Moving IT Service Management Beyond Tickets With Xurrent

[00:00:00] [SPEAKER_01] The leading issue of agentic AI in businesses right now is ensuring agents act with compliance guidelines. And Denodo applies guardrails across your entire data estate. By aligning your company's data infrastructure under one system, these guardrails perform consistently across your platform. So start scaling your business and start with Denodo. Simply visit denodo.com to learn more.

[00:00:29] [SPEAKER_01] What happens when an IT team hits every service metric while the rest of the company remains deeply frustrated with the technology? Well, in today's episode of IT Infrastructure as a Conversation, I'm going to be speaking with Phil Christensen, Chief Product Officer at Xurrent. And we're going to talk about moving service management beyond tickets, queues and SLA reports. Yeah, I'm going to encourage you to think a little bit bigger today.

[00:00:59] [SPEAKER_01] And my guest will also share a costly example where two teams working with different pricing data actually cause Google to remove a major retailer's ads for two days. And we'll also discuss why shared context matters, especially during incidents, how service outcomes reveal what ticket counts miss. And why yesterday's processes built for a different era can quietly become tomorrow's employee burnout.

[00:01:28] [SPEAKER_01] I want to also cover how AI can improve routing and support, why virtual agents need human auditing, and how the role of an IT admin could change as repetitive requests continue to decline. All of which is a useful reminder that numbers never tell anything. There's a great book about statistics called How to Lie with Statistics, which perfectly sums up that problem. But that is another rabbit hole to go down on another day.

[00:01:59] [SPEAKER_01] Today, I want to talk about how the smallest communication gap can create the largest operational bill. And I have the perfect guest to talk about it. So enough from me. Let me introduce you to Phil right now. So thank you for joining me on the show today, Phil. Can you tell everyone listening a little about who you are and what you do?

[00:02:20] [SPEAKER_00] Thanks, Neil. Great to be here. Excited to have a conversation today. I'm Phil Christensen. I'm Chief Product Officer at Xurrent. Xurrent is an ITSM platform as well as an incident management platform. We've been around for about 15 years or more. We focus on the middle market primarily, operating with a large customer base in Northern Europe and recent expansion into North America.

[00:02:44] [SPEAKER_00] My job here at Xurrent, as you might expect with a title like CPO, is to establish the strategy and vision and ultimately the roadmap for where we're going. You know, I think it's a really an exciting moment in the world for product managers as AI takes on more and more of an outsized position in engineering organizations. It seems that PMs are becoming even more and more relevant these days as it relates to turning requirements and impact into functionality. So exciting moment, I think, to be in product.

[00:03:13] [SPEAKER_01] Yeah, I completely agree with you. I'm glad you have the time to sit down with me today. A lot I want to talk with you about. And when we look at IT service management, it's often associated with things like tickets and workflows, maybe even a sigh and a groan thrown in there as well. So what does it mean, though, when we put the technology to one side and maybe treat infrastructure as a conversation instead?

[00:03:37] [SPEAKER_00] Yeah, I think that in my experience with infrastructure in particular is that generally speaking, as opposed to even into traditional IT service, infrastructure is one that is often a crowdsourced area where wherein when issues occur, when there are problems with infrastructure, the solution can often come from a lot of places.

[00:04:00] [SPEAKER_00] And I think ultimately the sort of the way you go about determining the solutions, the way you go about triaging issues with infrastructure is often much more of sort of a collaborative process where it's not about figuring out, did you meet an SLA? Did you do something wrong last week? Did you not follow instructions? It's all about the now. It's about let's get this thing fixed. Let's get the website back up or whatever has broken.

[00:04:25] [SPEAKER_00] And let's put aside all the baggage that might have come with that and let's work together to make this happen. And I think the best companies I've worked for are ones in which you can create that sort of safe conversational environment where everybody feels like, look, they're not we're not here to point fingers and to talk about our skills from the past. We're here to problem solve. And I think that's what's different. And I think specifically about managing infrastructure and large enterprises.

[00:04:54] [SPEAKER_01] It's interesting you said about pointing fingers there. I think very often when things go wrong in IT, each team will dig their heels in. It feels almost like that Spider-Man meme with IT ops, security applications, teams, developers all pointing to each other at the same time. It's not the network. It's not the application. It's not my code. And one word that we're hearing a lot in AI at the moment is context.

[00:05:18] [SPEAKER_01] But bringing that word context into, let's say, IT infrastructure, do you have any examples or stories from the field there where better context between all those technical teams I just mentioned could collectively change the outcome of an incident or even service requests?

[00:05:34] [SPEAKER_00] Yeah, when we chatted before, I had seen this question that I had. I was like, yes, I have a great example for this from my past. I worked for many years in e-commerce and in e-commerce, it was a large website for buying furniture and a very large organization and selling millions and millions of dollars of furniture every single day. And when you work at that scale, you ultimately have to specialize. It's the only way these large organizations can function.

[00:06:01] [SPEAKER_00] So I was a member of the pricing team, but the actual experience of buying something on this website brought together hundreds of different teams, whether you're dealing with the imagery on the site, whether you're dealing with the actual checkout experience, the credit card processing, the ads that go up into Google to make sure people are drived into your site. You're dealing with many, many, many different teams. It's really quite difficult to have context, frankly, for the entire story of what's going on during that experience.

[00:06:31] [SPEAKER_00] We had an incident wherein we were providing prices to the website and those prices were changing on a daily or weekly basis as economics, underlying economics change, things like shipping costs and costs of support, etc. So we would change the prices and unbeknownst to us, the ads team would then take that information and they would send it over to Google to put up PLAs or product listing ads they would buy. It's a common process in e-com. You buy these ads and drive traffic to your website.

[00:07:01] [SPEAKER_00] And we didn't realize, but they were using a stale version of our pricing data and they were sending the wrong prices to Google. And Google would then turn around and they scan your website and they make sure the prices on your website match what you sent them. They want to provide a solid experience to their shoppers so that when they click an ad that says, oh, I like this $100 couch, they don't go to the site and it says $150. There's a trust that they've built and that's why people use Google Ads.

[00:07:26] [SPEAKER_00] And we basically broke that trust, not intentionally, of course, but we broke it because we had this disconnect between these two teams that weren't talking. We didn't have the context and Google actually blacklisted the site for two days and took down all of our ads. And this is a company that spends literally a billion dollars with Google to make sure those ads stay up. That's how serious they take the trust. And the idea, the ultimate problem was quite simple.

[00:07:55] [SPEAKER_00] And it was just about these two teams coming together and talking and saying, no, no, don't get the prices from there. Get them from here. And here's when they change is simple context would have eliminated this entire problem. But these two teams weren't talking. And I think it's just a good example of when you do create environments where you can have these conversations, where you can know the broader context. It can really help eliminate problems.

[00:08:21] [SPEAKER_01] Wow. Incredible story there. I think it really brings to life what we're talking about. And another mantra that maybe all IT teams can unite on and follow is you can only improve what you measure. And knowing what to measure, knowing what metrics to use, that is the tricky part. So what measures have you seen that show that an IT service platform is improving the business rather than just simply processing more tickets? I think it's great to dig a little bit deeper than that. Yeah.

[00:08:50] [SPEAKER_00] Yeah. One of the beautiful parts, I think, about the Xurrent platform in particular, not to get too much into us and I'll get to your question, but it is that it is designed to be a service oriented platform. Meaning rather than just having a list of ticket types, you know, broken laptop and add memory to server, we orient around the service that's being provided, whether that might be employee productivity or employee onboarding or desktop support.

[00:09:17] [SPEAKER_00] These are designed to empower the IT teams to know that they are providing something bigger than just a ticket. They're providing a service that is actually connected to the business and not just ultimately a cost. And the quality of that service then is going to impact those bigger ideas like employee productivity.

[00:09:37] [SPEAKER_00] If you are an IT leader and you're charged with making sure employee productivity is paramount at your organization, instead of, for example, closing 100 broken laptop tickets, it's going to inspire, I think, people in a different way. It's going to bring a different sort of level of understanding to everyone on that team from top to bottom as to what they're doing.

[00:09:59] [SPEAKER_00] And then you can start to look at things like, and we do this, we measure this in our system, things like customer satisfaction and MPS. And are you providing a service that people are liking? I've talked to customers who we have, we have a very friendly CIO and a large customer in MSP in Europe. And we invited him recently to talk to the team and they asked him a similar question about KPIs. And he said, I don't ever look at KPIs.

[00:10:28] [SPEAKER_00] And he sort of said it facetiously. He was sort of joking that he's like, you can have the best stats in this IT system. They can show you no SLA is breached. You can look at ticket volume going down through first call ticket resolution. You can see all the right numbers and the entire company can still hate the IT team. So true. Because you just, frankly, yes, these are these sort of old school numbers around SLA breaches and can be, they're important.

[00:10:58] [SPEAKER_00] You need to run your business and look at these. But ultimately, it doesn't replace that sort of feel that IT has in the company. And for that, I think you actually, in some respects, you need to look at these broader metrics and you just sort of think outside of the KPI box a little bit.

[00:11:14] [SPEAKER_01] And before we jump to the future vision and including AI into everything that we're talking about here, most businesses, especially people listening, will have a fair amount of technical debt. It's a phrase that we're still hearing all these years later. But before we even get to technical debt, I mean, where do legacy processors, where are they creating the greatest friction when organizations modernize service management?

[00:11:39] [SPEAKER_01] Because very often I've seen a few examples of people implementing AI, but just adding them on to legacy processors that were built for an entirely different era. So what are you seeing here?

[00:11:50] [SPEAKER_00] Yeah, I think that legacy processes come from a lot of sources. Oftentimes businesses grow and the processes they put in place were designed for an organization that doesn't exist anymore. I think that'd probably be the biggest area I see where, you know, the problem you decided not to document because you only get one a week and, you know, Joe knows how to fix it. Suddenly you grow, two years go by, and guess what? You're getting 100 of those a week.

[00:12:20] [SPEAKER_00] And, you know, going to Joe doesn't work anymore. He doesn't even work here anymore. And that's sort of the theme I often see with our customers. It's really about thinking about the future, thinking about scale early on so that the processes you put in place, frankly, grow with the company. And, you know, having some systems in place, having an IT service management product in place that can help you establish those best practices early, I think, is really important.

[00:12:48] [SPEAKER_00] I mean, at the end of the day, you also can't avoid this. I think that having a realistic understanding that, look, every few years we need to think about what's in place. Think about where we are burning people out. Maybe Joe's still here and guess what? He's working weekends to keep up with that. That's not a good situation for anybody. You know, you're not going to keep him. He's not going to be happy. And you can imagine that sort of growing across the whole company.

[00:13:13] [SPEAKER_00] So we also encourage customers to every, say, 18 months to reevaluate what do they have in place? Do they make the same decisions about the processes they put in place now that they put in place 18 months ago? And not let it get, not let it fester, you know, going four years without ever looking at the request templates that you have in place and the services associated to those. Like, you're going to wake up four years from now and, like, and hate whatever system you're using just simply because the business has changed.

[00:13:43] [SPEAKER_01] Yeah, and I suspect people listening all around the world, every organization has a Joe there, that go-to person when they need when something goes wrong and they tap on their shoulder. But, I mean, when we started bringing technology into this and automation, et cetera, how should teams decide which kind of requests that they can automate and which still need that very human judgment? Is there a right or wrong answer to this?

[00:14:09] [SPEAKER_00] Well, I think what we've done, the approach we've taken is that generally we sort of let the end user make that decision. So we will, you know, we are attempting to provide, we have a knowledge base as a lot of tools do. We're attempting to provide information that's gleaned. Like, LLMs now and AI is so good at looking through large data sets and being like, oh, this person's asking this question.

[00:14:34] [SPEAKER_00] Guess what? It's in section nine of this knowledge article that I never looked at, but I found it in a millisecond. So, you know, let's serve it up. And, like, let's see, does this solve your problem? And if it doesn't, then give the customer an opportunity to say no. Sorry. Thanks. Try again. And then, you know, as fast as we can get you to a human. So we've sort of taken a similar approach both with end users as well as the tools that we provide to our actual IT teams.

[00:15:02] [SPEAKER_00] Those are all about bringing you efficiency, bringing you tools and trades to do your job faster, better, you know, in less time. So things like summarization, routing to the teams intelligently, not needing a tier one person to route that ticket out to the right person, having, you know, automatic detection of duplicates or automatic detection of poorly written requests that need more insight. Like, basically bringing tools.

[00:15:29] [SPEAKER_00] So I wouldn't say we are, I don't know, I kind of reframe it a little around, like, there aren't specific functions that we are replacing. It's more about continuing to bring more and more efficiencies to all the excellent people out there that are doing these IT jobs day in and day out.

[00:15:47] [SPEAKER_01] And, of course, if we were to take a peek in any enterprise anywhere in the world, they just cannot help creating silos, whether it be locking down shared drives or access to data, etc. So what gets lost, though, when service management tools are designed around those infamous departmental boundaries rather than the employee's problem that they're looking for a solution?

[00:16:10] [SPEAKER_00] Yeah, I think the biggest thing we see when companies have these silos is, I guess, two things I'd call out. One is that I do think it becomes harder and harder to maintain accountability. You know, ultimately, the reality is that while you have silos between HR and IT and others, they often are working together. And someone, you're onboarding an employee, they're going to need to get help from HR, probably from facilities, from other organizations. And ultimately, it's not as siloed as you might think.

[00:16:38] [SPEAKER_00] So passing things around and making sure that, you know, you get this from that team within a reasonable time frame can get harder and harder when you have silos. I think the other one that maybe is a little tangential, but I think is still an interesting component, is that it actually can become quite expensive for organizations. Because what ends up happening is that these different organizations solve the same problem in different ways.

[00:17:03] [SPEAKER_00] And often the same problem, maybe oftentimes, excuse me, the solving of that problem involves spending money, buying some other system, buying some other tool, buying a subscription to a service. But guess what? Whatever you just bought this subscription to solve this problem, this other department already solved that problem, but they bought it from a different vendor. So it actually can become a very expensive prospect for organizations. Beyond the efficiencies lost, it can become quite expensive to have silos, especially as you get larger.

[00:17:32] [SPEAKER_01] And I've been to a lot of tech conferences this year. I've seen multiple vendors all promising big AI solutions, AI that can easily summarize and reroute work automatically. But for many techies listening, I've spoken to a few of these in the audiences and chats on the show floor. What risks appear when it might act on incomplete or poor quality service data or missing context even?

[00:17:59] [SPEAKER_00] I'll answer this with an example, because we obviously, as you would expect, we've added a lot of AI to our tools. We've also joined a lot of customers on this journey of enablement. And we've discovered that, honestly, the enablement side can be just as hard as building these capabilities in the first place. And one enablement anecdote I'll share with you is around the use of virtual agents and the use of these quote-unquote chatbots to help answer questions.

[00:18:26] [SPEAKER_00] What ends up happening with these tools is that you have this AI talking to your customers. And since it is AI and it's leveraging these amazing LLMs that have gotten better and better and these great frontier models that are getting better and better, you actually don't know what it's saying. And you don't really have much control as an IT administrator over manipulating that output. Like, you sort of are giving in to the trust that this thing is good at what it does.

[00:18:51] [SPEAKER_00] But ultimately, having that trust is – and yes, they're getting better and better. And there's some truth in having trust in just this black box that is these AIs. But most IT teams that I've interacted with, they want to have some control. They want to know that, okay, here are the common questions that are being asked in the chatbot. And here are the responses. And yes, check, those are the right responses. And oh, guess what?

[00:19:18] [SPEAKER_00] We found that here in this situation, it's referenced a knowledge article that was wrong. And it's because that knowledge article wasn't written well or was written three years ago. We don't agree with it anymore. And so we've built tooling around this exact idea of let's bring administration and control to IT teams so that we can expose these gaps, we can expose these blind spots as opposed to keeping them hidden.

[00:19:43] [SPEAKER_00] So we can say, look, here are the 50 common things your customers are asking for. And here's the response that we gave them. And do you agree? What do you, Mr. or Mrs. IT administrator, think the right response should have been? And you can fill it out in our system. You can say, here, I think it should go here, here, and here. And then you can run it. And you can run it as many times as you want.

[00:20:04] [SPEAKER_00] And you can see, you can sort of build this roadmap, if you will, for your company to get better and better at using these AI tools. So I think it comes back to there's been a lack of enablement capabilities. There's been a push to bringing AI in. But there hasn't been a push to helping the administrators use that AI.

[00:20:27] [SPEAKER_01] And before you join me on the podcast today, I was doing a little research on you guys. You just mentioned AI there. And I think when attempting to get context right, you need to bring everything together. And I know you've created an AI-powered platform that unifies ITSM, ESM, incident management and response and status pages, and putting them all into a single system of action. And the result of this, from what I was reading, was you bridge silos between service desk operations and DevOps,

[00:20:56] [SPEAKER_01] much of what we've talked about already, and create this almost shared source of truth there. And the results that I was reading as well, I think this embedded intelligence fabric can automate 50% of service desk tasks and route 60% of requests automatically. Some big numbers in there. But tell me more about that and what you've seen.

[00:21:18] [SPEAKER_00] Yeah, I think that having, there's a couple things. One, you know, ITSM and incident management generally is just very well positioned, I think, to take advantage of a lot of the tooling that I've talked about, broadly speaking. Just these tools are sort of a perfect match for these capabilities. I also think that as organizations grapple with how to use AI and how to enable their organizations to AI,

[00:21:44] [SPEAKER_00] they're going to need tools like ITSM systems that sit in the middle and have a view of what's happening to really make sure that you're doing it in a compliant and sort of auditable way. So that when you do make changes, just like when you used to make changes, they're still being controlled in a system where you can sort of see them or you can roll them back, where you have some visibility into them. So we've had great luck. I think the impact it's having on our customers is pretty obvious. You quoted some of the metrics.

[00:22:12] [SPEAKER_00] A few others I'd throw out there is that we've really taken an approach here where we want everybody using this. I think this is partly why we're getting some of these metrics. We want all of our customers enabled, and we are not monetizing AI, which is quite unique in this space. We're specifically not tokenizing it or making it hard to use or putting wallets on people. We want every one of our customers to use it at every step. I think it's been very hard with the push for monetization.

[00:22:40] [SPEAKER_00] Many companies are having to evaluate on the same level, do I want to have a summarizer versus do I want to have an agent that takes an action? Those are two very different tasks for an ITSM to do, but they've all been, you're having to choose whether you want to pay for those through these monetization efforts. We're not doing that, and we want everybody to use it. We have over 90% of our customers who have enabled AI in their environments. I think that's on the back of they trust the futures that we're building from a security and compliance standpoint.

[00:23:07] [SPEAKER_00] They also believe in the fact that we're not going to suddenly pull the rug out from under them and look for a way to monetize. I think that's pushed us in this direction. I think it's a matter of our approach to AI generally, the fact that we've gotten broad adoption. Then ultimately, just these AI capabilities are very strong.

[00:23:30] [SPEAKER_00] Giving one more stat to fill it out, our virtual agent that I've been referring to is in use at scale at a number of our customers. You talk about things like ticket resolution, first call ticket resolution. I was blown away. One of the stats they shared with us was that when they rolled this out, they have a bunch of grocery stores, and they have a fair bit of hourly labor at the grocery stores. Fairly not that technical. Think like cashiers and things like this.

[00:23:59] [SPEAKER_00] They used to have to go into our system when they had a problem. The cash register is broken, and they'd have to navigate through the service hierarchy, find, okay, cash register support, broken keyboard, and they put a ticket in. Now with AI, and oftentimes, unfortunately, as they navigated that, they wouldn't end up at the right ticket. They'd go down, oh, you know, broken monitor instead of broken keyboard just because they're moving fast.

[00:24:22] [SPEAKER_00] Now with AI, they told us that the AI is more likely to get them to the right team and the right request template than the human, which I think is just like a somewhat terrifying but also really impactful idea that we can make the experience better for our customers through these capabilities.

[00:24:43] [SPEAKER_01] There will be some techies listening that will quickly jump to the conclusion that, hang on a minute, AI is removing my role here, but I would argue it could also create more opportunities for things like being a project resource because those projects are always backed up, and it's better to be part of a value add than just fixing the same thing every day. But from your point of view, how does a well-designed service experience, how's that changing the working lives of those IT teams and the employees that they're supporting?

[00:25:13] [SPEAKER_00] Well, it's exactly that, and that's why I was about to interrupt you. I was excited to share, but you asked the same question. The tasks that we've introduced on this enablement side that I was talking about is changing the landscape for what IT teams are expected to do. And this is the part that I think can be missed in this belief that, oh, suddenly AI is going to take jobs or replace people. I think it's quite the opposite. I think if anything, it's bringing new tools that need to be managed, that need to be enabled.

[00:25:41] [SPEAKER_00] You don't just turn on a chatbot and it works tomorrow. You need it, as I was describing before, you need someone there who is auditing it and making sure it's sending people to the right places. We showed we have this AI, we call our AI Sarah. We have a Sarah studio that I've basically have been describing that allows you to see all these responses in the direction that it's sending people. And then it's got all these tasks. You can see all those red and yellow and green about here's where you have good data. Here's where you have bad data.

[00:26:11] [SPEAKER_00] We've had customers tell us, oh, my God, this is perfect. This is what I need to turn. This is what I need to get this going. But I don't have anyone on my IT team that has time for this. This is nobody's job. We have no one's job responsibility includes configuring the AI layer. Because guess what? That job didn't exist six months ago. So they were not saying they weren't going to do it. They're just saying, well, hang on. We need to sort of adjust goals and expectations of our employees.

[00:26:37] [SPEAKER_00] And maybe as inbound tier one tickets go down, AI administration goes up. So I think it's a rebalancing and a new set of expectations. But the IT administrator is just as necessary now as it was three years ago.

[00:26:55] [SPEAKER_01] I think that is a thought-provoking moment to end on. I have thoroughly enjoyed talking with you today. I think we busted a few myths along the way as well, which is always good. But for anyone listening who wants to find out more information about anything we talked about today, we mentioned your AI-powered platform there. Where should they go if they want to find out more?

[00:27:13] [SPEAKER_00] Well, if you didn't get the name off my t-shirt or my background, the company is Zerent. You can go to www.zerent.com. You can also find me on LinkedIn, Phil Christensen. I'm always happy. I love to have conversations with customers and prospects. So thank you, Neil.

[00:27:30] [SPEAKER_01] So I will add links to everything that you mentioned there, plus a few other things. I've run the show notes and the blog posts associated with this episode at techtalksnetwork.com. I encourage people to check it out. There's some great things over there, and I, for one, have learned a lot from talking with you today. So thanks again for sharing your story. Thank you, Neil. Appreciate your time. I think Phil's examples today show why modern service management cannot be reduced to just faster ticket closure. We need to think bigger than that.

[00:27:59] [SPEAKER_01] Because yes, an organization can report a healthy SLA performance while employees, they continue to experience slow support, disconnected teams, duplicated tools, and answers based on outdated knowledge. In short, that's why your IT team is not very popular in the organization, despite you ticking a few boxes on that SLA report. But there's a bigger lesson here, and that is to connect each request with the service

[00:28:29] [SPEAKER_01] and business outcome behind it. Then review the process before growth turns one person's workaround into a company-wide dependency. And AI, yeah, that can help people find answers, root work, and remove repetitive admin. But somebody must still inspect what the system says, correct weak data, and decide when a human should take over.

[00:28:55] [SPEAKER_01] And that question reaches beyond any single platform or support team. So massive thank you to Phil for sharing his experiences and lessons that he's applying to ITSM there. But is your service desk measuring completed work still? Or the quality of the experience it creates? Big talking point, this one. So techtalksnetwork.com. You can find out more about me. 4,000 interviews over there.

[00:29:24] [SPEAKER_01] You can meet me on the road. So many, or send me an audio message. So many different ways of contacting me. But that is it for today. So thanks for listening as always. Bye for now.