Revolutionizing Customer Service with AI with Edwin Wong – Atlassian Team ’25 NA

This is Textron tv. We're back at Atlassian, team 25. Um, I'm here with Edwin Wong, who's a distinguished product manager, customer service management for Atlassian.

Welcome to the show. Thank You, John. Thanks for being here.

And, um, yeah, I, I don't know much about what you all do in this space. I'm wondering why Atlassian's entering the customer service management space to begin with. That's my opening question.

Yeah. Well, I'm super excited because today we're announcing our new CSM app, customer service management app that's really geared towards, uh, customer service, right? So one of the things that we've consistently heard from our Atlassian customers is that, um, they really wanted to break down these silos, these barriers between their development teams, their operations teams, and their customer support.

Now, for the last sort of two decades of our sort of existence, Atlassian, I think, has done a good job to bridge operations teams and dev teams together. But oftentimes we've heard that customer service teams are sort of a bit left out in a cold. Yeah.

We feel a bit disconnected, and we see the symptoms of that. So when we speak to our customers, many of them come to us to say, oh, I really wished that you had a way to bring customer service together. And so that's why I'm so excited to talk about this app today that we're launching, because that's what we, when They, when they talk about the silos, for instance, I, I'm sorry to interrupt.

No, that's okay. But is there a common scenario or a common pain point that they often experience that that is Yeah. Something that you consistently hear about?

Yeah, absolutely. So, um, you can look at it from both sides, right? So for a customer service team, for example, um, sometimes they just don't know what's actually happening, right?

A customer may be telling them, Hey, something doesn't work, or something's broken, it may be caused by a bug. And they're like, ah, well, we'll just, um, tell the development team. And then it sort of goes into a bit of abyss, right?

And they don't hear back much about it. They don't understand what the progress is. On the flip side, the development team is like, oh, there is this sort of random tickets that is coming up from customer support.

We want to build our own thing, but we are not really sure what customers are saying about this. And the benefit of bringing them all together is that you solve both of those problems, right? So for customer support teams, you now get really good visibility into what's happening in the dev team.

For example, the bug's actually being worked on right now, it's about to be deployed, it's actually being deployed, and I can go back to the customer and tell them it's been fixed, right? For the development team, on the other hand, they'll be understanding a lot more about, okay, this bug actually is really important because right now, at this very moment, there are 20 customers that are affected by this. So I should really get going and work on this.

And may this, maybe this one here, that one's a little bit less important and I shouldn't focus my time on it while we are working through the actual sort of task or development at hand. What it also allows you to do is to actually get more context and say, Hey, I actually can truly understand what's happening, what are the steps that's happening? Because all the, in all the sort of information coming from the customer is, is all linked back to the development task.

So it's a very powerful way to bring development teams and customer service together into one platform. You want me to go in, I interrupt you. I'm sorry for that.

You were mentioning some of what you announced today during the what, during the keynote address? I, I assume? Yes, That's right.

So yeah, so that's the c uh, customer service management app that we are announcing. Obviously very, very eager to bring this to the market as a way to break down those silos that I just mentioned earlier. But on top of that, it's also, I think, um, our latest foray into, um, really bringing together AI and human collaboration in a space that we think is a ripe for that opportunity, right?

So if we look at the adoption of AI right now, there are lots of different agents that are out there, but what we see is that, uh, one of the best opportunities to actually get high ROI on your AI investments is in this particular use case, because there is a level of repeatability. Some of these, there's a level of, I guess, predict predictability to some of this. And what we've been able to do is actually marry our investments in ai, in vo, right?

To provide a great, uh, customer experience, uh, AI driven customer experience for end customers, as well as using the AI to help your customer service teams for those, uh, requests or those questions that aren't solvable by a AI in its first instance directly for the customer for self-serve, right? So take the example of say a ticket that's come through and now, uh, requires human attention. AI can actually pull in a lot more of the context of what's happening and actually suggest to you, for example, maybe you should try these three steps with the customer.

Maybe you should talk to Bob over there on the development team, right? 'cause he seems to be an expert here, and that's the power that we bring with our CSM app. Yeah, that's interesting that you're saying this because there's this, there's just this cade like a stampede for AI agents, and, and there's no doubt it's gonna be adopted and used certain tasks, but then that kind of creates this friction or dynamic between how the, where the agent and the human interact and how that affects the customer.

And I think there's maybe a little bit of, uh, um, fear on the part of the customer of how their experience might be changed or altered. I think, I think ironically, from what I've heard so far, the experience is gonna be a lot better if you're a customer, especially with an AI agent assisting you in some ways, even down to the, to it talking to you voice to voice. It, It, it has John, it's interesting sort of you raise that fact because, um, as one of the things we do in product development, as we are developing this new customer service management app, we've been dogfooding ourselves heavily.

So we've been deploying it for our own support teams. And one of the things we were very worried about when we first started rolling it out is, Hey, are customers going to react negatively to this? Because it's gonna sort of give them a, a less sort of human sort of experience.

Do, Do you think they almost feel as if they're kind of like a secondary citizen and that they're, that they're, they've been relegated to second class status by talking to a, So that's what we were afraid of when we started this sort of dog fooding role, as we call it. But what we found after sort of launching it into our own, uh, support use case is the reverse actually happened. We actually saw an improvement in our CSAT 6% improvement actually, which is quite substantial, a big jump from not using AI to using AI because of the timeliness and because the accuracy of the responses that we're giving, it's actually made people more happy that they're getting a better response and quicker.

And that's actually turned out to be a win, both on the cost and the productivity and the efficiency side, as well as the customer satisfaction side. So it's actually a win-win on both sides. That's, Yeah, that is, that is interesting because in sense, you're, if you're on the receiving end and you're frustrated, like if you have a problem as a customer, you wanna know what the, what the issue is.

Sometimes there's a back and forth or they call you back later, or you get an email here. Is this, are we talking about a response within seconds, conceivably in a casual conversation where it's assessing all the, uh, information it has and and relaying that to you accurately? Is that essentially going on?

Yeah, So I mean, leveraging the ai, I mean, the response is almost instantaneous, obviously. Okay, there's gonna be some time for the AI to respond, so it's within seconds, right? And it could be drawing from a pretty big corpus of knowledge, your knowledge basis, um, other sources of information.

And one of the big benefits Atlassian bring is this notion of teamwork graph that we talked about. So take for example, um, the idea that you may have a feature request that we are capturing in our product, your product discovery right now, we could tell you this is a known feature request. Why don't you go watch it here and vote for it as a way to respond, right?

Customers like, okay, well great, now I can engage on that particular feature request as opposed to continuing to think about this as a bug and get frustrated about it. So it's actually quite powerful to be able to do that very, very quickly within seconds. To your point, John, Internally, when you, when you use this customer service management involving ai, how, what is the dynamic between humans and the AI elements?

Yeah. Do they co, do they work together or does, I'm just, can you maybe explain to me how that, how that yeah. Works?

So our philosophy is really to think about AI agents as a teammate for you, right? So I think traditionally some sort of, uh, uh, general understanding would be you're pitching human against ai. We actually don't believe in this whatsoever.

We actually believe it's about human AI collaboration. When you are working on this internally, what we believe is AI could actually be a great, um, orchestrator and a great collaborator to the work that you are doing, right? So rather than, for example, say, Hey, go chase down all the information that you need or understand everything about all the domains, AI can now be suggesting to you what you should be doing, suggesting to you who the knowledge experts should be orchestrating for you, what the response is.

Helping you automate those responses back to your customers, making sure you're on top of your workflow so that you're actually much more effective. So we are effectively empowering those humans with AI as opposed to saying, well, we're getting rid of a replacement. Yeah, yeah, exactly right.

That, that always usually comes down to the fear of being displaced. I, so this is a really interesting program. Is there a name for the program?

And what I want to know is, is, is this new app, what, when will it be available and how could you sign up for a closed beta program? Um, so this is the, the new app is called customer service management. Last thing's, trying Keep it simple here.

Yeah, simple enough, nice and descriptive in terms of, uh, when it'll be available. That's probably the hardest question you have asked me, John, because with developments one of those things that we, uh, don't really know at this point. So we are just, uh, working with our customers and we will know as we firm up our product.

C Yeah. Can I ask you what type of customers are, are testing it right now? I mean, what type of Yeah.

Uh, industries they're in, or can you name them? I don't know if you can. No, I don't have clearance names.

That's fine. And we have a couple that are right now, uh, very, very early, uh, that we've reached out to, and they're eager to actually, um, try our product into production. Um, but, uh, the type of customers, uh, that we typically focus on, uh, what we would be calling, um, um, technology driven, right?

So, um, it's not exactly the same as, uh, maybe software companies or technology companies that you're familiar with in the sense that their core business is technology, but increasingly we're seeing even bricks and mortar businesses, right? Dominoes, for example, was at the keynote this morning, and I would argue that Domino's is a technology driven company in the sense that so much of their offering today is based on technology. And that's the sort of customers that we are seeing that really benefit from this.

Because the very strong value proposition that we bring to the table is being able to connect the dots across that whole life cycle for technology driven companies. Alright, That sounds cool. I mean, this sounds fantastic.

It's gonna help people and it's gonna make them more open to working with AI and trusting it. So, um, I'm all for that. And, you know, it's something inevitable.

We have to, we're gonna have to work with it one way or another. And so you might as well take advantage of its possibilities and enjoy it. So thank you Edwin, for being here.

Thank you. Thank You, John. I appreciate your time.

Uh, that's it for this, this, uh, segment. We'll be back with some more from Atlassian, team 25, very shortly.