Latest / The DEX Show: A Show for IT Change Makers / Reality Bytes Is BACK: ft. Marc Petter on the Future of IT Jobs

Transcript

Reality Bytes Is BACK: ft. Marc Petter on the Future of IT Jobs · The DEX Show: A Show for IT Change Makers

0:02
Speaker 1: Hey it change Makers, and welcome to the.
0:05
Speaker 2: Deck show with Tomac gra Let's get into it.
0:14
Speaker 3: Hello, change Makers, we're back.
0:16
Speaker 4: You didn't think Reality Bites was was going to disappear
0:20
Speaker 4: along with our departed co host Tim Flower, did you.
0:25
Speaker 4: It was never our intention to leave leave it so
0:28
Speaker 4: long to record again, but we we did want to
0:31
Speaker 4: take the opportunity to to recalibrate the Reality Bites format
0:35
Speaker 4: a little bit. If you recall, we used to call
0:39
Speaker 4: one Reality Bites for show within a show, which was
0:42
Speaker 4: when the whole team had had a bit of a
0:45
Speaker 4: conversation together, and we called the other Reality Bites for
0:48
Speaker 4: show within a show, which was when we welcomed sorry,
0:51
Speaker 4: the show within a show within a show.
0:54
Speaker 1: Which was when we welcomed next thing is to join
0:57
Speaker 1: the reality by its panelists, and we were we would
1:00
Speaker 1: have a conversation with them about their area or areas
1:04
Speaker 1: of expertise, and we're kind of combining.
1:07
Speaker 4: These two formats going forward. So in place of Tim,
1:11
Speaker 4: if you like, we're always going to welcome a guest
1:14
Speaker 4: once a month from inside the organization to talk to
1:17
Speaker 4: us about something fascinating, something intriguing, and we're going to
1:22
Speaker 4: do so with our regular reality Bies panelists, and today
1:27
Speaker 4: we welcome back after far too long, Dina and Oriana. Dina,
1:32
Speaker 4: first of all, how are you?
1:33
Speaker 3: How you been keeping all good?
1:35
Speaker 2: Oh lood? Been sharpening my AI skills?
1:38
Speaker 4: Have you have you what does that look like?
1:40
Speaker 3: Have you been doing it?
1:42
Speaker 2: Just exploring, practicing, practicing? I think yeah, I'll leave some
1:46
Speaker 2: for the show, but it's more more practice other than
1:49
Speaker 2: this all.
1:50
Speaker 4: Good shouting and Oriana, welcome back. How are you? How
1:53
Speaker 4: have you been?
1:54
Speaker 2: Hello?
1:55
Speaker 5: Happy to be back. And actually Mark and I have
1:58
Speaker 5: worked on some AI things together recently as well.
2:03
Speaker 4: Fascinating and foreshadowing our guest, who has known my pleasure
2:06
Speaker 4: to introduce. We last saw him on the Deck Show,
2:09
Speaker 4: side by side with his assist robot AI sidekick. He's
2:13
Speaker 4: Mark Petter, senior product manager here at Next Think. Mark,
2:17
Speaker 4: welcome back to this show on your own today. We
2:19
Speaker 4: assume no no, no AI sidekick with you this afternoon.
2:23
Speaker 6: Unfortunately, yes, the assist is busy today so he would.
2:27
Speaker 3: Not be able to truly do to join you know
2:30
Speaker 3: this podcast.
2:31
Speaker 4: Yeah, well, we were very happy to have an all
2:35
Speaker 4: human human panel, but not necessarily an all human topics.
2:40
Speaker 4: So I'm gonna I'm going to tear it up and
2:41
Speaker 4: we're going to come to you first of all in
2:43
Speaker 4: the second mark, but the floor is absolutely open Orianna
2:46
Speaker 4: and Dina for your insights. This is this is what
2:50
Speaker 4: sort of inspired the idea, and actually the idea came
2:53
Speaker 4: in part from our very own co panelist, Sean Malvey,
2:57
Speaker 4: who I should take a moment to congratulate for the
2:59
Speaker 4: birth of this second child. That's why he's not with
3:02
Speaker 4: us today. Congratulations to Sean and family. This was this idea, right.
3:07
Speaker 4: You often hear let's say, you've got the AI Optimists
3:10
Speaker 4: out there, and the AI Optimists are always saying things like, well, yes,
3:13
Speaker 4: of course some roles will be automated, some roles will
3:16
Speaker 4: disappear in the AI revolution. But what you always find
3:20
Speaker 4: with these technological revolutions is some jobs disappear and new
3:24
Speaker 4: ones emerge, and some new ones we won't be able
3:27
Speaker 4: to imagine, and some new ones we can sort of predict,
3:30
Speaker 4: but we can be confident, we can be sure that
3:32
Speaker 4: human work itself won't disappear. It will merely reconfigure itself,
3:36
Speaker 4: as we saw in the indust dual revolution, et cetera,
3:39
Speaker 4: et cetera. And it's a kind of common take, but
3:42
Speaker 4: what they rarely do is start to tell us, start
3:45
Speaker 4: to describe, whichever industry of the discussion is taking place inside,
3:51
Speaker 4: what those future jobs might look like. So we thought
3:55
Speaker 4: we would come together as a team and have a
3:58
Speaker 4: bit of a stab at articulating the future of vi
4:01
Speaker 4: T job, the future of vi T career. And so
4:06
Speaker 4: let's start with you, Mark. There's no accident that we
4:09
Speaker 4: invited you to discuss this. It's obviously an area you're
4:12
Speaker 4: working quite a bit. But let's let's begin with some
4:16
Speaker 4: cold hard facts first of all, and let's look at
4:18
Speaker 4: what's passing away before we look at what's to come,
4:22
Speaker 4: what i T tasks, and with it, let's be honest, roles.
4:25
Speaker 4: Do you think our most vulnerable Mark and most worth automating?
4:30
Speaker 3: You might say yes, that's actually a very good question.
4:34
Speaker 6: And what I want to maybe just start with the
4:36
Speaker 6: other disclaimer and you mentioned it in your production, is
4:41
Speaker 6: I think the eyelids KIP is changing so fast. The
4:44
Speaker 6: technology is evolving so fast that everything that I'm going
4:48
Speaker 6: to share today will be mostly based on on some
4:51
Speaker 6: trends that I see and some feedback that I get
4:53
Speaker 6: from our customers, and also how I use AI, right,
4:57
Speaker 6: But it's not necessarily going to be exactly like this
4:59
Speaker 6: in the future, because maybe I is going to shift,
5:02
Speaker 6: you know, in a year time, and maybe it's going
5:04
Speaker 6: to do something else. But anyway, with this discimer out
5:08
Speaker 6: of the way, I mean, what I see as what
5:11
Speaker 6: could be automated is first of all, something that is
5:14
Speaker 6: quite exciting because I think AI will be able to
5:18
Speaker 6: transform a lot of jobs even beyond it in terms
5:21
Speaker 6: of what people are doing in the day today in
5:23
Speaker 6: the sense that if you think about what you're doing
5:26
Speaker 6: today as part of your work, you're spending a lot
5:29
Speaker 6: of tasks and time on operational day to day things
5:33
Speaker 6: that you have to do but are maybe not so
5:35
Speaker 6: core to the jump descriptions that you have. For example,
5:39
Speaker 6: could be, you know, you need to update status about
5:42
Speaker 6: a task, you need to document something to answer a
5:44
Speaker 6: question about one of your projects, or a product that
5:48
Speaker 6: you own that is in the documentation.
5:50
Speaker 3: And so on.
5:50
Speaker 6: So all of that are you know when when it
5:54
Speaker 6: pipes up right and it takes time from your time,
5:57
Speaker 6: And I think this is the type of activities and
6:00
Speaker 6: asks that we will be able to automate, which is
6:03
Speaker 6: I think very exciting because if you can automate these
6:06
Speaker 6: kind of things, then you can free up time for
6:08
Speaker 6: more added value time and activities that you can do
6:11
Speaker 6: that are more core to your job descriptions per se.
6:15
Speaker 6: So I think this is really great, and you know,
6:17
Speaker 6: I even had examples where me personally, I'm starting dedicating
6:20
Speaker 6: a lot of things to clode to AI that sits
6:24
Speaker 6: on top of my device and really takes a lot
6:27
Speaker 6: of things for my to do list, my daily to
6:28
Speaker 6: do list, for me to be able to focus on
6:31
Speaker 6: other more strategic items. And I was also just going
6:35
Speaker 6: to mention in terms of of other types of automations
6:39
Speaker 6: that I see happening among our customers when I interact
6:42
Speaker 6: with them, and how I see them actually using AI
6:45
Speaker 6: to automate things. I think, I for all the time being,
6:48
Speaker 6: I see maybe three main main levels of automations.
6:54
Speaker 3: I see that people are doing.
6:56
Speaker 6: You know, they're trying to automate things that are more
7:00
Speaker 6: to productivity, so they create AI agents for them to
7:04
Speaker 6: find the right time slot to organize a meeting, or
7:07
Speaker 6: to take an expense note that you have and fill
7:10
Speaker 6: that into a expense form that you have in a system. Right,
7:13
Speaker 6: So all those smaller use cases where AI actually held
7:17
Speaker 6: them it's automated there. But then I see also two
7:21
Speaker 6: other levels that are a bit bigger. Let's say the
7:24
Speaker 6: first one is trying to delegate some parts of their processes.
7:27
Speaker 6: And here, for example, if we talk about netting workspace right,
7:32
Speaker 6: so what we're referring is we're offering a way for
7:34
Speaker 6: IT teams to be able to delegate the whole data
7:38
Speaker 6: analysis side of things to manage digital.
7:41
Speaker 3: Pre experience too AI.
7:43
Speaker 6: So suddenly instead of spending a lot of time and
7:46
Speaker 6: efforts and its kids and expertise on doing one activity,
7:50
Speaker 6: they can actually delegate that to AI. They are still
7:53
Speaker 6: in the loop, right, but they consume the output of that,
7:56
Speaker 6: they gain a lot of time and then they can
7:58
Speaker 6: make the right decisions and do their process is right.
8:01
Speaker 6: And the last level is the level where we start
8:05
Speaker 6: seeing also customers fully dedicating processes food processes actually also too.
8:10
Speaker 3: And here, for.
8:11
Speaker 6: Example, we have nets in spark right that we recently
8:15
Speaker 6: released that is a perfect example of that, where now
8:18
Speaker 6: we have customers that use netsing Spark and they use
8:21
Speaker 6: this advanced agent to really automate how to resolve incidents,
8:27
Speaker 6: and the AI is actually really taking care of that
8:30
Speaker 6: aspect manage incidents and in rare occasions, is credated to
8:35
Speaker 6: a human when needed.
8:36
Speaker 3: But really a huge part of the process is actually
8:39
Speaker 3: taking care by the AI.
8:41
Speaker 4: Yeah, that's a really good idea. Mark And just before
8:44
Speaker 4: we get into what comes next or what remains, is
8:49
Speaker 4: there a difference in your view between what can be
8:52
Speaker 4: automated and what should be.
8:56
Speaker 6: Yes. So that's a very interesting question, because that's a
8:59
Speaker 6: question everybody should ask themselves, right, I think I think
9:04
Speaker 6: there's different things that people need to ask themselves to
9:08
Speaker 6: select the right tasks to automate.
9:11
Speaker 3: Right, So first of all, is what's the gain if
9:17
Speaker 3: you automate it? Right? Is it going to really unlock anything?
9:20
Speaker 6: Is it going to free upsot some of your time
9:22
Speaker 6: and then you would be able to do something or not? Really?
9:25
Speaker 6: And also what's the expected quality that you have for
9:29
Speaker 6: that task? Is AI really going to meet your requirements there?
9:33
Speaker 3: Right?
9:34
Speaker 6: And you need some somehow to think about this and
9:39
Speaker 6: not jump into the AI brand and say, okay, let's
9:42
Speaker 6: automate everything that we're currently doing, because you may spend
9:45
Speaker 6: a lot of time trying to automate it configure your
9:47
Speaker 6: AI systems with very little written.
9:50
Speaker 3: On investments at the end, on top of that. So, yeah,
9:54
Speaker 3: that's my take.
9:55
Speaker 5: Care I would say on this, Yeah, I think I
9:59
Speaker 5: really agree with this idea of you know, there's a
10:01
Speaker 5: lot of things you can automate, and the craik here
10:04
Speaker 5: you should used to hear it if you should, are
10:06
Speaker 5: things like quality and what kinds of systems can you
10:10
Speaker 5: build around it. I think a lot of the questions
10:13
Speaker 5: that Mark is bringing up are really questions of systems
10:16
Speaker 5: and of discernment, and that's you know, my favorite soapbox
10:21
Speaker 5: is that AI works best if you apply human discernment
10:24
Speaker 5: to it, and that means sometimes choosing the right workflow,
10:28
Speaker 5: and sometimes it means also saying, actually, here we want
10:32
Speaker 5: to have you know, the the ungloved. Shall we say,
10:36
Speaker 5: apologies for the weird metaphor, but rather than white glove
10:40
Speaker 5: or you know, robot hand experience, the human hand experience
10:46
Speaker 5: that we want to provide for whatever reason.
10:50
Speaker 2: Totally, yes, I'm totally in alignment with more thoughts and
10:55
Speaker 2: oreanas as well. Just to sum it up with the
10:58
Speaker 2: should I automated or could it be automated? I try
11:01
Speaker 2: and some simplify it, put it in one framework. Whenever
11:05
Speaker 2: there's a task in front of me, I think on
11:08
Speaker 2: three pillars, efficiency, quality, and compliance. If I'm able to
11:13
Speaker 2: answer with the SSS to all of them, then I
11:15
Speaker 2: go to how and how much and then try and
11:17
Speaker 2: balance it out between these three pillars. If that is
11:20
Speaker 2: the case, then let's go ahead. We start to automate
11:23
Speaker 2: what we need to and then we move on to
11:25
Speaker 2: the next one. And the other part with Oreana really
11:29
Speaker 2: highlighted is the human fingerprint. You can do these things,
11:32
Speaker 2: the efficiency, the quality, and compliance, but you always need
11:35
Speaker 2: to be on top of it. You need to make
11:36
Speaker 2: sure that it has a human fingerprint. If it doesn't
11:40
Speaker 2: and it's way too aid, then it's going to lose
11:44
Speaker 2: its meaning it's not going to be as valuable as
11:46
Speaker 2: it would be, and we might be falling into replications
11:49
Speaker 2: of aislab even.
11:51
Speaker 4: Lots of talk mark of gloves and fingerprints and ungloved hands.
11:54
Speaker 4: So far maybe pointing us towards the question of what
11:58
Speaker 4: stays the most human, Oreana kind of flagged this notion
12:02
Speaker 4: of discernment as a kind of critical component in successful
12:06
Speaker 4: AI transformation for you, where the trust, judgment context still
12:11
Speaker 4: matter most in it.
12:13
Speaker 6: Yes, I think it does because we're still a human society.
12:18
Speaker 6: We're still having companies with humans in them, and I
12:21
Speaker 6: think we need to step profoundly human.
12:24
Speaker 3: In the ways that we're doing it.
12:27
Speaker 6: And you're bringing the concept of trust, for example, and
12:30
Speaker 6: I think this is something very key. It's I think
12:32
Speaker 6: more difficult to trust an AI the same way you
12:35
Speaker 6: will trust a human when you have a relations You know,
12:38
Speaker 6: humans are looking for relationship, right when you're doing stickholder management.
12:41
Speaker 6: Could it be discussing with the business, Could it be
12:44
Speaker 6: having a meeting with your manager, and so on? You
12:46
Speaker 6: need that trust to make sure that the relationship and
12:50
Speaker 6: the collaboration works correctly, right, And I think this is
12:54
Speaker 6: something that is very difficult to delegate to an ELM, right,
12:58
Speaker 6: because this is just a machine at the end. And
13:01
Speaker 6: I think there's also some elements that are also profoundly
13:04
Speaker 6: human that is difficult for them to pick up, things
13:07
Speaker 6: such as, you know, the organizational context. Could it be politics,
13:14
Speaker 6: internet politics, rules, how to behave.
13:17
Speaker 3: And so on and so on. And I think that's
13:19
Speaker 3: something that also.
13:20
Speaker 6: Contributes to those relationships that we're building and would be
13:23
Speaker 6: difficult to delegate to THEI.
13:29
Speaker 2: Right.
13:29
Speaker 4: This is where we get to what really fascinates me
13:31
Speaker 4: for most in this whole topic. Okay, everybody, and that's
13:35
Speaker 4: the topic of what does the new IT ladder look like? Mark,
13:40
Speaker 4: Let's begin with you. If if entry level work used
13:43
Speaker 4: to mean ticket triage, basic troubleshooting, what does it mean now?
13:48
Speaker 4: What will it mean tomorrow?
13:50
Speaker 6: So to speak? Yes, I think and that goes back
13:53
Speaker 6: to something that I shot earlier on is I think
13:56
Speaker 6: AI would allow us to actually do more.
13:59
Speaker 3: Things that what we were able to do so far,
14:01
Speaker 3: in the sense.
14:01
Speaker 6: That it's going to take away some of the tedious
14:05
Speaker 6: tasks that we're currently doing as part of our jobs,
14:08
Speaker 6: and we had free up time for us to then
14:10
Speaker 6: start experiencing more testing, more being more out there, being
14:15
Speaker 6: more trying to understand the business, being more trying to
14:18
Speaker 6: understand you know, we are somehow the things that are
14:24
Speaker 6: really going to move the needle as part of the
14:27
Speaker 6: IT that you're delivering to the business right And I
14:31
Speaker 6: think this is quite exciting because not only that allows
14:34
Speaker 6: you to spend more time there, but it also gives
14:37
Speaker 6: you tools to experiment way more, so you as an
14:42
Speaker 6: IT professional can start actually exploring solutions for the business
14:48
Speaker 6: to issues that they have and so on in ways
14:50
Speaker 6: that you could never before. Now it's very easy for
14:53
Speaker 6: anyone in IT to go to an AI and to
14:56
Speaker 6: prototype something, to create a productivity app, you know, to
15:00
Speaker 6: do something with AI that you can then somehow look at, Okay,
15:07
Speaker 6: there is there any value for me to bring that
15:09
Speaker 6: to the business. Does that make us as part of
15:11
Speaker 6: my job to do this and that somehow? So I
15:15
Speaker 6: think going to make people do things that maybe today
15:20
Speaker 6: don't have as part of their scope for their job description,
15:23
Speaker 6: I think I will give them the potential to also.
15:28
Speaker 3: Well also somehow blend a bit.
15:29
Speaker 6: More roles in the IT environment, in the the IT departments,
15:34
Speaker 6: people that today are in the box and only doing
15:37
Speaker 6: one thing will be able to actually start collaborating more
15:40
Speaker 6: with other IT professions than in your organization, doing more
15:43
Speaker 6: collaboration and so on. And I think this is also
15:47
Speaker 6: something towards the evolution of IT. But I think any
15:51
Speaker 6: types of jobs in general in the future.
15:56
Speaker 4: Yeah, I think there's an interesting topic in this area,
15:59
Speaker 4: as in some many others, mark of like if how
16:03
Speaker 4: do people really learn it? If the first trust of
16:06
Speaker 4: everything AI generated, and it has particular I think pertinence
16:09
Speaker 4: to coders, you know inside inside computing, and is there
16:13
Speaker 4: a risk with automation of kind of hollowing out that
16:17
Speaker 4: apprenticeship layer or having a kind of human talent based
16:21
Speaker 4: that kind of lack itself, lacks a foundation of awareness
16:25
Speaker 4: and knowledge.
16:27
Speaker 6: I think there's a risk, but I think it's not
16:31
Speaker 6: necessarily always negative and bad in the sense that I
16:36
Speaker 6: think with the I, what we're actually doing is we're
16:40
Speaker 6: reducing the barrier for anyone to be able to do
16:44
Speaker 6: again more.
16:45
Speaker 3: Than what they were able to do so far.
16:47
Speaker 6: Examples of that is, even though you're not an expert
16:51
Speaker 6: of an AI system, you can start inter interacting with
16:56
Speaker 6: AI that knows all about that system, and you can
16:59
Speaker 6: start actually leveraging that system to do your work.
17:03
Speaker 3: Before that would have taken.
17:04
Speaker 6: Maybe months, depending on how complex the system is to
17:08
Speaker 6: fully master it, fully understand all the bits and bites
17:10
Speaker 6: and on. Now, because you have that intermediary layer, which
17:15
Speaker 6: is AI, you can go to it and you can
17:17
Speaker 6: somehow start just sharing the outcome that you expect, the
17:22
Speaker 6: intent that you have right and AI will be the
17:26
Speaker 6: middleman somehow and we'll be able to interact with that
17:29
Speaker 6: system and you will get the insights, the value, the
17:32
Speaker 6: action that you want that system to perform in a
17:35
Speaker 6: more easy way, so that means that anyone, even potentially
17:40
Speaker 6: more junior people could start doing more than today would
17:44
Speaker 6: have tach cares for them to get as an experience. Right,
17:47
Speaker 6: So I think this is something that is quite good.
17:50
Speaker 3: Now.
17:50
Speaker 6: Another dimension associated to all of that is a lot
17:54
Speaker 6: of things that we do today are really based on experience.
17:57
Speaker 6: You need to try something, you need to see it
17:59
Speaker 6: fail for you to learn something, and next time you
18:02
Speaker 6: will be a bit more of an expert of that
18:04
Speaker 6: domain and then be able to.
18:08
Speaker 3: To be faster and and and better somehow.
18:10
Speaker 6: Right. If now we're delegating all of that to I,
18:15
Speaker 6: you're not always learning, so that that's a risk, that's
18:17
Speaker 6: a challenge, right, So that's something that we need to
18:23
Speaker 6: see how it's going to evolve. What's interesting is we
18:26
Speaker 6: see different types of AI systems out there by the way,
18:28
Speaker 6: Some of them, like clothes, are quite collaborative with the
18:32
Speaker 6: users in the sense that if you go there and
18:35
Speaker 6: you ask it to perform something, it will give you
18:38
Speaker 6: a lot of feedback as us are like, oh, okay,
18:41
Speaker 6: you want me to do this, I will try.
18:43
Speaker 3: That's that's my plan. I will try to do this, this, this,
18:46
Speaker 3: and if it fails, tells.
18:47
Speaker 6: You how Okay, it feels no, you know what, maybe
18:49
Speaker 6: the best is to try to do that, that that
18:51
Speaker 6: and that and that could be also a way for
18:53
Speaker 6: people to learn with the AI somehow. So it's because
18:57
Speaker 6: you see what's happening and you see what works best,
18:59
Speaker 6: and maybe next time you can prompt it in a
19:01
Speaker 6: different way to guide it to make sure that it's
19:03
Speaker 6: doing better.
19:05
Speaker 4: Dina, any thoughts on this version of like that skill,
19:08
Speaker 4: that learning layer that Mark.
19:10
Speaker 6: Is distributed to.
19:11
Speaker 4: I know you're very you're very you're very dedicated team nurturing.
19:16
Speaker 2: Yes, yes, I am, and learning. I think it's a
19:21
Speaker 2: bit of everything that Mark mentioned, the trial and error,
19:23
Speaker 2: the fact that this is how we mostly learn in
19:26
Speaker 2: it and in life. We're going to do that quicker
19:28
Speaker 2: now with AI, we're starting from knowledge. We're starting from
19:32
Speaker 2: information that's already being gathered and most probably trianged first layer.
19:37
Speaker 2: So we should use that as human beings in our
19:39
Speaker 2: own benefit to learn faster with a I, to learn
19:42
Speaker 2: quicker and actually compact the amounts of experiences that you
19:46
Speaker 2: would take generally without a I maybe years to scratch
19:50
Speaker 2: the surface and understand it. You can do that in
19:53
Speaker 2: a much much shorter amount of time. So I think
19:56
Speaker 2: we're more empowered. We should raise the bar, we should
19:59
Speaker 2: raise our targets. We should say, Okay, this task used
20:02
Speaker 2: to be finishing with this amount of quality and x
20:04
Speaker 2: amount of time. With the use of AI, you can
20:07
Speaker 2: do that much more efficiently, much more quicker. And I'm
20:12
Speaker 2: pretty sure this will increase the collaboration between people. We
20:16
Speaker 2: will work more on our own emotional intelligence and collaboration
20:21
Speaker 2: with each other. So I think we can use AI
20:23
Speaker 2: in our own benefit. It's just one more tool that
20:26
Speaker 2: humans made that will still be used. But however it's
20:30
Speaker 2: going to be prevailing or not, that's all up to us.
20:33
Speaker 2: We just need to change our perspective on how our
20:36
Speaker 2: current roles are looking like within IT and in non
20:40
Speaker 2: IT environments as well, so we can embrace that change.
20:43
Speaker 2: The quicker we play with AI, practice with AI, learn
20:47
Speaker 2: with AI, the more we're going to be able to
20:50
Speaker 2: master it and integrate it in our ways of working
20:53
Speaker 2: every day.
20:54
Speaker 4: Super And let's think about future roles and from current
20:58
Speaker 4: roles to future roles. Mark what the what?
21:01
Speaker 2: What?
21:01
Speaker 4: What roles are starting to emerge in response to this
21:05
Speaker 4: new kind of backdrop of technology.
21:10
Speaker 3: Or what roles do you expect to see.
21:12
Speaker 4: Emerging in the coming two three years, but maybe some
21:16
Speaker 4: people haven't imagined yet even.
21:19
Speaker 6: Yeah, this is this is actually something interesting to look
21:23
Speaker 6: at from what's happening right now. There's there's roles that
21:27
Speaker 6: started emerging I would say are not very surprising, so
21:30
Speaker 6: I would not spend too much time on them. But
21:32
Speaker 6: I think we see you know, uh, new roles around
21:35
Speaker 6: how to make sure that you have the right infrastructures
21:38
Speaker 6: or the right operations around AI.
21:40
Speaker 3: So if you start, you know, skinning up.
21:42
Speaker 6: Your agents, you know in your environment, you need to
21:44
Speaker 6: make sure that you have the right set up in place,
21:46
Speaker 6: the right infraud or right governance and so on for
21:49
Speaker 6: those agents to to operate right. So I think this
21:52
Speaker 6: is something that we see emerging.
21:56
Speaker 2: Uh.
21:57
Speaker 3: We have also new you.
21:58
Speaker 6: Know things around secure for A systems and so on,
22:01
Speaker 6: some new new skids on new competencies around that that
22:04
Speaker 6: are somehow required. But there's one role I want to
22:09
Speaker 6: spend a bit more time on because I found it
22:12
Speaker 6: very interesting. And this is also closer to the core
22:16
Speaker 6: mission of net Sinc. Somehow, which is this concept of
22:23
Speaker 6: designing the right AI processes okay, and making sure that
22:28
Speaker 6: the experience of those processes are the best ones. What
22:33
Speaker 6: I mean by that is, if you start having more
22:35
Speaker 6: and more of those AI agents doing workflows in your governments,
22:39
Speaker 6: doing you know, business workflows and so on, you need
22:43
Speaker 6: to make sure to understand, Okay, what are the current
22:45
Speaker 6: processes in place, how should you turn them into an
22:51
Speaker 6: AI ready process where AI agents actually will do some
22:55
Speaker 6: steps of that process. Maybe it's not just sporting the
22:58
Speaker 6: existing process right, maybe it's also perstforming those processes into
23:02
Speaker 6: something new, but also how an AI agent is going
23:06
Speaker 6: to go from one system to the next one as
23:09
Speaker 6: part of the process.
23:10
Speaker 3: Is there any friction? Is there any latency?
23:13
Speaker 6: Is there anything that is really impacting how this agent
23:17
Speaker 6: is productive.
23:18
Speaker 3: As part of the process.
23:19
Speaker 6: And as you can see, this is actually echoing what
23:23
Speaker 6: we're currently offering as capabilities through net Sinc. But for
23:27
Speaker 6: employees today were focusing really on the digital employee experience.
23:31
Speaker 6: But somehow maybe in a world where you have a
23:33
Speaker 6: lot of AI agents running in your environment, you want
23:37
Speaker 6: also to handle their experience. Even though maybe experience is
23:41
Speaker 6: not the right word for an AI agent, the concept
23:44
Speaker 6: is very similar, right, And I think this is something
23:47
Speaker 6: that will emerge as a role well.
23:49
Speaker 5: I think there's an interesting sort of analogy to be
23:52
Speaker 5: made to any other sort of outsourcing, right, Like you
23:56
Speaker 5: can distribute skills across a variety of different workforces in
24:04
Speaker 5: different ways, and one of the things that you have
24:11
Speaker 5: to do with that is is really think about your
24:14
Speaker 5: policy and your process and where you're using those different resources.
24:21
Speaker 5: I'm wondering if you can elaborate a little bit on
24:23
Speaker 5: that last part of you were talking about, because I
24:25
Speaker 5: think that's the part that I find really interesting in
24:28
Speaker 5: terms of, you know, how we're flying through with the
24:33
Speaker 5: idea of building skills. I'm also thinking about sort of yeah,
24:38
Speaker 5: the just sorry, this is a little hopping around Beason
24:42
Speaker 5: thinking about all the different things that you said, But
24:44
Speaker 5: I was thinking about the skill component, especially because I've
24:49
Speaker 5: been thinking about how do we train models to have
24:54
Speaker 5: a specific set of skills, right, whether it's like a
24:57
Speaker 5: formal skill capital s in something like Claude or custom GPTs,
25:04
Speaker 5: the things where we think about, like how do we
25:07
Speaker 5: make things more tailored to a particular tool and task. Yeah,
25:16
Speaker 5: I'd be interesting hearing more of your thoughts on how
25:21
Speaker 5: that dovetails with that experience of also learning and getting
25:25
Speaker 5: better as both an it expert and an AI user.
25:33
Speaker 6: I think you're touching upon something interesting because it has
25:38
Speaker 6: a parallel to the humans.
25:40
Speaker 2: Right.
25:40
Speaker 6: So when you're talking about skills, I think you're referring
25:42
Speaker 6: to the skills as probablyanthropic concepts, right that an anthropic
25:46
Speaker 6: poosht and now all the major eye players are are using.
25:50
Speaker 3: Right.
25:50
Speaker 6: So the skill in that context is really sets of knowledge, instructions,
25:57
Speaker 6: tools that you plug into your agents for them to
26:00
Speaker 6: perform specific tasks. Right, And like humans, I think what's
26:07
Speaker 6: interesting is you need to Okay, you need to train
26:10
Speaker 6: your agents and and you need to define sets of skills,
26:13
Speaker 6: but maybe those skills you need to maintain them, maintain them.
26:16
Speaker 6: You may realize that actually the instructions that you give
26:19
Speaker 6: to your agents to perform a task is maybe the
26:22
Speaker 6: most not optimized one. So maybe you need to train
26:25
Speaker 6: them again by updating, if you like, their their skill
26:28
Speaker 6: again with a better up to date sets.
26:31
Speaker 3: Of instructions, for example. So all that management of how
26:36
Speaker 3: you performing processes needs to be it's not fire and forget.
26:45
Speaker 6: This is something that you need to make sure that
26:47
Speaker 6: you maintain, assess all the time, evolve when required, and optimized. Right,
26:54
Speaker 6: So this is really something that you would do continuously
26:56
Speaker 6: as like what's currently happening in position. Just with humans,
27:01
Speaker 6: you need to keep on treating your employees, reviewing your processes,
27:05
Speaker 6: making them better, and then sharing all those changes with
27:09
Speaker 6: your employees for them to start diverging those new processes
27:11
Speaker 6: and so on. So I think here it's really the
27:14
Speaker 6: analogy and the management of that is exactly the same
27:17
Speaker 6: for e agents for par to humans.
27:21
Speaker 5: Yeah, and I think maybe part of that top of
27:22
Speaker 5: mind is that that's what you and I have been
27:24
Speaker 5: talking about with the syste and how do we make
27:27
Speaker 5: sure that we are continuously building those new skills and
27:32
Speaker 5: refinements into the tool and make it really easy and
27:36
Speaker 5: possible to continue to grow those capabilities in the right
27:40
Speaker 5: places where they're needed.
27:43
Speaker 4: Correct and oreanor I think I might give you the
27:48
Speaker 4: last question directly before we give it to Mark for
27:51
Speaker 4: the last word, which is, what do you think responsible
27:54
Speaker 4: transition looks like? How to leaders pursue automation without treating
27:59
Speaker 4: it as a headcount strategy? What does the reskilling process
28:04
Speaker 4: we've alluded.
28:04
Speaker 6: To here look like?
28:05
Speaker 5: In part Well, I think this goes back to some
28:08
Speaker 5: of the conversations we had much earlier on reality bites
28:14
Speaker 5: where we talk about policy in terms of what are
28:18
Speaker 5: you trying to accomplish? Right, Like, yes, of course you
28:20
Speaker 5: can use AI as a headcount reduction tool. Will it
28:25
Speaker 5: be good for your organization? Well, it depends how do
28:28
Speaker 5: you define good, right, And this is something that we
28:34
Speaker 5: talked about quite a bit with Tim, so not to
28:36
Speaker 5: you know, beat the same drum too many times, but
28:39
Speaker 5: I think if we're thinking about what makes your company
28:43
Speaker 5: most effective in the long run versus what allows you
28:47
Speaker 5: to save a few dollars in the next few months,
28:50
Speaker 5: those are often different policies and different strategies, and those
28:54
Speaker 5: are business questions, right. And you can replace workers with
28:59
Speaker 5: an outsourced team, and you also have sometimes a dip
29:03
Speaker 5: in quality and you know, cost saving, and that's something
29:08
Speaker 5: that organizations did a lot over the past few decades.
29:12
Speaker 5: And we've seen a redistribution of those skills. And I
29:16
Speaker 5: think that's what we're going to continue to see, is
29:19
Speaker 5: companies redistributing different levels of skill thinking about, you know,
29:23
Speaker 5: where do we really want to invest and also where
29:26
Speaker 5: do we want to invest long term? Where do we
29:28
Speaker 5: want to make sure that we are growing the talent
29:31
Speaker 5: that will help us be successful in the long run
29:33
Speaker 5: and really having genuine knowledge and skill that we're cultivating.
29:40
Speaker 2: Yes, it's not about reducing the headcounts. We're getting time
29:44
Speaker 2: back to do more, to do actually the stuff that
29:48
Speaker 2: we always dreamed of doing, rather than spending more time
29:51
Speaker 2: into exploring, trial and edit and all of that. So
29:55
Speaker 2: we should look at this from a strategic perspective as
29:58
Speaker 2: now we're going to be more productive or will reach
30:01
Speaker 2: higher targets in quicker amount of time. So this is
30:04
Speaker 2: basically like the shortest value statement that might fit most
30:08
Speaker 2: of the organizations. It will just need to be a
30:11
Speaker 2: little bit customized and tailored. The upskilling and keeping people
30:16
Speaker 2: up too poor. We just need to make sure that
30:18
Speaker 2: every employee knows what is their KPI, is what they
30:23
Speaker 2: are being asked for, and how they can really put
30:25
Speaker 2: the use of AI into useful day to day use
30:28
Speaker 2: cases end to end that are really useful in their
30:32
Speaker 2: productive aspect and it would really contribute into the growth
30:36
Speaker 2: of the business at the end of the day.
30:39
Speaker 4: And Mark has some pretty good reflections there on transition.
30:44
Speaker 4: You get the last word in this discussion. I know
30:46
Speaker 4: we didn't touch upon absolutely everything, but we're almost out
30:49
Speaker 4: of time. So if there's anything that we had to
30:51
Speaker 4: skip over you that you want to come back to,
30:54
Speaker 4: or indeed you want to add your own two cents
30:55
Speaker 4: to that question of op scaling, re scaling and transition
30:58
Speaker 4: for floor is yours?
31:01
Speaker 6: Yes, I think it's I think it's a very exciting
31:04
Speaker 6: time to be honest, because again, I think the AI
31:07
Speaker 6: technologies are evolving so fast that there's so much potential
31:13
Speaker 6: behind it that my advice is for people not to
31:18
Speaker 6: be left behind and really start actually embracing those technologies,
31:23
Speaker 6: having a better grasp at what they can do, what's
31:26
Speaker 6: their potential, and then try to translate that into how
31:31
Speaker 6: can they can that transform how they are working today,
31:35
Speaker 6: how can they actually be more productive that even more
31:38
Speaker 6: value for the business.
31:39
Speaker 3: Which is not always easy so far.
31:42
Speaker 6: I think it's a lot of experiments and people trying
31:47
Speaker 6: to tailor to this very specific use case.
31:49
Speaker 3: I think there's going to be more major convergence.
31:52
Speaker 6: About how to use AI really in the professional words
31:55
Speaker 6: in maybe the coming year, but for the time being, yes,
31:59
Speaker 6: really have fun with it, have a look at what
32:02
Speaker 6: it can do and how you can apply it to
32:04
Speaker 6: your to your work.
32:06
Speaker 4: Great stuff, Mark, And a final final final question for you,
32:09
Speaker 4: I've been curious, a spark and assist getting along. Okay,
32:12
Speaker 4: there been any friction between them.
32:14
Speaker 3: They are the best budies. There are the best buddies.
32:17
Speaker 4: Yes, there the best buddies. That's a happy note to
32:20
Speaker 4: end on.
32:20
Speaker 7: Well, look, I hope everybody enjoyed our rebooted, relaunched Reality Bites.
32:26
Speaker 7: It was a delight for me to have Oriana and
32:28
Speaker 7: Dina back, and indeed Mark to come back on the show.
32:30
Speaker 7: And I didn't even find I didn't even could have
32:32
Speaker 7: find point on what's a bit different about it. So
32:34
Speaker 7: we used to have the guests and the kind of
32:36
Speaker 7: interrogative approach to the guest.
32:39
Speaker 3: And we wanted to open it up and have a.
32:40
Speaker 4: More of a fluid conversational format. And look, I think
32:45
Speaker 4: that worked really really well. I really enjoyed it. I
32:47
Speaker 4: thought we covered some really interesting territory. And I guarantee it,
32:52
Speaker 4: after that long hiatus, we're going to be back next
32:55
Speaker 4: month with Reality Bites panelists that you know and love
32:58
Speaker 4: and a new special guest and a new topic to discuss,
33:01
Speaker 4: So be sure to tune into that. And it just
33:04
Speaker 4: remains for me to say thank you so much Oriana,
33:06
Speaker 4: Dina and of.
33:07
Speaker 3: Course Mark, see you all very soon live.
33:13
Speaker 2: That's it for this week's Deck Show.
33:15
Speaker 4: We'll be back next week, and if you don't already,
33:18
Speaker 4: be sure to follow the show so you don't miss
33:19
Speaker 4: an episode until next time.