Unlocking AI Advancements in Jira Service Management – Atlassian Team ’25 NA

This is Techstrong tv. We are back in Anaheim, California, near Disneyland. This is Atlassian, team 25.

I'm John Swartz with Techstrong tv. I'm the senior content writer, and with me is Shemic Sharma, who's the head of Product IT Solutions at Atlassian. Welcome to the show.

It's good to see you. Thanks For having me. Sure.

It's a pleasure to have you. Uh, we're gonna start with something kinda just basic, but I wanna ask you about the most important AI advancements being introduced in Jira service management at this show. Yeah.

We're announcing a whole bunch of, uh, AI advancements, uh, but I'd like to highlight just four of them. Sure. First is that inside the Jira service management, um, there is something called the virtual service agent that helps agents solve the problems and the tickets that come in.

Um, now we are making the virtual service agent available across a lot more surfaces than it used to be. So previously it was available on the help center and in Slack. Now it's available in Microsoft Teams.

You can access it over email and you can also embed a widget, including the virtual service agent in any website. So that's one that more s second, your agents are only as useful as the knowledge that they can access. So we are expanding the knowledge sources that these agents can access to Google Drive and SharePoint as well.

So they get a lot more powerful. And also we're allowing the agents to suggest what knowledge you lack, so you can go and improve that. And it even drafts these knowledge articles for you so you can review and expand your knowledge base.

So that makes the agents much better. The third thing we are doing is adding agents into your portfolio so that these agents can actually do the work for your agents. So, for example, um, when service requests come in, the service triage agent helps you prioritize the requests so that you can work on the most important ones.

It fixes the subject headings so that it's more relevant, it fixes the direction and the priority of those requests and so on. So it just makes the work of your human agents a lot more easier. They don't have to work on the, um, the, the, the peripherals and they can actually get focused on the actual tickets content.

And the last is on the operations side. We are doing a whole bunch of things to help people identify the root cause of an incident so that they can get to the problem, um, very quickly. You know, the interesting, uh, theme I got get from this show is that with all the AI agents that are flooding the zone, so to speak.

Yeah. And I know Google made a bunch of announcements. There'll be more from ServiceNow, from Zscaler, from whomever, Adobe, Salesforce.

It, your approach as a company to me seems to be more about a bottom up a feeling of this is a pragmatic solution. We understand what you're facing as a customer. We wanna make it simple, simple for you.

We don't wanna be overwhelming. We, we want to, uh, uh, make this more productive and we want you to trust the agent. We don't want you to feel threatened by it.

Is that kind of a theme or an undercurrent of what, of what you're doing or Trying to do? Exactly. Um, we've always come from the place that, um, we want to connect dev and IT teams to all the business teams, right?

And to help connect them so that the work gets done better. And we see as agents as coming and helping in those connections between these teams. So the root underlying, you know, framework that all of these agents rely on is what we call the teamwork graph.

And the teamwork graph takes all the data from all of these tools and make, makes all of these agents much more powerful. And that teamwork graph has data from all the dev tools that you have, from all the knowledge tools that you have, and from all the IT tools that you have, as well as from third parties. And once it has all of this information, it just makes those agents work seamlessly inside your business workflow rather than requiring you to learn a whole new workflow that you may not be used to.

So it just kind of interjects itself in the right place at the right time so that you, you know, you're doing the work that you want to be doing, but it's a, it's a panel on the side that's suggesting things for you. I think that's always been the big fear, is like having to relearn. I mean, that's one reason why people are so afraid of technology.

Exactly. It's this idea like, oh my God, this thing called A VCR or a microwave or whatever, a smartphone. Now I've gotta learn, I've gotta start all over again.

And in a sense, what you're doing is you're making an easy transition based on what they already understand, so they don't have to start from the ground up. Exactly. So in my opinion, agents should come into the workflows that you're already familiar with and come and help you within there, right?

Mm-hmm. Rather than requiring you to learn a whole new business workflow. Okay.

I was gonna ask you the, how do these new features that you mentioned differentiate Atlassian from folks like ServiceNow in the ITSM space? And I'm asking you that because ServiceNow, it seems to me is making, uh, acquisition after acquisition after acquisition in IS is a competitor. I'm wondering how you differentiate.

Yeah. Our main differentiation is the underlying data that we rely on. And, um, what we do is that we take all your IT systems data, which, you know, other companies and competitors also do, but we connect it better to the dev information and all the product information that is there in Jira and Confluence.

So typically when a uh, issue comes up, that issue did not start just because something went down. There was usually some change that had happened. Mm-hmm.

And that change was because you were trying to pursue some goal and some team was trying to do something. So when you want to resolve that issue and fix it, you want to have that context of how did this issue come up in the first place? Right?

There must have been some change that we were doing or some business goal that we are trying to pursue that caused a configuration change that then led to this issue. So being able to identify the cause of a problem requires you to be able to look to the left of your code, left of the operational system that you have, and that's the information that's already encapsulated inside Jira, confluence and so on. So we tie those things together to give you a better answer.

Okay. I, um, maybe we can talk about VO agents. I'm wondering if you can, can elaborate on how those agents, uh, their role in incident management and how they contribute to faster resolution times?

Yeah. This is a area we've been investing a lot in. Um, so when a alert comes in or something goes wrong, usually it's not just one alert, there's usually a flood of alerts.

And we've always had, we've announced before that we allow you to alert group these things so that it actually, uh, separates the signal from the noise and you can see what the gist of the problem is and not have to dig through every alert. That's a duplicate of everything else. But now what we are announcing beyond that is an ops agent as well, that goes beyond all of that to help you identify what the root cause of the issue is.

So it looks across all your systems, all your prior incidents, as well as the information that you have in Jira and Confluence, and suggests that this is where we think the problem might be. And it also suggests resources that might help you resolve that issue. Playbooks, for example, that you should execute.

And further it goes, um, a step further by allowing you to take the action and say that, yes, I want to run this playbook, and then it act automatically runs the playbook to fix the issue. And one of the other thing it does, which people love, is that after the issue is resolved, it helps create a post incident report for you so that you don't have to write it yourself about what just happened. Right.

So all of this comes with the operations agent that we are launching along with Jira service management. Yeah. It's just like the takeaway for me, always, again, consistently is that this agent is actually helping me.

It's not threatening my job. Exactly. It's kind of trying to ease my burden.

Exactly. And not in an intrusive way, just like, and, and kind of taking a little bit of the edge of the time off of these little, these meddlesome tasks or these important tasks that need to be done. But you know, you just, you have, you feel as if you have somebody who has your back.

Exactly. And it's not another person, it's somebody who is, uh, gonna probably be very accurate each time they, they do this, It's, it's the robin to your Batman. Right.

You Know? Yeah. There you go.

That's the killer quote of so far of the show. Yeah. Um, what are some of the expected impacts of these AI features on HR service management, and how do they improve employee experiences?

Yeah, so hrs uh, departments are unique in the sense that they have a lot going on. Um, solving employee questions are obviously something that's top of mind, but it's just one part of their job. They have so many other things going on.

So what they're really looking for is to get started quickly and to be able to resolve these, uh, employee questions very fast. The advantage they have is that typically they have very good knowledge base. They, they're pretty good about, uh, writing down their policies, their leave policies, their compensation policies, and so on.

So this is an ideal environment for AI agents to actually participate because they have the knowledge and the requests that come in are typically well structured and frequently the same that have been seen before, right? Mm-hmm. So, um, we, when we deploy AI agents into the hr, we find that hr, um, HR folks love them because this is exactly the stuff that they don't want to be doing.

They want to be doing higher end tasks of handling helping their employees, right? So we find like HR SM agents are, are, are getting great uptake in the, in the HR world, they primarily help with things like that are very structured in the process. Link employee onboarding, for example.

There's a sequence of about 20 things that you typically have to do. And once you tell the AI agent, go create the employee records, update this, uh, uh, form, that will then trigger, uh, a laptop procurement or so on, it can do all of that for you in a sequence. Okay.

Yeah. I was gonna ask you, that's my final question is what's next for a Jira service management? Um, we con we want to continue investing in bringing, um, dev it and business teams together, and we are gonna expand those connections.

And we want to have more and more agents that work across these boundaries to help your information that's stored in one place, to be used in other places so that they become, the agents become much more powerful. So things like employee onboarding, things like, um, solving an incident that might cut across many systems across your department, these are the kinds of problems we want to solve. You'll be hearing a lot more about these kinds of issues and these kinds of, uh, innovations that we are building in the next few quarters.

Oh, That's awesome. Hey, thanks, Jamica. It was great meeting you.

Uh, thanks for your insight. Uh, very interesting, very interesting show to me. It's like a very practical show about collaboration, teamwork, uh, productivity, and then just, uh, eliminating clutter.

Yeah. Eliminating unnecessary things that I find myself repeating repeatedly. Doing like trying to find data or trying or trying to get answers to something that I got the answer to months before.

Yeah. So, um, thanks again. Thank you.

Thanks. And so that's it. Um, we're gonna have more interviews today from, uh, Atlassian team 25 in Anaheim for now.

Uh, we'll see you in a few minutes.