Latest / Women Disrupting Tech / Navigating AI and Leadership as Trengo CEO with Marili 't Hooft-Bolle | Ep. 116

Transcript

Navigating AI and Leadership as Trengo CEO with Marili 't Hooft-Bolle | Ep. 116 · Women Disrupting Tech

• 0:10
  Welcome to Women Disrupting Tech.
• 0:13
  My name is dear Kanupkas, and on this podcast, I give the women
• 0:16
  who work in tech the platform they need to inspire and empower
• 0:20
  each other. Join me on a journey into the
• 0:23
  world of entrepreneurship, funding biases, and role models
• 0:28
  to discover the strategies for success in the tech industry.
• 0:32
  In other words, how women can disrupt tech.
• 0:36
  Are you ready to listen? Hey, welcome back to Women
• 0:45
  Disrupting Tech and thank you for tuning into episode 116.
• 0:51
  We're living in an age of AI that feels like the land of
• 0:54
  confusion. There are new tools everyday
• 0:58
  that promise to change everything, but for founders,
• 1:02
  the real challenge remains the same.
• 1:05
  How do you grow your business without losing what makes it
• 1:08
  work? How do you stay hyper agile
• 1:11
  without burning out your team? And how do you put structure in
• 1:15
  place without killing everything in the process?
• 1:19
  My guest today is Marley Atov Polo.
• 1:22
  She's the CEO of Trango. She helped scale We Transfer and
• 1:27
  now guides Trango through this AI driven world.
• 1:31
  Having LED companies through rapid growth, she knows the
• 1:35
  tension between building structure and keeping A-Team
• 1:38
  agile. In our conversation, she shares
• 1:41
  why scaling is never just about adding people, how
• 1:45
  self-awareness makes or breaks leadership, and why
• 1:49
  complimentary teams always deliver stronger results.
• 1:53
  Her biggest lesson is this Growth requires structure, but
• 1:59
  not too much. If you get it right, you'll
• 2:02
  navigate change with confidence even in the age of AI.
• 2:07
  So when you're ready to listen and learn how to navigate growth
• 2:11
  in the age of AI, please welcome to women disrupting tech.
• 2:15
  Marie Lee at Holt Paula Marie Lee, welcome to women disrupting
• 2:19
  tech. Good to have you.
• 2:22
  Question I always ask at the start is how you got into your
• 2:25
  current position and in your case as CEO of Tringo?
• 2:30
  Yeah. Well, I've been at Trango for a
• 2:32
  little over 2 years now. And I started speaking with the
• 2:37
  founders actually quite a few months before that.
• 2:40
  And they were initially looking for a CEO position.
• 2:44
  But when I dug a little bit deeper, it felt as if they
• 2:46
  needed a CEO and we actually started out on an interim
• 2:53
  contract. And I think that is a very good
• 2:56
  thing to start, right, Because it's kind of the founders
• 2:58
  handing over their baby to an external person.
• 3:01
  You can talk about that in a room and interviews in theory
• 3:04
  for a long time and like each other, but I feel that you
• 3:06
  really need to be on the floor to make sure that you're on the
• 3:09
  same page and it actually works in practice.
• 3:11
  And then secondly, it also allows for a bit of time to form
• 3:15
  a plan and a strategy and make sure that founders, investors
• 3:19
  and a new person, me coming in are all on the same page in
• 3:23
  terms of strategic direction. And so that's, that's worked out
• 3:29
  in a in a matter of a few months.
• 3:31
  And so then we made it, we made it final.
• 3:34
  And here I am, 2 years a little bit in, and I think this setup
• 3:39
  really allows everyone between the founders and myself to play
• 3:42
  to our strengths. Yeah, I'll get into paying into
• 3:47
  each other's strengths in a bit. I just was wondering, I seem to
• 3:50
  remember that there was a little hiccup in the hiring process
• 3:55
  where you sort of went your own separate ways and then came
• 4:00
  back. Yeah.
• 4:01
  So I think that was probably at the start of the year, a little
• 4:06
  before the end of the year before I started.
• 4:10
  And indeed, right, it's a difficult decision as founders
• 4:14
  to make to find a replacement for yourself in certain roles.
• 4:19
  And so at that point, I also said it's, it's got to feel good
• 4:23
  for you, right? If you're not fully on board and
• 4:25
  you're not sure, then maybe you should think about it a bit
• 4:27
  longer. And so indeed, there was a bit
• 4:29
  of radio silence. I moved on.
• 4:31
  I actually took on a on a different project and then a few
• 4:36
  weeks in one of the founders came back to me said can we go
• 4:39
  for lunch after all because I really think it would be a
• 4:41
  really good fit and I think we should start speaking again.
• 4:45
  And so I think I probably started with a quarter delay
• 4:48
  because of that, but everything happens for a reason.
• 4:51
  So I'm sure that part of the decision making process was was
• 4:55
  required. You know why he came back,
• 5:01
  besides thinking that it was a good fit.
• 5:03
  Well, no, mostly that and making sure that I was still keen, of
• 5:07
  course. So it was, I mean, there were
• 5:09
  three founders, so I spoke to them on different occasions and
• 5:13
  yeah. OK, So what does Trango do?
• 5:18
  Trango is a customer engagement platform.
• 5:20
  It's pretty much a place where you can handle all the
• 5:23
  communication with your customers or potential
• 5:26
  prospects. And in the old way you could say
• 5:29
  that is a ticketing system and an inbox where all your channels
• 5:32
  come in, so your WhatsApp, your Instagram, your e-mail,
• 5:35
  etcetera. But actually, in today's world,
• 5:38
  customer communication, kind of the support inbox and the
• 5:42
  marketing automation are starting to blend through each
• 5:45
  other. And so we made the platform in
• 5:49
  such a way that it's very easy to build your AI agents on them.
• 5:53
  And so that kind of has three components.
• 5:55
  Is that what can the AI agents talk about?
• 5:58
  So that is your knowledge, the information, but also which
• 6:01
  systems can the agent actually go into?
• 6:03
  The second part is how do you want your agent to behave?
• 6:06
  So what can they talk about? What can they not talk about?
• 6:09
  Do they use emojis? What form of speech, etcetera.
• 6:12
  And the third one is actually the whole orchestration,
• 6:15
  orchestration logic around it as to what agent actually steps in
• 6:19
  where. And if they can't solve it,
• 6:20
  where do they escalate to? And so all conditions around
• 6:24
  which agent operates, how and when and that we do for a not
• 6:30
  super tech savvy segment, we we sell into the consumer services
• 6:35
  kind of mid market companies. So you should think about a
• 6:38
  chain of beauty clinics or a chain of hotels or an amusement
• 6:43
  park. And they make it really easy for
• 6:46
  them to automate, let's say between 40 to 80% of all their
• 6:50
  customer and prospect conversations so they can focus
• 6:53
  on driving revenue and delivering value.
• 6:56
  OK. So what's in it for them in
• 6:58
  terms? Is it also cost reduction or
• 7:02
  time reduction? I think anytime where you use
• 7:05
  AI, it's like the question is, is it just an efficiency point?
• 7:10
  But I would say it actually the knife cuts both ways in this
• 7:13
  one. So on the one hand, you can
• 7:14
  handle your incoming and let's say customer support function
• 7:19
  more efficiently. But that actually means that
• 7:21
  you're freeing up people who know your product and your
• 7:23
  services in your market really well to do more revenue driving
• 7:26
  or more guest satisfaction type of activity.
• 7:30
  So whilst you're reducing your costs, you are also actually
• 7:35
  freeing up the resources to drive value.
• 7:37
  So I see it as a two way impact. OK, so can you give an example
• 7:44
  of how this could work both proactively and reactively?
• 7:48
  Yes, reactively is very much let's say you are going to a
• 7:52
  festival and you've booked a campsite and and you have
• 7:56
  questions on whether you can bring your dog in or maybe you
• 7:59
  have questions around whether you can bring your own drinks
• 8:02
  and foods to the place. Well, usually there are quite
• 8:06
  standard regulations per campsite or per let's say a
• 8:11
  festival. Maybe that is a reactive way
• 8:13
  where the customer comes in and says, OK, I need to know these
• 8:16
  things and then your help straight away.
• 8:19
  Or it can even be your on site concierge the the AI agent
• 8:23
  whilst you're at the festival to say, OK, what's the latest again
• 8:26
  on the program? Do I have AVIP package?
• 8:30
  Where do I go? And then they can answer those
• 8:32
  questions and the proactive can be the company knowing that
• 8:37
  you've just arrived and say, hey, listen, because you've
• 8:39
  booked your VIP package, did you know you're also getting a 30%
• 8:43
  discount on the merchandising? And hey, guess what, here's a
• 8:46
  link where you can go and buy it.
• 8:48
  Or hey, here's a coupon that you can redeem at the stand at the
• 8:52
  festival, for example. So those are two different ways
• 8:55
  of where the initiation of the communication stream starts.
• 8:59
  OK, so in that sense it could also does it then also replace
• 9:04
  apps? Where is this in app?
• 9:06
  No, it can be in app and it can is, it doesn't necessarily have
• 9:11
  to replace the app, right? The thing is, if you put a
• 9:14
  website or an app and you have really elaborate information,
• 9:18
  usually consumers are quite lazy.
• 9:20
  So they'll see a number and they press on it and they call,
• 9:23
  right? They don't actually go and
• 9:24
  browse everything. And if the search functionality
• 9:26
  isn't great, then it becomes harder.
• 9:28
  So now this, this AI agents can be your, let's say, little
• 9:32
  helper there. If you can't find the
• 9:35
  information intuitively in app, it can still be part of the app
• 9:38
  where that actually happens. OK, so then it would be sort of
• 9:43
  like you have with ChatGPT where in the old days with Google just
• 9:47
  search or a chat screen. Yeah, it would be a chat screen.
• 9:53
  That wouldn't be a search functionality then, because the
• 9:55
  search functionality is what you then prompt and try and find and
• 9:59
  then you get a few links and it may not be accurate.
• 10:01
  Whereas with the agent, obviously you get the exact
• 10:04
  answer to what you're looking for.
• 10:06
  OK. I see.
• 10:07
  So mid market being like really big hotel chains, no.
• 10:14
  No, I wouldn't. Say that type of thing.
• 10:16
  No, we do have a few of the bigger names, but then maybe
• 10:21
  it's very specific part. For example, we have Ritz
• 10:24
  Carlton as our customer. We do their boating, their
• 10:27
  yachting business. But I would say maybe more the
• 10:30
  hotel chains that would have between 40 and a few hundreds
• 10:35
  locations or the Tier 2 ones I would say.
• 10:39
  OK. Is that is that on purpose or?
• 10:43
  Well, that is a segment that we know that we can really provide
• 10:46
  value for. If you're going into the other
• 10:49
  one, you get into an enterprise market and that has other
• 10:52
  product requirements, which is also a deliberate choice, right?
• 10:55
  You can't serve all the segments at the same time.
• 10:58
  And so we feel that we have a very strong proposition for that
• 11:00
  part of the market. OK, I want to move back to you
• 11:05
  as a person. So I read, I think on LinkedIn
• 11:12
  that you want to work for fast growing companies where
• 11:15
  innovation is the name of the game.
• 11:17
  What makes innovation attractive to you?
• 11:19
  A. Good question.
• 11:20
  I think by nature I'm not very much a person of routine.
• 11:25
  So anytime I get into routine and predictability, I I start
• 11:29
  finding a little bit of boredom, I would say.
• 11:33
  And the tech industry just has a certain pace to it.
• 11:36
  So it keeps changing etcetera. And I feel that I like
• 11:41
  challenges both from a very intellectual point of view and
• 11:44
  on a people side. So the more of an IQ and EQ
• 11:47
  balance that those things marry up really well in in a fast
• 11:51
  moving environment. And I just think it's the
• 11:53
  excitement and complication of, for example, something like AI
• 11:58
  coming into the mix now that can turn whole industries upside
• 12:01
  down, which is super exciting. And then trying to figure out
• 12:04
  how to play with those new rules.
• 12:06
  So I think I prefer difficult and changing environments also
• 12:10
  in my as a personality over something that is predictable,
• 12:14
  steady and, and, and let's say more routine.
• 12:19
  And now maybe your next question is why is that so I find a
• 12:23
  difficult one to to to analyze, but I do know that that that
• 12:30
  that suits me really well. It's.
• 12:32
  You told me that you when you go to question what you wanted to
• 12:38
  be, you wanted to be a scientist and also a bit of a social
• 12:41
  worker, which is sort of a split that for well, if I when I first
• 12:47
  heard it, I was like, oh, OK. Yeah, that was back in school.
• 12:52
  So indeed it's quite schizophrenic.
• 12:54
  So it's on on a very social side and a very analytical, let's say
• 13:01
  what they didn't touch maybe called the beta studies.
• 13:04
  And so I like, I indeed like both.
• 13:06
  And so I'm trying to find spaces where both of those things have
• 13:10
  a challenge. So if, let's say, if it was only
• 13:13
  an analytical job, then I would miss the people aspect.
• 13:16
  It was only about people. I would miss the hardcore kind
• 13:19
  of analytical cracking your brain side.
• 13:22
  Yeah. Yeah.
• 13:23
  So that makes it managing a start up slash Caleb pretty much
• 13:30
  the ideal working environment. That was I've found so far, yes.
• 13:36
  Yeah. So before Triangle you worked
• 13:40
  also for a company called We transfer that most of the
• 13:45
  listeners will have used in at least in the past.
• 13:49
  I'm still using it. If I remember correctly you
• 13:53
  skilled helps the company skill from about 50 full time people
• 14:01
  to 200 plus. What?
• 14:06
  What is the secret behind doing that?
• 14:09
  How do your strong points play into that?
• 14:13
  I think first of all, I like when, when people say, oh, I've
• 14:16
  scaled from this to this, then they often mention indeed the
• 14:20
  people, but I'm always like, oh, maybe you could have reached the
• 14:23
  same result with less people, right?
• 14:25
  So adding more people is not necessarily a successful
• 14:28
  exercise, but indeed it's about scaling an organization, scaling
• 14:32
  a business. And I think on every kind of
• 14:35
  size, it's, it's, it's about adding enough structure, but not
• 14:40
  too much. So you can completely over
• 14:43
  engineer and try and run a 200% company like like an enterprise,
• 14:48
  but there's no real need for that, right?
• 14:50
  And the, the nice thing about to having only 200 people is that
• 14:52
  you still know the names and the faces and there's still a lot of
• 14:55
  mandate that you can give and you can move fast.
• 14:57
  So you don't want to kill that with, with process.
• 15:01
  Whereas in the very small companies, let's say up to 50,
• 15:04
  you don't need a lot of process whatsoever because people will
• 15:07
  be sitting at the same lunch table almost every day and then
• 15:10
  you overhear stuff. But once you reach the next
• 15:13
  stage, you need to define certain, let's say communication
• 15:16
  flows or decision structures in order to actually keep that base
• 15:22
  and and and get to get to results.
• 15:26
  Even with the that size of a company with however small that
• 15:31
  still is, you can still have silos occurring.
• 15:33
  And I think that is where stuff sometimes crumbles down when the
• 15:38
  left hand doesn't really know what the right hand is doing
• 15:40
  anymore. Yeah.
• 15:41
  So how do you prevent that from happening?
• 15:44
  Well, that is, again, it's that enough process to make sure that
• 15:47
  it doesn't happen. So figure out whether it's very
• 15:49
  T shaped in terms of the comms, right, so that the top layer and
• 15:53
  it goes down is creating transparency on, for example,
• 15:57
  senior leadership decisions and a place where people can go and
• 16:00
  actually see what's on the agenda for the week.
• 16:03
  It's creating weekly, monthly, quarterly, let's say all hands
• 16:08
  where people can ask their questions and hear about the
• 16:12
  state of the business. I'm always very much for a
• 16:14
  pretty transparent way of showing what the direction of
• 16:20
  the company is and why certain decisions have been made, so I
• 16:24
  think that's a good way to do it.
• 16:27
  OK, so really make it transparent to what is
• 16:29
  happening. Yeah.
• 16:31
  OK, how do you determine what is enough process?
• 16:42
  Yeah, that's a good one. I mean that's obviously not a
• 16:45
  something that's cast in stone. And for the one company that's
• 16:49
  maybe fully remote, you have to apply a lot more process to make
• 16:54
  sure that the information flows in the right way.
• 16:57
  Whereas if everybody Co locates and sits in the same room every
• 17:00
  day, there's more spontaneous knowledge sharing.
• 17:04
  So I, I think it's a, it's not a clear cut, let's say recipe, but
• 17:11
  I guess if you see that you're not moving as fast or there's
• 17:16
  lots of things that go counter intuitive or gets lost in
• 17:20
  translation between departments, you probably need to apply a
• 17:23
  little more. Yeah.
• 17:25
  So how did you do that for Trango Is Trango is not remote?
• 17:31
  No, we're not remote. We have a few remote employees,
• 17:34
  but the majority is Co located in our offices in Utrecht and
• 17:38
  people do come in several days a week.
• 17:40
  So that is is very nice. For example, we have a first
• 17:46
  Friday every first Friday of the month where we look at the
• 17:50
  previous let's say a month and what's coming for the for the
• 17:54
  next one. We do a more elaborate one on
• 17:57
  the break of the quarter. We have weekly all hand sessions
• 18:02
  that anyone can claim. So let's say we are going to do
• 18:06
  a new receipt expense claim process for example.
• 18:10
  Then finance can claim that session and stand there and
• 18:13
  explain to everyone what's going on.
• 18:14
  But it can also be what have we learned from let's say the the
• 18:20
  new AI suggestions coming into Google and how does it impact
• 18:23
  our marketing approach And then marketing claims a slot.
• 18:26
  So that's pretty much a slot that anyone can claim that has
• 18:29
  something that is insightful and interesting for the entire
• 18:32
  company to hear and learn about. OK, let's shift to music.
• 18:39
  I read a post from you about a song from Genesis, the Land of
• 18:46
  Confusion. Are we in a land of confusion
• 18:51
  with AI still and? More.
• 18:53
  Importantly, how do we get out? I would probably argue yes, so,
• 18:58
  but I'm not sure that is different from any let's say
• 19:01
  shift in, in technology. But I do feel that we're we've
• 19:07
  entered an era of unprecedented compute mostly that we haven't
• 19:11
  seen before, which can fully shake industries and put whole
• 19:18
  systems upside down again. And I think whoever was the
• 19:21
  dominant player in the last 10 to 20 years in the SAS world as
• 19:24
  we knew it is not necessarily well and best equipped for for
• 19:28
  the next decade. And so yes, we are trying to
• 19:32
  figure it out together. And actually that quote that you
• 19:34
  saw was from a speech that I did that said you don't always, you
• 19:38
  can't always, let's say, find the answering data, right?
• 19:42
  Some people just derive a lot of confidence out of hundreds of
• 19:47
  data models. But they say that the longer you
• 19:53
  torture data, it will confess to anything.
• 19:55
  You can always make the model work based on which assumptions
• 19:58
  and go in a direction, right? But there's also this other
• 20:00
  thing that. Based on experience, knowledge,
• 20:03
  customer discussions, etcetera, maybe even intuition that plays
• 20:07
  into decision making. And so in a land of confusion,
• 20:12
  spreadsheet won't always help you get out of it.
• 20:16
  And sometimes it's also I'm particularly now with AI, right?
• 20:19
  I don't even have to torture cha TPT, I have to say.
• 20:23
  That's true. But and I think it's also about
• 20:28
  placing bets, right. So now the VCs are putting money
• 20:32
  in and lots of AI initiatives and they also don't know whether
• 20:35
  those are the ones that are going to be, you know, holding
• 20:38
  the winning formula in the end. And so yes, I would say we are
• 20:43
  all still trying to figure out this new space and but we're all
• 20:46
  in the same boat and trying to figure it out together.
• 20:48
  Yeah, So what? What's talking about bets?
• 20:52
  What's the bet that Trango has taken that the other companies
• 20:58
  in the space have missed? Well, I'm not sure they would
• 21:01
  have missed it. First of all, it's kind of
• 21:03
  reimagining your entire product now on the basis of what AI can
• 21:08
  do, right? So don't build the next version
• 21:10
  of what you knew or what you had, but it's really like
• 21:12
  reimagine the value for your customer.
• 21:14
  But now in the Gen. AI space.
• 21:16
  I think secondly, internally on our processes, we have
• 21:21
  completely revamped our company values to also fit this new
• 21:26
  space. And one of them is we're dynamic
• 21:28
  and the other one is we're future proof.
• 21:31
  And that also means that we embrace working in a new way
• 21:35
  quicker than let's say that some of the other companies will do
• 21:39
  in the market, which means whatever we knew about how we
• 21:43
  did things, we can do them better.
• 21:45
  For example, we have all our customer calls, whether they're
• 21:50
  with a prospect, existing customer, whatever comes into
• 21:53
  support desk, etcetera, all going into one big system.
• 21:56
  And then we can query that for synthesis of, you know, out of
• 22:00
  this particular segment of all the deals we've lost, what was
• 22:03
  the main driver for them not choosing Trango over a
• 22:06
  competitor, for example. And that intelligence is so
• 22:10
  insightful and valuable. Yes, we had tools that had some
• 22:14
  of that before, but with AI it becomes you can, you can tie all
• 22:18
  of that together or for example, we've built a an assistant that
• 22:24
  helps us if we go to a new prospect.
• 22:26
  There's always certain things that you want to try and find
• 22:28
  out about that company before you speak to them, right?
• 22:31
  You want to understand their business model, you understand
• 22:33
  where their pain sits, etcetera. And now as assistant, you just
• 22:36
  give the name of the company and they'll do the research as we
• 22:39
  normally do across the several sources for you.
• 22:42
  And then you just get a whole brief as to, OK, who am I going
• 22:46
  to speak to? What is this about?
• 22:47
  Where is the angle that why would they speak to triangle
• 22:51
  etcetera. So I think becoming hyper agile
• 22:56
  is the one thing that we want to do in this new space that is
• 23:00
  going to hopefully set us apart from some of the other other
• 23:04
  companies is reimagine not built from what you have and going
• 23:09
  hyper agile because the cycles are only getting shorter and
• 23:12
  shorter in terms of learning. Yeah.
• 23:14
  How do, how do you do that in terms of balance with workload?
• 23:20
  Because you can keep sort of continue to, to, to work on
• 23:25
  things 24 hours if you want, but there's also something like
• 23:31
  sleep that people need and work life.
• 23:34
  Yeah, I'm a I'm a very big advocate for a good work life
• 23:40
  balance. I think switching off actually
• 23:42
  makes you more creative and more energetic as well.
• 23:47
  And I, I'm not so sure that hyper agile means doing more and
• 23:53
  spending your time less efficiently, right?
• 23:55
  It's actually about iterating faster.
• 23:58
  So it's a more about the pace than the volume necessarily.
• 24:01
  And I think actually with AI now as our our kind of external
• 24:05
  super brains, we can actually be even more efficient maybe with
• 24:09
  our time and still have 20 more learnings.
• 24:12
  OK. So one of the bigger, bigger
• 24:15
  fears of people with AI is about privacy and about biases.
• 24:22
  How do you look, how do you prevent that from happening at
• 24:26
  Triangle? Because you use external models,
• 24:29
  data models, right? Yeah.
• 24:31
  We can. We're LLM agnostic, so our
• 24:34
  product can point to various LLMS and I would say there will
• 24:39
  always be because these LLMS consume heaps and heaps and
• 24:42
  heaps of amounts of data and whatever biases in that data
• 24:45
  will also shine through in the model, right?
• 24:48
  So I don't think you can be completely, let's say, get rid
• 24:51
  of full bias. What you can still do is in your
• 24:55
  prompts as to before anything then goes out, you can still put
• 24:59
  boundaries as to what you want things to be spoken about or
• 25:04
  whether you want to use specific pronouns or if you generate
• 25:11
  images, you can also prompt as to what type of people are shown
• 25:16
  there, etcetera. So you can take exactly what
• 25:19
  comes out of the LLM or you can shape and and give boundaries to
• 25:22
  what how you then let's say, pass that information on.
• 25:27
  OK. And is does Trango do that or
• 25:30
  does the customer do that by defining tonal voice and use of
• 25:37
  emojis whatever? Yeah.
• 25:39
  So the customer can do that for their customer, right.
• 25:42
  So somebody who's a financial services company probably wants
• 25:44
  to speak very differently to their customers than someone
• 25:47
  who's an amusement park, for example.
• 25:49
  And we do have a prompt layer, but that's also more to make
• 25:53
  sure that things get escalated if something is off, right, if
• 25:57
  some, some answer isn't given or something gets stuck or there's
• 26:00
  a 22nd delay or whatever, that immediately an escalation can
• 26:04
  come in. So we do more of the security
• 26:06
  safety performance prompts. But the actual tone of voice and
• 26:10
  how they want to speak or whether they want to use a
• 26:13
  certain way of portraying, let's say their their customer base,
• 26:17
  that's obviously up to the client to to decide on.
• 26:20
  OK. So how do you then to the the
• 26:24
  protection base? We, we, we put logic in and also
• 26:29
  we learn by our own mistakes as well, because at some point we
• 26:33
  did an investigation was actually in a very different
• 26:35
  topic. It was just about what WhatsApp
• 26:36
  traffic. And then we learned there was
• 26:40
  one of our bots deployed at one of our clients that had received
• 26:45
  an incoming call from actually not a person, but a bot.
• 26:49
  And so R1 responded very nicely about, hello, I'm so and so,
• 26:55
  what's your name? Can I help you?
• 26:57
  And the other one then exchanged another pleasantry and said, hi,
• 27:01
  how can I help you? I didn't understand your
• 27:03
  question. Could you please repeat it?
• 27:05
  And that went back and forth and back and forth.
• 27:07
  And they were in this loop of, let's say, perfectly polite
• 27:10
  pleasantries. And then R1 was instructed if 25
• 27:15
  seconds into it you don't, there's something funky going
• 27:19
  on, and then they should ask the other one to please escalate.
• 27:22
  And then that didn't happen on the other side.
• 27:25
  So they just kept going. And that went on for weeks
• 27:27
  actually. So that was a big learning so
• 27:29
  that we had to put another boundary around what we thought
• 27:32
  was enough, a 25 second escalation rule, but we needed
• 27:36
  to be more specific as to what we and how and when we wanted
• 27:39
  that to trigger. Yeah, because I was.
• 27:43
  When we just told me this, I was thinking like the next step in
• 27:49
  AI will be both stocking to bolts and then it sort of
• 27:53
  becomes important to have, at least in your own bolts, good
• 28:00
  boundaries around it. Is it something that also has to
• 28:03
  be done on a more systemic level?
• 28:07
  Yeah, I think generally it's good to have checks and
• 28:10
  balances, right? And I think it also matters what
• 28:14
  the subject is. For example, particularly as now
• 28:18
  we're in the age of agentic AI, right?
• 28:20
  They don't just supply information, but they actually
• 28:22
  go do things in system. So we now have customers who
• 28:27
  let's say cancel an order and then the bolts actually cancels.
• 28:31
  So the AI agent cancels the order.
• 28:33
  And now the question is, does that then somatically trigger a
• 28:37
  refund or not? Because that's when money starts
• 28:40
  flowing. And so do you need an extra
• 28:42
  balance or if that's the automated next one, when would
• 28:46
  you, let's say police? So there needs to be a
• 28:48
  monitoring system. Suddenly the same amount goes
• 28:51
  to, you know, 25 bank accounts in this in the same 3 minutes.
• 28:56
  Should a red flag go up and say something is funky here?
• 28:59
  So where it used to be about managing people and how many
• 29:03
  tickets they resolved in the support desk, it now now becomes
• 29:06
  more of an overview of all these different ages and boats running
• 29:10
  and a monitoring, let's say dashboard as to, hey, this is a
• 29:16
  normal business, this is out of the ordinary.
• 29:18
  And what gets flagged where? And so where do I have to dig in
• 29:21
  to see if everything is running smoothly?
• 29:25
  Yeah. And as soon as it's money, bank
• 29:27
  accounts, etcetera, obviously those controls need to be a lot
• 29:30
  stricter than whether it's can I bring my dog to the festival.
• 29:35
  For these reasons, yeah, so we talked about the technical part
• 29:43
  and AI and about Triangle. I wanted to move a bit back to
• 29:49
  you as a person. What?
• 29:53
  Because we've had a number of conversations so far and what
• 29:56
  struck me really is that you, you have a passion for
• 30:02
  innovation and on the other hand, also people.
• 30:06
  But what I also found is that you really want to compliment
• 30:11
  A-Team. What?
• 30:12
  Why is that important to you? Well, I think I, I think nobody
• 30:19
  has all skill sets, right? And in order to achieve
• 30:22
  something, you need a lot of different skill sets to make
• 30:25
  that work. And particularly in the
• 30:27
  situation that I'm in. So as an external person coming
• 30:31
  into a founding team, that's even more important.
• 30:34
  Like if they were to hire Aceo with the exact same qualities
• 30:38
  that they have, then what are they solving for?
• 30:42
  And why did they even take an external person if they already
• 30:44
  have it? So I think that
• 30:46
  complementarities is, is super important.
• 30:50
  And that's also one of the reasons, you know why we said,
• 30:54
  let's do a trial period and work together to make sure that that
• 30:57
  is actually not just there in theory, but also there also
• 31:01
  there in practice. And the founders now are very
• 31:06
  much focused on products and on engineering.
• 31:09
  And so they're very much also people of detail.
• 31:13
  I'm much more of the strategic, longer broad strokes and, you
• 31:17
  know, the managing the people and so that works really well
• 31:20
  together. OK.
• 31:22
  And why is it important to you to fit in?
• 31:26
  To belong? You know, in the end, we're
• 31:30
  trying to build something that's super exciting for our
• 31:33
  customers. And let's say for me that's a
• 31:35
  true North Star is that you build something super valuable
• 31:38
  that your customers see the benefit of and are happy to work
• 31:42
  with and recommend etcetera. And I think the way to get there
• 31:47
  is, is to find that complementarity, make sure that
• 31:49
  everybody can play at their strengths, because then you move
• 31:52
  the fastest and you can probably develop that in the in the best
• 31:55
  way. And there's no point in working
• 31:58
  somewhere where you're not a fit, right?
• 32:00
  I think generally in careers, when you sometimes lose your job
• 32:05
  involuntarily, it can actually be a bit of a blessing because
• 32:08
  it's never about somebody not being good, but it's more of a
• 32:12
  mismatch in what is required at that point in time for the
• 32:15
  company and what does the the person personally thrive at.
• 32:18
  And if there's a mismatch between the two, then there's
• 32:20
  probably 2 unhappy sides, so. Can a mismatch develop over
• 32:26
  time? Uh, yes, definitely.
• 32:30
  Umm, So what makes total sense in a certain phase of the
• 32:34
  company, uh, may require, let's say somebody is ahead of, uh,
• 32:38
  marketing and then marketing between zero and then three in,
• 32:43
  let's say, if 10 is enterprise would require a different skill
• 32:47
  set, maybe then for someone who needs to bring it from three to
• 32:49
  seven and, and let alone from 7:00 to 10:00 again.
• 32:52
  So I do think that doesn't that mismatch can arise over time.
• 32:57
  OK. Does it also apply to CEOs that
• 32:59
  there's a sort of, let's say quote UN quote holding periods?
• 33:06
  I was, I'm thinking, best before.
• 33:09
  Yeah, I agree. I think first of all you see
• 33:14
  that there's only very few people who like Zuckerberg can
• 33:18
  bring it from a start up into a billion dollar company and still
• 33:22
  hold the reins. I think that's that's that's not
• 33:24
  very common. So you do see in and that's just
• 33:27
  growth phase. But I think also at some point,
• 33:30
  once you've been around the block for a number of years,
• 33:33
  you've done it 7 times. Sometimes a fresh perspective or
• 33:37
  a new look or is also so beneficial for the company.
• 33:42
  Yeah. How do you prevent, Because you
• 33:44
  told us at the beginning that you don't want to get bored and
• 33:50
  that you at some point you do get bored.
• 33:52
  How do you spot that when that happens?
• 33:56
  It's mostly about energy, I think if the, let's say, the
• 34:01
  energy that you get from your work and the energy that you put
• 34:05
  in need to be in a certain balance, right?
• 34:08
  And and there's always parts of your job that are not the most
• 34:11
  exciting bits, but hopefully it's 8020 of the stuff that does
• 34:15
  excite you. And if that balance shifts
• 34:17
  between 2080 of the 20 does excite and the 80% doesn't
• 34:22
  excite anymore, probably your outside of work energy will be a
• 34:26
  lot different as well. And so that's the point where
• 34:28
  you need to either pay attention and maybe redistribute the tasks
• 34:32
  or think about, let's say next horizons from a personal point
• 34:37
  of view, right? That's not necessarily the
• 34:38
  company side. No, but and what do you do then
• 34:44
  as a person? Well, first of all to see.
• 34:47
  Because it's sort of scary to a signal that yeah.
• 34:54
  And then B also take the take the leap to go for something
• 35:01
  about it. Yeah, to do something about it.
• 35:03
  Yeah. I mean, first one is
• 35:04
  acknowledging it and then trying to identify in which of the
• 35:07
  elements that sits. Then the second one is I think
• 35:10
  to discuss it and decide what you could change in terms of
• 35:16
  holding which part of let's say the ownership of the, let's say
• 35:20
  the, the company tasks or focus or etcetera.
• 35:24
  And it's also a discussion that I have with, you know, people on
• 35:27
  my team reporting to me because we're a small company.
• 35:31
  So it's not like there's 20 growth trajectories for everyone
• 35:35
  post the step, the threat, right?
• 35:37
  And there is a point where maybe you've outgrown your current
• 35:40
  seat. The seat above you is filled and
• 35:42
  you don't have a lateral move then yeah, the most logical move
• 35:46
  is, is, is somewhere else. And then I'll happily obviously
• 35:49
  help the person transition both in there and in the in the in
• 35:54
  the company's interests. So I think it's important to
• 35:57
  have those discussions quite openly and honestly.
• 36:00
  Yeah. And where for me that is on
• 36:03
  routine, boredom, etcetera, for somebody else's can be something
• 36:06
  completely different, right. I've mastered this skill or
• 36:09
  this, let's say job description. I now need to own more, but
• 36:15
  there might not always be more in a small company.
• 36:18
  No, in a small company the the number of things to do, number
• 36:24
  of areas to be in are still limited.
• 36:28
  It's not like your ING where the world is pretty much everywhere.
• 36:35
  Exactly. So I was wondering, yeah, you've
• 36:39
  done a lot of leadership. It's retransfer now at Trango,
• 36:45
  but also at other companies. What's 1 lesson that you've
• 36:49
  learned that you still apply today?
• 36:53
  Well, I touched upon one more generic one, I think previously.
• 36:57
  It's like for me, value should be the goal and value, I mean
• 37:01
  actually value for your customers.
• 37:03
  So not necessarily monetary value.
• 37:06
  I sometimes see people chasing like I'm starting a company
• 37:09
  because I want to be rich. I think you would have started
• 37:12
  the company because you want something in the market or you
• 37:15
  see an opportunity that isn't there yet and you you
• 37:18
  desperately want to bring that into, let's say.
• 37:22
  Yeah, you want to give a solution and yes, hopefully as a
• 37:25
  result of it earn money. Exactly, instead of making that
• 37:28
  the goal. So I would say that is 1 lesson.
• 37:31
  And maybe on a more personal level, I tend to go in
• 37:35
  solutionizing mode pretty quickly.
• 37:38
  If somebody comes to me with a problem, whatever, but sometimes
• 37:40
  that person just wants to fence and just get the emotional side
• 37:48
  of that out of the way. And so I think on a personal
• 37:52
  level, I learned to leave some space for that instead of going
• 37:56
  immediately into, OK, what can we do about it?
• 38:00
  Yeah, that's it's, it's common. It's a reflex I recognize for
• 38:04
  myself. But I always thought it was more
• 38:08
  because men tend to do this, tend to sort of run around the
• 38:14
  emotion and then focus to go straight to the solution mode.
• 38:19
  Well. Maybe.
• 38:20
  Maybe it is more predominant to men, that I don't know.
• 38:23
  I do know that I display that same behaviour.
• 38:27
  So how do you work around that? And I mean, it's, it's again,
• 38:32
  acknowledging it and being aware of it.
• 38:35
  And that comes through feedback or reflecting back on a
• 38:39
  situation where you think, you know, an hour later, a day
• 38:42
  after, oh, we could have done that in a different way and we
• 38:46
  would have been much more effective.
• 38:48
  Yeah. Yeah.
• 38:49
  Do you, you think about these situations a lot?
• 38:53
  Are you more of a defensive side?
• 38:55
  Yeah, absolutely. Because that's, I mean, it's
• 38:58
  this is you want, let's say the human interactions on the on the
• 39:05
  team to be as smooth as they can be, right?
• 39:09
  So if some of this is in the way, it'll take energy for
• 39:12
  things that you would rather have that energy spent on
• 39:15
  something else. So it's trying to understand.
• 39:18
  And every team has their own dynamics and every bilateral
• 39:22
  sort of relationship in that team has their own dynamics
• 39:24
  because people are all different personalities.
• 39:27
  So I do like actually going through some of those
• 39:30
  personality tests with the team and not to pinpoint or put
• 39:34
  people in a sort of quadrant somewhere, but for all of us to
• 39:39
  understand each other well as to where do I normally come from?
• 39:44
  Where does somebody else come from?
• 39:45
  Where are the sensitive parts for the one or the other?
• 39:48
  So you can read better how that person maybe responds to you and
• 39:53
  how you can how you can improve. Let's say if you want something
• 39:58
  done from someone, the one listens very much to rational.
• 40:03
  Explanation. The other one doesn't care at
• 40:05
  all about that, right? So and if you still want it
• 40:07
  done, then it does make sense that you know a little bit each
• 40:10
  other's manual or instructions. So yes, I do spend time on that.
• 40:15
  Yeah. Do you also coach people in your
• 40:18
  leadership team on that? Like like if you spot things
• 40:22
  like hey I see this discussion going off the rails for that's a
• 40:28
  non functional reasons. Yeah, that is definitely
• 40:31
  something that obviously as any manager you would also take a
• 40:36
  take a part in. But I also get external people
• 40:38
  sometimes to help us, let's say on more observing and then
• 40:43
  solutionizing for the team dynamics what where we might get
• 40:47
  stuck. And then they kind of also plant
• 40:50
  little seeds of, OK, how do you recognize a certain pattern of
• 40:54
  behavior, how can you change that or when do you put a stop
• 40:57
  to that, etcetera. So that's super helpful as well.
• 40:59
  Yeah. Does that also had the
• 41:04
  difference in manuals, let's say that is especially relevant if
• 41:10
  you have a diverse team, if you want to have build an inclusive
• 41:14
  business, let me put it that way, you focus particularly on
• 41:20
  also building those inclusive teams or is that something that
• 41:24
  sort of comes easy? No, I do.
• 41:28
  I find it very important. And I think the utopia is that
• 41:34
  it's happens naturally because if you just go for, let's say,
• 41:39
  the skill sets and everything that you need, you would by
• 41:41
  definition get it a diverse team.
• 41:45
  I. Also think that the reality
• 41:48
  isn't always exactly like that. So in every step of hiring,
• 41:51
  whether you put the job description out for once you do
• 41:54
  the screening interview, the actual interviews bias will will
• 41:57
  creep into it, right. So maybe you also need to be
• 42:00
  acting a little bit more conscious.
• 42:02
  I do know that at Trend now we have a really good gender
• 42:05
  balance across the board. So we, we're well above the
• 42:08
  industry trends as the overall company, but also particularly
• 42:12
  on the product engineering and data engineering type of roles.
• 42:16
  But I'd love to beat even closer to 5050.
• 42:18
  And we have on our 100% team probably 30 nationalities.
• 42:24
  And so I think we're, we're we're pretty good on, on
• 42:29
  diversity. And I see that it brings richer
• 42:33
  outcomes. Sometimes it takes a little bit
• 42:35
  more debate because sometimes people don't understand each
• 42:38
  other that well. Even the difference between the
• 42:41
  native English speakers and the non-native English speakers,
• 42:44
  right? It's you can't express to the
• 42:46
  exact same sentiment as the, as the native speaker.
• 42:50
  And then yeah, stuff might be perceived less nuanced than you
• 42:54
  meant it to be, but you just don't have the vocabulary to
• 42:57
  make it maybe more nuanced, for example.
• 42:59
  Yeah. How important is it when you're
• 43:02
  developing AI to have an inclusive team or to have a
• 43:06
  first? Story if, whether with or
• 43:08
  without AI has a, has a difference, I would say I, I
• 43:13
  think anything can benefit from it because we are selling to the
• 43:18
  entire universe and that also consists of a lot of different
• 43:21
  flavors and people from different walks of life.
• 43:24
  So the more you have, let's say on your team, the more you can
• 43:27
  also reflect to your, to your audiences.
• 43:29
  I'm not so sure that AI is a distinguishing factor there.
• 43:34
  Now while I was thinking along the lines of helping customers
• 43:39
  present bias or sporting biases earlier.
• 43:43
  Yeah, that could be, but that is not exactly the tool set that
• 43:46
  we're in, right. So I think if they felt very
• 43:49
  strongly about it, they could actually review all their data
• 43:53
  and figure out where they may or may not be biased in it.
• 43:58
  But that's not necessarily tooling that we provide.
• 44:00
  So maybe on a more informal way we would come across that team,
• 44:04
  but it's not, let's say the how our product is used or yeah.
• 44:11
  OK, so you posted on LinkedIn a while ago that you hoped to to
• 44:19
  live to tell that International Women's Day wasn't necessary
• 44:23
  anymore, that it wasn't necessary to celebrate it
• 44:27
  anymore. Why?
• 44:30
  Well, in my view, if you have a special Women's Day or the top
• 44:36
  50 women leaders or it means that that is the exception.
• 44:42
  So if it's only that one day, that's exception means and all
• 44:44
  the other day, the implication is almost that they are not
• 44:47
  women days. And if we were to be all equal,
• 44:52
  and then I would say there's no need to call it out because the,
• 44:56
  the, the, the, let's say the difference doesn't matter
• 44:59
  anymore. So as long as you feel that you
• 45:01
  need to call it out, it means there isn't a full balance yet.
• 45:05
  That's how. Yeah.
• 45:06
  So how do we get there? How do we get there?
• 45:09
  So I think we've come, we've come a long way.
• 45:12
  I do think very much improving. I also think having more role
• 45:18
  models is a good one. I also feel that we, let's say
• 45:21
  in terms of what we're seeing happening in medicine, in
• 45:25
  history, in inventions, right? Some of the truth finding
• 45:30
  exactly some of the inventions done by women were claimed by
• 45:33
  men over the past is how we position them in movies or let's
• 45:40
  say fiction, etcetera. All of these little things like
• 45:43
  starting with the with the the, the books you start to read when
• 45:47
  you first learn to read, right. So how much do we put a, a
• 45:54
  gender kind of bias in there already?
• 45:57
  And I think we, we are starting to become more, let's say,
• 46:02
  susceptible to that and more understanding of how all of
• 46:05
  these things kind of interact. I think from a personal point of
• 46:08
  view, what I can do is participate in sessions like you
• 46:11
  organize, but also in mentorship for, you know, women
• 46:17
  entrepreneurs or other women managers.
• 46:20
  But looking at, I have two daughters who are in college
• 46:23
  now. And I do think from when I was
• 46:27
  there, we've already made big, big strides from there.
• 46:33
  And so I hope, I don't know if it's that early, but by the time
• 46:37
  they have, let's say children who are in universities, maybe
• 46:42
  the Women's Day is not as required anymore.
• 46:47
  That'll be maybe another whatever 2530 years, but let's
• 46:50
  see. Yeah, what role can men play in
• 46:54
  that? Men can definitely play a role
• 46:57
  as well, right? It is also, again, knowing
• 47:03
  you're there's this thing called identification bias.
• 47:07
  So if you speak to someone who is a little bit like you and you
• 47:10
  recognize and you understand each other really well, that you
• 47:14
  feel that that person is, is equipped to work with you or
• 47:18
  maybe a very good person to add to your team.
• 47:21
  But if you understand that complementarity is maybe the
• 47:25
  better, let's say, elements to look for, it means that that
• 47:31
  identification bias is actually a negative thing in terms of
• 47:34
  finding that complementarity. So I think it's all about
• 47:37
  consciousness and, and really knowing yourself and, and
• 47:40
  understanding what that looks like and also realizing that we
• 47:44
  will all have some bias, right? You are brought up in a
• 47:46
  different way than I was brought up and you have had different
• 47:51
  expectations, maybe because you were a man from your parents and
• 47:54
  maybe mine were different because I was a daughter.
• 47:56
  So acknowledging that and then defining your own path through
• 48:00
  that, I would say. Yeah, yeah.
• 48:03
  And I think to your point, I spoke to a guest recently who
• 48:09
  called empathy a double edged sword in, in the sense that
• 48:14
  empathy, how often is really seen as good from a manager,
• 48:20
  that he or she has empathy, but that it's also callers, let's
• 48:24
  say the people you work with, the people you promote, the
• 48:29
  people you like, the people you hire in that sense.
• 48:33
  So that, yeah, that sounds really familiar.
• 48:36
  And I'm starting to see that as well.
• 48:39
  I have to say it. And I would say in general in
• 48:42
  business and particularly with performance review, etcetera,
• 48:45
  it's very hard to make the distinction between liking
• 48:47
  someone and thinking someone is very good for the role, right?
• 48:51
  You can like someone very much and still fire them or you can
• 48:55
  dislike them but really see how they're amazing in the role that
• 48:58
  they're in. And trying to separate those two
• 49:00
  is is is difficult, but I think that's what's required to do an
• 49:04
  objective performance review. Marie Lee, thank you very much.
• 49:10
  Thank you for. The for the interview, if people
• 49:12
  want to learn more about you and about Tango, where would you
• 49:17
  like to send them? Obviously to tango.com and to my
• 49:22
  LinkedIn. I think you will maybe also add
• 49:25
  some to. We'll include a link in the show
• 49:29
  notes. Yeah.
• 49:30
  And then I have an AI Voice Assistant where you can call me
• 49:35
  and then you'll speak to my alter ego and you can ask
• 49:37
  anything about Trango and what we might deliver for your
• 49:40
  company. OK, well, we'll include.
• 49:43
  That's a phone number I think. Yeah, it is a phone number.
• 49:46
  OK, I'll send it over to you. Yeah, you now I remember that
• 49:49
  you don't give out your number except for this one.
• 49:53
  Exactly. OK, Maggie Etalph, thank you
• 49:57
  very much for an amazing interview.
• 50:00
  Really enjoyed this listeners. Thank you very much for
• 50:02
  listening. And until the next episode, keep
• 50:06
  being awesome. Well, that's it for episode 116
• 50:15
  of Women Disrupting Tech with Marily at Toft Puller on how to
• 50:19
  navigate growth in the age of AI.
• 50:23
  Marili. And I hope that you found our
• 50:24
  conversation as empowering and engaging as we intended it to
• 50:28
  be. And if you have any questions
• 50:31
  for Marili, please leave them in the comments on Spotify, Apple
• 50:35
  or YouTube. I'll make sure you receive an
• 50:37
  answer. We want to connect with Marie
• 50:40
  Lee or learn more about Trango. Please check out the links in
• 50:43
  the show notes. That's where you will also find
• 50:46
  her phone number and a link to the music video of Land of
• 50:51
  Confusion by Genesis. As usual, you can find the show
• 50:55
  notes at Women Disrupting Tech dot Blog and the link is also in
• 50:59
  the description. If you've enjoyed this episode,
• 51:02
  please consider leaving a rating or review on Spotify or Apple.
• 51:07
  Leaving a rating or a review helps new listeners find a
• 51:10
  podcast and supports my guest by giving them a bigger platform.
• 51:15
  If you're a regular listener, please make sure to follow Wim
• 51:18
  Disrupting Tech on Spotify, Apple, Amazon or YouTube.
• 51:22
  Following Wim Disrupting Tech is a simple 0 cost way to support
• 51:26
  Marley and my older guests in making inclusion in tech to new
• 51:30
  normal by the end of 2032. For more about Women Disrupting
• 51:35
  Tech please follow the podcast on social media you can find us
• 51:39
  everywhere as at Women Disrupting Tech.
• 51:42
  Women Disrupting Tech is also on Sub Stack.
• 51:45
  On Sub Stack you will find in depth stories and behind the
• 51:48
  scenes insights plus trends that I've spotted across 100 plus
• 51:53
  conversations so far. So don't forget to subscribe to
• 51:56
  Women Disrupting Tech on Sub Stack.
• 51:59
  Finally, if you'd like to connect with me directly, please
• 52:03
  reach out via e-mail, Dear Kyon at Women Disrupting Techno blog
• 52:07
  or connect with me on LinkedIn. Just search for my name, dear
• 52:11
  Kyon, and you'll find me. With that, I wish you a
• 52:15
  wonderful day. I hope to see you in the next
• 52:17
  episode. And until then, as always, keep
• 52:21
  being awesome.