Latest / The DEX Show: A Show for IT Change Makers / The Spark Avengers Unite: Dispatches on the FUTURE of IT (w/ Matt, Moe & Denis)

Transcript

The Spark Avengers Unite: Dispatches on the FUTURE of IT (w/ Matt, Moe & Denis) · The DEX Show: A Show for IT Change Makers

0:00
Speaker 1: Okay, everybody, everyone ready listen. This is I'm going to
0:04
Speaker 1: count us in a second as a guitar tuned. Is
0:07
Speaker 1: everybody paying attention? All right? In three and two and one?
0:12
Speaker 2: And hey it change makers and welcome to the deck
0:18
Speaker 2: Show with tom Ac gra Let's get into it.
0:24
Speaker 1: Well, what do we think? I like it? I think
0:28
Speaker 1: it's I think it's I think it's cheerful. We actually
0:31
Speaker 1: asked one of our producers, Catherine, what she thought of it,
0:34
Speaker 1: and she said she liked it, but it sounds a
0:37
Speaker 1: bit vacation y. I kind of think that's a good thing,
0:39
Speaker 1: you know. I think we all like vacations. Think of
0:41
Speaker 1: we like to think of a deck show as a
0:43
Speaker 1: kind of as a dex vacation. And if that's the case,
0:47
Speaker 1: what a what a wonderful vacation, What a wonderful destination
0:51
Speaker 1: we have planned for everybody to day. Because listen, for
0:55
Speaker 1: the last month or so, all I ever hear about
0:58
Speaker 1: if I'm talking to next thinkers, if I'm talking and
1:00
Speaker 1: customers or the listeners of the show, all I'm asked
1:03
Speaker 1: about is spark next Thing's personal it agent for every employee.
1:08
Speaker 1: And I've never seen this level of buzz about anything
1:12
Speaker 1: Next Thinks released. I don't think I've seen this level
1:14
Speaker 1: of buzz for any product that's come out of the
1:16
Speaker 1: whole Deck space. And you know, some people will tell
1:19
Speaker 1: you when we talk about Spark, it's not so much
1:22
Speaker 1: just the future of Dex, it's more like the future
1:26
Speaker 1: of it. And so we wanted to do an extra
1:28
Speaker 1: extra special episode and I've brought together We've brought together
1:32
Speaker 1: something of I'm calling the Spark Avengers. All right, it's
1:35
Speaker 1: got we got Matt, Dennie and Mo. So let's let's
1:39
Speaker 1: welcome them. And if you guys all introduce yourself. First
1:42
Speaker 1: of all, I'm starting with you, Dinny.
1:45
Speaker 3: Yeah, thank you, thank you for having us.
1:48
Speaker 1: So.
1:49
Speaker 4: I'm part of the product organization as a product senior
1:53
Speaker 4: product director, being with the company for.
1:55
Speaker 3: More than seven yours.
1:57
Speaker 4: And before that there was more on the its always
2:00
Speaker 4: more the customer sides. I've been walking business quite some time.
2:04
Speaker 1: Fantastic. I think I met you my first weekend Next
2:07
Speaker 1: Thing and it's lovely to finally have you on the
2:10
Speaker 1: show man. Welcome and another first time guest. Mo introduce
2:14
Speaker 1: yourself to the listeners, please.
2:16
Speaker 5: Mate, Thank you, Tom. I'm happy to be here. I'm
2:18
Speaker 5: an all time in here at next Think. I've been
2:20
Speaker 5: around for ten years and the last few years I've
2:23
Speaker 5: been focusing on developing AI native products. I'm from the
2:27
Speaker 5: engineering team. I'm leading the engineering department that is building Spark.
2:31
Speaker 5: Prior to this, I've worked on a lot of different
2:33
Speaker 5: products from floor remote, actions, infinity and and everything around that.
2:39
Speaker 1: I mean, don't you don't look like a ten year
2:41
Speaker 1: Next Thing veteran.
2:43
Speaker 5: Take good care of my skin care.
2:46
Speaker 1: You must have come straight out of school man, straight
2:49
Speaker 1: in there. Excellent returning guests. Now, welcome back Matt, Matt Rose, Hey.
2:54
Speaker 6: Doing, Thanks you good Tom.
2:56
Speaker 3: How are you very well? Very well.
2:57
Speaker 1: We're not going to introduce you you've been on before.
3:00
Speaker 1: Listeners know you a little bit, or you're going to
3:01
Speaker 1: say a couple of works.
3:02
Speaker 6: I could say hello for those who you know, maybe
3:05
Speaker 6: first time listeners or forget who I am, Matt Rose.
3:10
Speaker 6: I'm on the product marketing side over at Next Thing.
3:13
Speaker 6: Been working on a myriad of different solutions over the
3:16
Speaker 6: course of my only three years with the company. But
3:20
Speaker 6: it's been you know, great, great endeavors over that period
3:24
Speaker 6: of time, and like you're saying, Tom, the buzz around
3:27
Speaker 6: Spark has been nothing short of extraordinary. So really looking
3:31
Speaker 6: forward to talking about it today.
3:33
Speaker 1: Yeah, one hundred percent. It's been crazy. What there they are?
3:37
Speaker 1: Everybody are are Spark Avengers. You see, they've got the credentials,
3:41
Speaker 1: as will be more than clear. But let's begin with you, Matt.
3:47
Speaker 1: Maybe for anybody, despite all the hype and all the buzz,
3:52
Speaker 1: who hasn't already heard about it, who doesn't know what
3:54
Speaker 1: we're talking about, maybe give us a fresh elevator pitch
3:59
Speaker 1: so that we just level set and everybody knows what
4:02
Speaker 1: it is where we're all so excited.
4:05
Speaker 6: Yeah, So for I guess twenty plus years now, right,
4:08
Speaker 6: next Think has had an digital employee experience platform that's
4:12
Speaker 6: been helping companies solve problems at scale, and that's been incredible. Right,
4:17
Speaker 6: We've seen such a momentous leap in terms of overarching
4:22
Speaker 6: experience that people have on an everyday basis with the
4:24
Speaker 6: technology that they use. But when you think about many
4:28
Speaker 6: of the different kind of individual issues and challenges that
4:32
Speaker 6: are still arising, they're still arising. Right, my computer is slow,
4:36
Speaker 6: how do I access my company printer? These kind of
4:38
Speaker 6: tailored situations that may not be solvable at scale, But
4:42
Speaker 6: what we know is that with the technology that next
4:45
Speaker 6: Think has today, pairing digital employee experience data with this
4:50
Speaker 6: artificial intelligence and the actual ability to full scale remediate
4:54
Speaker 6: it means that we can put a personal IT agent
4:57
Speaker 6: through next Think into the hands of every single employ
5:00
Speaker 6: at each one of our customers, and now, all of
5:01
Speaker 6: a sudden, those challenging experiences which take twenty thirty plus
5:07
Speaker 6: minutes to solve every single day when you reach out
5:09
Speaker 6: to the service desk now can be solved pretty much
5:12
Speaker 6: instantly by reaching out to next Thing Spark through your
5:17
Speaker 6: current kind of chat channels that you utilize today. So
5:21
Speaker 6: it's really pretty revolutionary. Experiences that we've seen our customers
5:25
Speaker 6: have already has been nothing short remarkable, and I'm looking
5:29
Speaker 6: forward to kind of talking more about them here today
5:32
Speaker 6: and obviously hearing from the team in terms of, you know,
5:35
Speaker 6: how we built this, while we built this and so
5:37
Speaker 6: on and so forth.
5:39
Speaker 1: Absolutely put a pinion in that we're going to come
5:40
Speaker 1: back to you in a bit. But we're excited to
5:42
Speaker 1: have Denny and Mode because you guys obviously instrumental in
5:45
Speaker 1: the development of it, and as we're going to hear
5:47
Speaker 1: from like the very earliest inception. But from your perspective,
5:51
Speaker 1: first of all, Denny, like, where do you see the
5:54
Speaker 1: genesis of spark as an idea originating?
5:58
Speaker 3: Yeah, and we do.
6:00
Speaker 4: Didn't start by a certainly wake up and having the
6:02
Speaker 4: ID but it's much more about observing and seeing will
6:07
Speaker 4: the employee or they are acting and or they are
6:09
Speaker 4: using the self service solution today end by having discussion conversation,
6:14
Speaker 4: but also looking at the employee we really rapidly so
6:18
Speaker 4: that they didn't get the right experience.
6:21
Speaker 3: Most of the time, those seb service solutions.
6:23
Speaker 4: Have been more put in place to make it more
6:27
Speaker 4: complex for the employee to contact supports. Let's agree on
6:31
Speaker 4: that I was part of it and I did it
6:34
Speaker 4: that in the past that will to help them and
6:39
Speaker 4: when it provided help. Most of the times is just
6:43
Speaker 4: an instruction just to tell to the employee to do
6:46
Speaker 4: their problem, but to try to solve the problem by themselves, which.
6:49
Speaker 3: Is not right.
6:51
Speaker 4: This is really what we would have to change and
6:54
Speaker 4: be able to kind of solve the things because what
6:58
Speaker 4: usually happened right now, are frustrated, they don't want to
7:02
Speaker 4: quote contact support and whether they are having issues, and
7:07
Speaker 4: they have to accept it. So it's what we were
7:10
Speaker 4: always saying that we are struggling in silence. Nobody knows
7:14
Speaker 4: that they have that, and they are impact their productivity.
7:17
Speaker 1: That's a really good point. Yeah, there's sort of the
7:19
Speaker 1: the constant wrestle inside decks, specifically around the notion of
7:24
Speaker 1: self service. Right, And in a way, what AI has
7:27
Speaker 1: done is completely reconfigured our whole approach to interacting with computers,
7:32
Speaker 1: like in a proactive way. It's become easeful, pleasurable. It's
7:37
Speaker 1: become something that gives us a sense of like pleasurable
7:40
Speaker 1: control over things to actually give the prompts ourselves to
7:43
Speaker 1: be involved in that process. Previous to that, it was
7:46
Speaker 1: always a bit too much like hard work. But Mo,
7:49
Speaker 1: you were kind of that. I think you were there
7:50
Speaker 1: the when the idea was specifically formed, right, and when
7:55
Speaker 1: the vision of it was kind of struck. Tell us
7:58
Speaker 1: a bit about that, and was there a moment when
8:00
Speaker 1: you suddenly realized, Wow, this is this vision is going
8:03
Speaker 1: to be extremely impactful in reality.
8:07
Speaker 5: I mean definitely. I mean, going back into memory lane
8:10
Speaker 5: to twenty twenty four, I mean, when lllm's became a
8:13
Speaker 5: thing and we kind of had access to this type
8:16
Speaker 5: of technology, we knew that we could definitely leverage this
8:19
Speaker 5: to kind of fulfill the mission of next thing, which
8:21
Speaker 5: is really delighting the employees and making really removing all
8:25
Speaker 5: of the obstacles that they have to navigate during their
8:28
Speaker 5: work day and just focus on the most important things.
8:31
Speaker 5: In twenty twenty four, we decided in R and D,
8:34
Speaker 5: with Pedro's direction, to kind of have a focused team
8:37
Speaker 5: looking at this LM technology and kind of discovering and
8:41
Speaker 5: validating our vision about all of this. So we started
8:44
Speaker 5: very quickly, we assembled the team, I let the team
8:47
Speaker 5: started to kind of explore what can we can be done.
8:51
Speaker 5: And the first version of what we've done is basically
8:53
Speaker 5: took this amazing technology that's known as L and m's
8:56
Speaker 5: and we wanted to see if lms are able to
8:58
Speaker 5: kind of provide trouble shooting insights to our two employees.
9:04
Speaker 5: So we created a chat interface and we attached it
9:07
Speaker 5: to knowledge based articles, and we first our first kind
9:10
Speaker 5: of good indication that this will work is we saw
9:12
Speaker 5: that these lms are able to kind of understand this
9:16
Speaker 5: extra context and reason.
9:17
Speaker 3: About them and good good answers.
9:20
Speaker 5: This is back two years ago when people were still
9:22
Speaker 5: debating lms, where their intelligence do they reason or not?
9:26
Speaker 5: So we were kind of experimenting with all of this.
9:29
Speaker 5: So we connected it with the with the with the
9:31
Speaker 5: nexting data and we did an experiment.
9:33
Speaker 3: We broke the laptop.
9:34
Speaker 5: I broke my laptop and I blocked the port, and
9:37
Speaker 5: then we connected it with the next thing data and
9:39
Speaker 5: then I asked, I asked the agent, like I can't
9:42
Speaker 5: access this website. How can I resolve it? And it
9:45
Speaker 5: actually told me and it was super impressive. We gave
9:47
Speaker 5: it thousands of data points and it was able to
9:50
Speaker 5: painpoint exactly where the root cause is that the reporters blocked.
9:53
Speaker 3: Please fix it.
9:54
Speaker 5: This was the first moment that I personally felt that
9:57
Speaker 5: we have something there. Sent an email to a couple
9:59
Speaker 5: of people in the organization like, look, it's working, let's
10:02
Speaker 5: double down on this. And the natural thing we did
10:05
Speaker 5: next is why do we want to stop about stop
10:08
Speaker 5: at troubleshooting. Why couldn't we autonomously resolve it? So we
10:13
Speaker 5: connected it with our act module with remote actions. We
10:16
Speaker 5: had too with the remote action that could resolve this problem.
10:20
Speaker 5: So we connected it and we read on the experiment
10:23
Speaker 5: and it was the first time but next thing that
10:25
Speaker 5: we've seen an autonomous resolution of an IT problem. This
10:29
Speaker 5: was a good indication that we have a product that
10:33
Speaker 5: fulfill our mission, which is to create body light employees
10:36
Speaker 5: and create autonomous agents, which fulfills the mission that Pederal
10:40
Speaker 5: from the start kind of gave it direction about like
10:42
Speaker 5: we really want to create these agents that help employees,
10:45
Speaker 5: and I think the rest is history. We decided to
10:49
Speaker 5: double down on this. We took this team, we created
10:52
Speaker 5: the whole department they need, joined the project, and later
10:56
Speaker 5: on joined the product. We've evolved a lot since then.
11:01
Speaker 1: That's a great account, Anthony, from your perspective. Then developing
11:05
Speaker 1: that timeline, what was it like working with AI? I
11:08
Speaker 1: guess you know, like it's going to be for you
11:10
Speaker 1: coming into the project kind of it's a novel it's
11:13
Speaker 1: a novel factor, right, and the developer product and obviously
11:16
Speaker 1: integral to it being so effective.
11:19
Speaker 3: Yeah, No, for me, it was also part of the learning.
11:21
Speaker 4: As more mentioned, I've started a bit more than twelve
11:25
Speaker 4: months ago, and what I've started to learn very rapidly,
11:28
Speaker 4: that's building an AI product is very different in terms
11:32
Speaker 4: of product management because normally when product management, you think
11:37
Speaker 4: about something that you have to do. Then after you
11:39
Speaker 4: work with a team, then after his sheep, and then
11:42
Speaker 4: after you are pushing in terms of adoption, making sure
11:46
Speaker 4: that is good and potentially making better. What is really
11:50
Speaker 4: different in AI is that sometimes you don't know so
11:55
Speaker 4: that you have for everything that we are doing, we
11:58
Speaker 4: have to continuously evaluate the performance, making sure that you're
12:02
Speaker 4: getting better. It went through that sometimes we have added
12:07
Speaker 4: new capabilities and certainly it gets less good, and then
12:11
Speaker 4: after we have to adapt it. So really this logic
12:14
Speaker 4: of having this constant measuring may of the performance of Spark,
12:20
Speaker 4: its ability to resolve the first contact resolution that are
12:24
Speaker 4: very very important, but also the empree experience.
12:27
Speaker 3: This is really what we our doing is getting.
12:29
Speaker 1: Us very cool. No, I mean where do you think
12:33
Speaker 1: what do you see as the significance of Spark. I
12:37
Speaker 1: mean we alluded at the top of a conversation to
12:39
Speaker 1: it maybe being the future of it. Do you think
12:41
Speaker 1: that's a bit much? Where does its potential set in
12:43
Speaker 1: your view?
12:44
Speaker 5: For me, Spark is basically as the technology of the future,
12:47
Speaker 5: not only of it, it's just implementation in it. I mean,
12:51
Speaker 5: if I was to describe what a Spark for me,
12:54
Speaker 5: I would say it's basically achieving one of our childhood
12:58
Speaker 5: dreams of having this personal assistance that are very smart
13:01
Speaker 5: and intelligent that will do work for us and then
13:03
Speaker 5: we can focus on enjoying life or work or whatever
13:06
Speaker 5: youone enjoy. So for me, what's essential about Spark is
13:09
Speaker 5: its intelligence is a reasoning capability. It's basically the power
13:13
Speaker 5: of LLM, but it's actually a system of technologies working
13:16
Speaker 5: together to autonomously help an employee. It's not only that
13:20
Speaker 5: LM itself, but it's the intelligence of the data that
13:23
Speaker 5: the next thing. Infinity Platform provides the toolings that we're
13:26
Speaker 5: giving Spark access to so that it can remediate it.
13:29
Speaker 5: The memory and the personality that we're equipping Spark is
13:33
Speaker 5: a fundamental piece of this system that will allow Spark
13:36
Speaker 5: to continuously grow and become better. I see it today
13:40
Speaker 5: as you higher than a junior, and in turn the
13:44
Speaker 5: very quick future, you're going to have an expert that
13:48
Speaker 5: is the best IT agent out there that is able
13:51
Speaker 5: to can remember everything about you, your enterprise, etc. And
13:54
Speaker 5: use all of your problems and eventually Spark will be
13:57
Speaker 5: silently making sure that you're work environment is healthy, is
14:01
Speaker 5: you don't have issues and focusing on your most important
14:05
Speaker 5: tasks and.
14:06
Speaker 1: Can you give us any more, like just really impressive
14:09
Speaker 1: examples of what it's already making possible.
14:12
Speaker 5: Definitely, I mean Matt mentioned it in the beginning. Sometimes
14:15
Speaker 5: these small problems that are very annoying to employees, Spark
14:18
Speaker 5: is very good at really solving them. Even we developed
14:22
Speaker 5: Spark and we released it nearly a year ago. Internally,
14:25
Speaker 5: Next Think and every Next thinker is using it, and
14:27
Speaker 5: we've had early adopters with our customers. So we hear
14:30
Speaker 5: a lot of the use cases that Spark is very
14:32
Speaker 5: good at. And I'm very always happy and I feel
14:35
Speaker 5: the spark in myself when somebody tells me, Like I
14:38
Speaker 5: had the problem in my printer, I just told Spark
14:40
Speaker 5: and Spark resolved it. Or I was in a video
14:44
Speaker 5: call and my camera pros and I couldn't solve it,
14:46
Speaker 5: and I talked to Spark and Spark was magically able
14:49
Speaker 5: to detect that it's a problem in the driver and
14:52
Speaker 5: resolve it for me and I got access to my
14:54
Speaker 5: video again. So these small things that are very annoying
14:58
Speaker 5: to employees, Spark is able to of them. We've seen
15:01
Speaker 5: very complex troubleshooting that Spark is able to do. If
15:04
Speaker 5: you complain about an application, but the underlying problem is
15:08
Speaker 5: something completely else. It would have taken it support days
15:13
Speaker 5: or weeks to kind of really pinpointed where spark and
15:16
Speaker 5: within seconds can understand the correlation between the event that
15:20
Speaker 5: your application is not working and a driver or your
15:24
Speaker 5: rum or maybe you have your laptop overheated and it's
15:28
Speaker 5: not functioning well, and we're able to kind of advise
15:30
Speaker 5: the user and what to do to make it better,
15:32
Speaker 5: or even do it on his behalf.
15:35
Speaker 1: I do love for printer use cases so much hate
15:37
Speaker 1: I hate printers so passionately.
15:40
Speaker 5: It's the technology that annoys everybody, right, It's an essential
15:43
Speaker 5: technology annoys everybody, you know.
15:47
Speaker 1: Like you describe it more brilliantly earlier, it was almost
15:50
Speaker 1: like you just unleash the power of the LLM on
15:53
Speaker 1: all of next thing's data and just say, go mad,
15:56
Speaker 1: you know, see what you can do, and it just
15:58
Speaker 1: brings back all of these incredible insight some functionalities. But
16:01
Speaker 1: whether also like bumps in the road with any specific
16:04
Speaker 1: challenges in tuning it and fine tuning it properly.
16:09
Speaker 5: Yes, definitely. And I would say, what's challenging about Spark,
16:12
Speaker 5: and I can kind of divide my answer into two,
16:14
Speaker 5: is actually making it enterprise ready. We're developing an agent
16:17
Speaker 5: for enterprises. So we're very serious here at next thing
16:20
Speaker 5: about responsible AI and making sure we're developing an agent
16:23
Speaker 5: that's very secure and reliable. So it is a challenge,
16:27
Speaker 5: but an important challenge for us to solve, and we're
16:30
Speaker 5: solving it. To make sure that these agents are reliable,
16:33
Speaker 5: they're secure, we have the proper access management that everything
16:37
Speaker 5: that they do are safe for the employees and their devices,
16:41
Speaker 5: and so this was a very fundamental part of what
16:43
Speaker 5: we've been focusing on in the last year to make
16:45
Speaker 5: sure that people trust Spark and use it, because this
16:49
Speaker 5: is the hardest part of adoption of a product is
16:51
Speaker 5: really trust, and we're putting in everything we should be
16:53
Speaker 5: putting in so that employees trusted. The second part, and
16:58
Speaker 5: I was actually discussing this with the Knee over the week,
17:01
Speaker 5: is actually how can we make Spark right? How can
17:05
Speaker 5: we tailor Spark to everybody right? The tonality of Spark,
17:10
Speaker 5: the level of detail that we need to give to
17:12
Speaker 5: the employees, because employees, they don't really care too much
17:16
Speaker 5: about the details. Not everybody cares about the details, but
17:19
Speaker 5: maybe some people care about the details. So really, me
17:22
Speaker 5: and deny. We're working a lot with the products and
17:24
Speaker 5: the GM teams to find tune that understanding what is
17:27
Speaker 5: the good information to provide to the user so that
17:29
Speaker 5: we don't distract them too much, but we give them
17:31
Speaker 5: trust that Spark is on their problem and solving. That's
17:34
Speaker 5: something that we're continuously evolving and that's something that we
17:37
Speaker 5: will continue to improve in the next few weeks and
17:39
Speaker 5: months so that we can have a park that's good
17:43
Speaker 5: enough for everybody. And later on we're going to start
17:45
Speaker 5: to kind of specialize Spark or specific profiles so that
17:49
Speaker 5: it can learn from what you like to know and
17:51
Speaker 5: it can give you the information if you want it.
17:53
Speaker 3: If you don't want it, it won't give you this information.
17:55
Speaker 5: So that's that's that's a challenge that we'll continuously investing.
18:01
Speaker 4: If I can, just just to add on what Mode said,
18:04
Speaker 4: so because you mentioned about the personality, and this is
18:08
Speaker 4: something that I was not expecting to get as well feedback.
18:12
Speaker 4: We the positive feedback that we're getting from our customers
18:15
Speaker 4: that are testing the product, that are using it is
18:17
Speaker 4: really about the personality. They love it, they love what
18:21
Speaker 4: it is or its act because certainly it creates these
18:24
Speaker 4: human interactions because they have a bit of.
18:27
Speaker 3: Sense of humor. Is they able to kind of get it?
18:29
Speaker 3: And this is super.
18:32
Speaker 1: Super important. We're all we're all building these relationships, aren't
18:36
Speaker 1: we with our llms. It's a fact you know, we've
18:40
Speaker 1: had We've had conversations on the show before about Spark
18:42
Speaker 1: when it was launched, and some of this material is
18:45
Speaker 1: going to be familiar to people. But we are coming
18:48
Speaker 1: back to the public now with first hand accounts with
18:53
Speaker 1: uptake among customers. What's that been like and are there
18:57
Speaker 1: any metrics, proof points, value indicators we can concretely share
19:01
Speaker 1: at this stage.
19:02
Speaker 6: Yeah, So we've been Denimo and I have been talking
19:05
Speaker 6: to customers about Spark for months upon months now, meeting
19:08
Speaker 6: with them and getting a feel for, you know, what
19:11
Speaker 6: would Spark look like in their environment. And we've been
19:13
Speaker 6: working with probably twenty plus over the course of the
19:16
Speaker 6: last six months now to get it deployed, get it
19:19
Speaker 6: into the hands of you know, test groups, a lot
19:21
Speaker 6: of their employees and say okay, have at it right,
19:24
Speaker 6: see what it can do for you. Response has been
19:26
Speaker 6: overwhelmingly positive. The first thing that I think we always
19:29
Speaker 6: find is people are surprised, right, People are surprised by
19:33
Speaker 6: the fact that it can answer many of these different questions.
19:36
Speaker 6: That isn't pulling information and background data from you know,
19:41
Speaker 6: knowledge bases that they've previously created in the past. Hey,
19:44
Speaker 6: how did Spark know that it needed to do this, that,
19:46
Speaker 6: and the other thing in order to solve that particular problem. Well,
19:49
Speaker 6: not only has next Thing trained it, but also we've
19:52
Speaker 6: worked with foundational knowledge in order to make sure that
19:54
Speaker 6: it is as smart as possible. So that's been you know,
19:57
Speaker 6: incredible to see. We've had a couple of customers who've
20:00
Speaker 6: deployed it to you know, a thousand plus people and
20:02
Speaker 6: then done kind of testing analysis, right said, okay, see
20:06
Speaker 6: what it's like, solve your problem, and then send out
20:09
Speaker 6: a sentiment survey after to see how is your experience
20:11
Speaker 6: with Spark? Did it solve the problem that you are
20:13
Speaker 6: looking at? From a couple of those different customers, they
20:16
Speaker 6: actually saw seventy percent plus first contact resolution rate and
20:21
Speaker 6: equally as high satisfaction in the actual engagement that took
20:25
Speaker 6: place and their overarching experience, which is huge because you know,
20:28
Speaker 6: we want people coming back and utilizing Spark over and
20:30
Speaker 6: over and over again, because that's going to be their
20:33
Speaker 6: new as we talked about the future of IT support
20:36
Speaker 6: and how they actually contact effectively the service desks in
20:38
Speaker 6: order to be able to get their problems handled.
20:40
Speaker 1: And obviously bunch of AI agents are sort of flooding
20:45
Speaker 1: the market a little bit of the moment. What distinguishes
20:48
Speaker 1: Spark from one's others in the space, or more broadly.
20:52
Speaker 6: Yeah, I think it's perfect. I guess trifecta of things, right,
20:55
Speaker 6: one of which is the digital employee experience data that
21:00
Speaker 6: effectively every single Spark agent has because it's tailored to
21:03
Speaker 6: the individual employee themselves, and the fact that next link
21:05
Speaker 6: has been deployed across their device, right it's pulling this
21:09
Speaker 6: real time DEX data to be able to see what
21:11
Speaker 6: kind of hardware are they unising, what operating system are
21:14
Speaker 6: they on, what network are they connected to, so on
21:16
Speaker 6: and so forth. All the things that Moe and Denny
21:18
Speaker 6: have talked about thus far that have really kind of
21:21
Speaker 6: fed into point two, which is the intelligence factor and
21:24
Speaker 6: marrying up the foundational knowledge that we've already built, paired
21:27
Speaker 6: with a lot of their knowledge bases and information that
21:30
Speaker 6: really tailors sparked to their organization. And then on top
21:33
Speaker 6: of that the fact that Spark can go through because
21:36
Speaker 6: of next think the full scale remediation that many other
21:39
Speaker 6: solutions out there may just redirect you to another source
21:43
Speaker 6: to be able to have your problem fixed. We don't
21:45
Speaker 6: want that. We want it solved as fast as possible
21:47
Speaker 6: in the moment, and that's what we're able to do
21:50
Speaker 6: with Spark because of the remediation capabilities that it provides
21:53
Speaker 6: that end to end experience. So overall, I think that
21:56
Speaker 6: perfect kind of combination of three is really the differentiation
21:59
Speaker 6: between Spark and everything out there in the market.
22:01
Speaker 1: But Danny, you already mentioned the kind of chemistry between
22:05
Speaker 1: sparking users, right, any other ways that sparked I guess
22:09
Speaker 1: delivering value in the real world that have surprised you
22:12
Speaker 1: or delighted you.
22:14
Speaker 3: Yeah, yeah, not youre really surprise me. But what I'm really.
22:18
Speaker 4: Fascinated fascinated about Spark is how capable of solving problems molays,
22:25
Speaker 4: mentioning about this junor it support or intern and things
22:29
Speaker 4: like that, and we are seeing it growing. So in
22:32
Speaker 4: Spark we have this ability for Spark Supervisor to be
22:36
Speaker 4: able to review the conversation and see exactly what Spark
22:40
Speaker 4: did it so to get to the brain of spark,
22:43
Speaker 4: we say, the reasoning and really when you look at
22:47
Speaker 4: it as multiple terms, I say, wow, this is really impressive.
22:53
Speaker 4: We are able to kind of get the knowledge based article,
22:56
Speaker 4: understanding what it is, getting the data you see, the reasoning,
23:00
Speaker 4: and the ability to kind of course step by step,
23:03
Speaker 4: and it really start any support person performing the job.
23:08
Speaker 4: At the end we say, we know that he's an
23:10
Speaker 4: AI agent. So this is really impressive. I say, I
23:14
Speaker 4: like the analergy. Isn't a bit like when the first
23:17
Speaker 4: time for the one that have kids, when your first
23:20
Speaker 4: time your kids are doing bicycle by themselves, you have
23:24
Speaker 4: this kind of wow moments where you're always.
23:28
Speaker 3: That is cool.
23:29
Speaker 1: Completely agree. How is it fitting in Denny into you know,
23:35
Speaker 1: into in practical terms, inside existing workflow systems teams.
23:41
Speaker 4: Yeah, this is one of elements that was very important
23:44
Speaker 4: from the beginning for us. Is key that we are
23:49
Speaker 4: directly integrated into the IT processes because we know that
23:54
Speaker 4: if we would like to transform all employee employees are
24:00
Speaker 4: requesting or receiving support, you cannot really change and certainly
24:06
Speaker 4: add a new tool for them. To do it, you
24:08
Speaker 4: need to fit into the process that they have, and
24:12
Speaker 4: so is why we have fully designed and Spark from
24:15
Speaker 4: the beginning to make sure that they can integrate with
24:18
Speaker 4: all the different channel that exists, either the collaboration tools,
24:23
Speaker 4: existing chat, but that they have it could be also
24:26
Speaker 4: interactive for its response systems. But also by doing that,
24:30
Speaker 4: we also saw that sometimes they are not optimized for
24:34
Speaker 4: IT support. When you're speaking about the chat, the chat
24:38
Speaker 4: is not the best interface to have a contact with support.
24:42
Speaker 3: So we are also building our own.
24:43
Speaker 4: Interface that will be first that is designed to be
24:48
Speaker 4: the IT support assistance for every employee. Think from the
24:53
Speaker 4: beginning as that not as a conversation. And this is
24:56
Speaker 4: more of the things. But the other part that are
24:58
Speaker 4: also important is to really integrate into the existing support
25:02
Speaker 4: organization processes.
25:04
Speaker 3: They have a way to operate.
25:06
Speaker 4: And I mentioned that, yeah, you need to train Spark,
25:10
Speaker 4: you need to teach, you need to them what is
25:13
Speaker 4: good or not an here is where we are fitting
25:16
Speaker 4: really like in the service desk. Many service discordinator organizations,
25:19
Speaker 4: they have this service disc coordinator that in fact are
25:23
Speaker 4: reviewing about the performance of these service desk agents in fact,
25:28
Speaker 4: and here we fit perfectly. So we are providing a
25:30
Speaker 4: new agent to the service desk and then after they
25:35
Speaker 4: have to do the same type of a job, looking
25:38
Speaker 4: at the host park perform and.
25:40
Speaker 3: Then dig it.
25:41
Speaker 4: I like to say also that the best things we
25:43
Speaker 4: Spark is for the service discordinator or they just have
25:46
Speaker 4: to date once on what is right and he will
25:49
Speaker 4: remain super cool.
25:50
Speaker 1: And we kind of covered then a little bit more
25:54
Speaker 1: a reaction from customers. What about like partners, vendors, technology
25:59
Speaker 1: partner is how important is the surrounding ecosystem to the
26:03
Speaker 1: automate success of a realization, the full realization of sparks
26:07
Speaker 1: and potential and what partnerships and resulting functions are already blossoming.
26:13
Speaker 5: I think it's a foundational I thing for kind of
26:16
Speaker 5: to be part of this ecosystem. And I'd always say
26:19
Speaker 5: that partners and our vendors and partners, they're essential for
26:22
Speaker 5: our success because we're building Spark at scale for twenty
26:26
Speaker 5: twenty five to thirty million people, so it's a big challenge.
26:30
Speaker 5: So we've partnered with our vendors like AWS and with
26:34
Speaker 5: foundational model providers or labs like Coantropic, and every time
26:37
Speaker 5: we talk about Spark, they're very excited going to sponsor
26:40
Speaker 5: this type of work and help us and achieve our mission.
26:43
Speaker 5: And it's super critical because as you know, llm's are
26:46
Speaker 5: a rare resource and if you want to power twenty
26:49
Speaker 5: million people, twenty five million people, we need the support
26:52
Speaker 5: of AWS and Entropic to be able to provide the compute.
26:55
Speaker 5: And we have this strong partnership. We work on weekly
26:58
Speaker 5: basis with them to make sure that we have the
27:00
Speaker 5: proper inference for what we're doing in Spark, but also
27:03
Speaker 5: on the newest technologies, the guard Rails, the enterprise ready ecosystem,
27:08
Speaker 5: so it's very critical to what we're doing. Also thinking
27:12
Speaker 5: about non traditional partners for next thing that we've kind
27:14
Speaker 5: of also started working with last year about AI observability,
27:19
Speaker 5: like partners like Langsmith and there wait some biases that
27:23
Speaker 5: are helping us in both in observing AI and doing
27:26
Speaker 5: experiments and evaluations and benchmarks. So their tools and technologies
27:30
Speaker 5: are very essential for us to make sure that Spark
27:33
Speaker 5: is doing what it needs to be done and to
27:35
Speaker 5: continuously monitor it and experiment with a tend and wait
27:38
Speaker 5: some biases to can help us with our fine tuning
27:41
Speaker 5: and experimentation. So this I think partners allowed us to
27:44
Speaker 5: move quicker actual rate what we're doing and if I
27:47
Speaker 5: can kind of summarize it in one sentences. You can
27:50
Speaker 5: walk to a place or take a jet. Having partners
27:53
Speaker 5: and vendors supporting you is allowing you to take a
27:57
Speaker 5: jet to arrive to your ultimate destination fast.
28:01
Speaker 1: So many good analogies today, guys, analogy avengers, not just
28:05
Speaker 1: a Spark cool. Let's just look at a roadmap the
28:10
Speaker 1: technology development across the remainder of twenty twenty six. Maybe
28:13
Speaker 1: start with you, Denny, what are you excited about? What
28:16
Speaker 1: have we got to look forward to and seeing West
28:17
Speaker 1: Park goes next?
28:19
Speaker 4: Yeah, I think there are multiple different things that we're
28:22
Speaker 4: excited that we all have in twenty twenty six.
28:24
Speaker 3: I think what definitely is one of our.
28:26
Speaker 4: Key focus is to continuously increase the first core or
28:30
Speaker 4: first contact posolution, offering new tools, new capability for Spark
28:34
Speaker 4: to be able to troubleshoot and reserve issues. And the
28:37
Speaker 4: other aspect is also on the learning, which is kind
28:40
Speaker 4: of the same together, so all we can make Spark
28:44
Speaker 4: be able to learn, identify what there are some knowledge gaps,
28:48
Speaker 4: some actions gap. Provide this information to the Sparks provisor
28:53
Speaker 4: who cannot make Spark better every day.
28:56
Speaker 1: Amazing. Anything you'd add to that, man.
29:00
Speaker 5: Yeah, and here really extending the capabilities of Spark, I
29:03
Speaker 5: mean then covered the immediate things that we're working on
29:07
Speaker 5: for me, I mean, twenty twenty six twenty twenty seven
29:09
Speaker 5: is a great opportunity to kind of to evolve Spark,
29:12
Speaker 5: give it more capabilities and tools. We're looking into how
29:16
Speaker 5: Spark can see what you see and do what you
29:19
Speaker 5: can do. And we'll keep these all secrets for the
29:22
Speaker 5: future about what we're going to be providing for our
29:25
Speaker 5: customers that will make Spark much more efficient. And also
29:29
Speaker 5: the vision that we briefly talked about before is making
29:32
Speaker 5: Spark proactive, not only reactive. So this is an important
29:36
Speaker 5: part of our future, the future of Spark where you
29:39
Speaker 5: don't need to wait for an employee to reach out
29:42
Speaker 5: to Spark. What Spark can be proactive and if it
29:44
Speaker 5: solves the problem for one employee, it can solve it
29:46
Speaker 5: for all other people that are impacted, for example. So
29:50
Speaker 5: these are things that we're actively looking into, and then
29:52
Speaker 5: the next let's say eighteen months, we're going to start
29:55
Speaker 5: to see them more in production.
29:57
Speaker 1: Be a natural direction for next thing, a proactive anticipatory element.
30:04
Speaker 1: Matt leave us with some thoughts on the sort of
30:06
Speaker 1: big picture implications of this sort of crazy opening salvos
30:14
Speaker 1: of Spark and what it points to and means for
30:16
Speaker 1: the way IT itself is going to evolve in the
30:19
Speaker 1: coming years. What do you see it as all pointing
30:21
Speaker 1: in the direction of.
30:23
Speaker 6: Yeah, we talk a lot about some of the different
30:25
Speaker 6: kind of components of autonomous IT operations in the direction
30:28
Speaker 6: that that's going, and I think we're fully behind it. Right,
30:31
Speaker 6: there's elements of every single individual challenge that an employee
30:38
Speaker 6: faces in the very short term here should be able
30:41
Speaker 6: to be addressed by Spark. From an IT standpoint, we
30:45
Speaker 6: have this notion of okay, as of right now the
30:48
Speaker 6: employees raising that issue, but moving forward becomes proactive, like
30:52
Speaker 6: we've already talked about. So now all of a sudden,
30:55
Speaker 6: we've really doubled down on driving that digital employee experience up,
30:59
Speaker 6: which has been our mission since the start. Right, how
31:01
Speaker 6: do we ensure that we're enabling people with their full
31:04
Speaker 6: potential with technology? If we have everybody has a personal
31:09
Speaker 6: IT agent in their hands that's always providing them with
31:11
Speaker 6: the best possible experience. Think of how much additional work
31:14
Speaker 6: gets done, how much how little time gets lost as
31:18
Speaker 6: a result of technology, getting in the way and just
31:21
Speaker 6: you know where things go in terms of ensuring that
31:25
Speaker 6: everybody can be operating at peak speed. So I think
31:28
Speaker 6: that's the biggest thing, is unlocking more from your workforce
31:32
Speaker 6: through the utilization of a personal it agent, And that's
31:35
Speaker 6: kind of the direction that we're going and providing that
31:37
Speaker 6: overall frictionless experience. And it doesn't matter if it's a
31:40
Speaker 6: problem that you're facing, problem that you will face, or
31:43
Speaker 6: you're a new employee just jumping into an organization and
31:47
Speaker 6: learning how to get started. Those are all things that
31:49
Speaker 6: we want to cover and make sure that you know
31:52
Speaker 6: you're able to do so in the easiest possible way.
31:55
Speaker 1: Awesome, guys, thanks for coming on. Excellent conversation and excellent work.
32:00
Speaker 1: Excellent start to a year with this really inspiring. Best
32:04
Speaker 1: of luck with the remainder of the year and making
32:07
Speaker 1: this really really count that everybody in the industry. It's
32:10
Speaker 1: super exciting.
32:11
Speaker 3: Thanks guys, Thank you.
32:13
Speaker 5: Thank you.
32:14
Speaker 2: That's it for this week's Deck Show. We'll be back
32:17
Speaker 2: next week and if you don't already, be sure to
32:20
Speaker 2: follow the show so you don't miss an episode until
32:22
Speaker 2: next time.