to the um we have call it help it slack Channel and I'm like um this is my problem Zoom tech support I trying to help me with a technical problem I've got on Zoom um they want access to my machine to install an app to track the logs am I allowed to do this and this is this is this is how literally how I typed it it wasn't I didn't type it in technical I typed it in for batim that's what and he came back and said looks like you're having problems with zoom I'm like yeah looks like you want to install an app to track Zoom logs yeah um we fully support what Zoom tech support are trying to do you've got admin rights on your machine please go ahead and do that and let us know if you anything else that was like 15 seconds Gary yeah it it was and this is it yeah and now I feel reassured and I could have probably made that decision myself but I wanted to make sure that I wasn't falling fou of any privacy or security issues that in the company and and that helped me out um and this is where to your point it it's an enabler yeah yeah and if you I mean if you flip this around I mean you made a really good point by saying well by putting a chat bot in place helps the company deflect issues to telephony straight away you make it's going to show you oh great we got no telephon calls coming out we must be working doing where the it's the classic water Bellon of the green on the outside right in the middle yeah exactly but I mean I think the thing is if you look at it from a service management point of view if you're a service manager you can be using at lasting intelligence now to be really looking at what is it in in the JSM space what is what tickets are you getting coming through yeah you've been given a challenge of reducing your service desk headcount you're not just going to cut your headcount you need to be really intelligent about how you're going to do that so use the alas intelligence part of JSM to triage and and and and categorize your requests are coming through so straight away you can then start seeing where your problem areas are and talk about problem areas you can turn these incidents automatically using that that's intelligence into PS post incident reviews for your problem managers so you can then really start get into the cruxs of what the what's causing these incidents you're going coming through and you can see what's happening in ter there was a volume of requests coming through the beauty of of of AI and L intelligence it takes massive amounts of data and it put us into a format that you could easily use yeah I mean one of the things and I was I was looking through this cuz I'm like right what can I what can I talk about and I'm just looking through my notes now yeah some of the things that I and I even found this out today um so we you said to me at the of the call were you away last week yeah we had a we had an offsite in San Francisco really good session and I'm like I need to share this with the team but I need to do it in such a way that it's quite succinct without giving them too much verbiage but the ability to JW down into it so me I'm I love conference pages so I write conference Pages I'm writing all this down I look at it I'm going it's to there's too much here yeah so straight away using at lasting intelligence you've got the the ability to shorten what you're writing you can summarize it you can summarize it as a viewer and you can summarize it as a Creator so you got two options there you can improve your tone so you can make itly it's amazing isn't it spell check I mean sometimes we come across when we talk Cally but when you write you want to be a bit more professional so you can professionalize it for you and make it a lot more a lot more visible and and easy on the eye but the other thing as well that it did is I was I was going to work yesterday and I'm like I've got jet lag I don't really want to work on Sunday if I can so I'm like let me do it Monday and I wanted to put it into a presentation that I could share with the team very quickly so I looked at the lasting intelligence because there's always new features coming out and had make this into a presentation and I clicked that and I and I kid you not what would have taken me I guess about two hours to write a presentation slide deck Etc it did it in about 10 seconds there was a couple of tweaks I had to make to it just to make it something that so you're still in control ites it doesn't take over you still got the ability to offer what you were trying to say and the message you want to put across but it put what I want into very succinct bullet points onto US onto presentation page on conference that I could share with my team easily and that they straight away can see that and as you're presenting this to your teams you can then go into more detail should you need to on on what that looks like now that that I think G is a that's a Time say yeah that's not taking my job away from me that's helping me and and our customers who use who who use our Solutions get the most out of what they're trying to do definitely definitely and um and you know that that um um the ability to change the tone of the language I think that's really import important for um kind of help desk teams where they're corresponding to a customer through the through the help desk through the portal because you can kind of write your response and then get it to of shape it into the format you like or even translate it which you know both those things you know that's just absolutely magical that you can do oh yeah the translation part is amazing yeah yeah and again this is just an enabler you know you you have said you know you've said what you want to say but you've just changed you've just made it more professional and more comprehensive to the person that's reading it so that hopefully the interaction you have with that customer is is better the other recent Innovation which um at last you have included and a couple other vendors started to as well is this concept of uh sentiment so when a ticket's raised this comes back to the experience manager I say it's my favorite subject um it's about getting to the heart if a customer asks you for something and you're providing a service sometimes it's not always obvious what the customer wants and also it's not always obvious how the customer feels about that particular thing are they frustrated that it's broken or it's not working are they you know are they in anticipation of a deadline and they need something urgently are they fairly relaxed about it and they're just sort of making a general request because they need something and that ability for the AI to kind of read what the customer asks and just to provide you a little flag to say what their sentiment sentiment is as a you know as a service manager that's fantastic for your team who doing the triage because they can immediately see the tickets that they probably need to be the most concerned about they can even organize it by sentiment and those sorts