EP. 265 AI for Happier Teams: L10's Real-Time Feedback Revolution

Is waiting months for employee feedback a thing of the past? Dive into the future of work with Shane McAllister and L10 founders, Zyrian Chung and Keith Chan. Discover how their AI-powered platform collects daily insights to transform workplace culture, empower managers, and address issues like silent quitting before they escalate. Learn how real-time feedback and AI can create more engaged and productive teams, and get a peek into the tech (including MongoDB) that makes it possible. Stop guessing, start understanding your team today!

TAGs:

#EmployeeEngagement, #AIinHR, #WorkplaceCulture, #FutureofWork, #EmployeeFeedback, #RealTimeInsights, #TeamManagement, #Leadership, #HRTechnology, #MongoDB, #L10, #ZyrianChung, #KeithChan, #SilentQuitting, #EmployeeMorale, #TeamProductivity, #PerformanceManagement, #AITools, #EmployeeSurveys, #PeopleAnalytics

Episode Transcript

(00:06):
You're very welcome to another MongoDB podcast.
I'm Shane McAllister, and today we're diving into the future of
work and how AI is helping us rethink the way we engage with
our team's. Annual surveys for employees are
very outdated. Waiting months to hear how your
team is feeling is a thing of the past.
My guests today are Zirian and Keith, founders of L10A platform

(00:30):
that's using AI to collect dailyfeedback and deliver real time
insights that helps managers before problems escalate.
We'll talk about how L 10 is transforming workplace culture,
why silent quitting is a symptomof deeper issues, and how AI,
when done right, can empower both managers and employees as

(00:52):
well too. And we'll get a a little bit of
an understanding of how Mongo DBsupports their real time
infrastructure at L 10. So without further ado, let's
get on with it. Zirian and Keith, you're very
welcome to the MongoDB podcast. How are you?
It's been great. Thank you for having us.
It's great. And you're both joining us from

(01:13):
Hong Kong. So I do appreciate you joining
me. What is late in your day?
And for our regular viewers, this is a much earlier episode
than I usually do in the podcast, but hopefully we'll
garner a a different type of audience than we we've had
before because we're in a much more kind of AIPAC friendly time
zone. So I always love to start the

(01:33):
show with a kind of understanding of kind of how our
guests got here, their career path to date.
And it's brilliant to have two founders on board.
But if I could delve a little bit, you know, starting with you
first Syrian, what will you know?
What did you study? What was your career?
What led you to founding L10? That that's a good question

(01:54):
because that's usually what we share with our friends.
So maybe I should start with howwe met.
So we are both from a very smallisland in Hong Kong called Cheng
E. So that's why you can imagine we
study middle school, high schooltogether.
We study university together. We study computer science at
University of Hong Kong. So I would say that oh, and also
when we sit, when we're in high school, we sit together.

(02:15):
So the teacher put us together because I'm a bit more
extroverted and keep this bit introverted.
So it's kind of like, you know, the teacher love to put like
different personality people together.
So we've been brilliant. So you've got rid of one of my
questions I usually ask Co founders.
It's like, where did you meet? You guys have been together all
the way up, right? Yeah, it's been a very, yeah,

(02:36):
it's been a very long time. And as we said, we studied
computer science together. So we've been doing a lot of
different projects, obviously our final year project and so
on. But but then because of that, we
went into different paths. So Kiev has been in tech like in
company like big cop like HSBC or startup like Google van and

(02:57):
I've been more I I've done a fewEA programming as well in
London. But later on I did a lot more in
consultancy and design. OK, OK.
And Keith, obviously you know, the we heard about the
beginnings there and your pathway then through and and
when did you both decide to do something together?

(03:18):
Yeah, I think it's maybe users of goals and basically we have
different career path before andfor me, I typical software
engineer and become a tablet andin in the software company or
stuck up. And we found that we can do
something for manager of the themiddle management to help them

(03:40):
to kind of empower the team and build the better culture in the
company. So we.
I think keep is slightly humble.So he joined a start up when
it's 100 people and they call Google fan and they went IPO in
2022. So keep experience from 100

(04:02):
people all the way to like a IPOcompany.
So he got a lot of stress because he used to manage like a
small team later on, you know, manage a bigger, bigger team as
a tech lead. And later on he realized, oh, I
really have to not only look at the skills of the people, but
also the culture because we're areally good friend.
Back then, I was in Berlin, I working for a company called

(04:23):
Bane and company. And I was apart from my client
work, I leading a really small team.
We call it culture chapter internally basically is
innovating the culture for internal consultant within Bane
EMEA for the innovation team. So what I experienced is

(04:43):
probably what key experience as well, which is a lot of time.
We spend so much time on performance, on doing our work,
but later on realise the biggestproblem of why a team is not
efficient is because the culture.
We also shared that as well in the phone call saying that for
for tech lead or for manager, your first priority should be
make sure you're happy and performing well.

(05:05):
But a lot of time because our KPI is not is about revenue, is
about sales, is about between and outcomes.
And for him, it's all silly. During the IPO, there are a lot
of other KPI come in like investor expectation and so on.
So we realized why as a manager when we managed a lot of people,
but we don't put their power test #1 and make us really

(05:27):
struggle. And because of that reason, quit
Keith, quit Google Van right after IPO.
And I decided to quit together to found this company with him.
And that's roughly a few years ago.
Excellent, excellent. So it was really borne out by
world experience, the, you know,the companies that you both had

(05:47):
been in, the companies that you had both worked for, you had
seen essentially, you know, thisproblem or there's pattern that,
you know, they don't check into regularly.
Other things become more prominent than employee
satisfaction or welfare or thosesort of things.
So you founded L10 when 2023 then was it or or roughly?

(06:11):
Roughly that. So we decided it started our
company late 2022, but we reallystarted working full time.
I think during our time in Chilein 2023.
I think a lot, a lot of foundersdid that as well because we've
been working for 15 years in Europe and Asia.
So we decided to find a space that we are a bit away from our

(06:31):
girlfriends and our friends. I'll be more, you know.
You don't get some work. Like all the story you're trying
to hide somewhere to build your product, taxi product.
So this is what we did before. We never, we said that we've
never been to Latin America. Why don't we find somewhere we
stay in a very small apartment trying to build something out of
nothing so that that what we did.

(06:52):
So we went to Chile in 2023, I think around March and that's
where we started our product. OK, wow.
I mean, like we all too often hear of startups and they're,
you know, young, early 20 somethings founders who can
exist on a diet of pizza and coke and live together,
etcetera. But you guys were a bit older

(07:12):
than that and a bit wiser. But you still said right, if
we're going to do this, we kind of need to do this 24/7.
Let's abstract ourselves away from the the interruptions and
the distractions and go somewhere to do that.
I think that's a that superb idea.
L10 What? What's the meaning behind the
name of the company then? That's also an an other story.

(07:34):
So now our first name we have called team culture dot AI
because we believe that culture of a team is the most important
part. So if you search TL10, you will
still see that we incorporate a name as Team dot AI.
The problem is a lot of manager,they think that culture is not
my first party, but people are. That's a really interesting

(07:57):
concept that people think that culture not equivalent to people
and they think that as a manager, culture is tech.
Chao is my leadership. But as we mentioned before, our
main goal is to help every single manager, even the mid
level or high level, they shouldfocus on their people.
So we like this brain name is really hard to sell.

(08:17):
And at that time we're using a system called EOS.
It's basically a basically, it'svery simple.
It's not really a system. Basically every week we should
talk about our most important topic of our startup.
And in this system, we call the most important meeting called
L10 Level 10 meeting. OK, meeting was talk about
everything we think is importantfor our company.

(08:38):
And at that time we Highland High, we hire some freelance
workers and realize we start to not focusing on the because we
are founders. Well, the main goal is to make
the product work. But later on we realise even as
founders, we prioritise our people.
And we say that at our L 10 meeting, we have to prioritise
our people. And it makes us feel like that
maybe for other company at theirL10, they shouldn't, our

(09:00):
business wins. Like what we did before is also
talking about the people. And that's how we think that
maybe we should use a name that is not about culture, but it's
more generic. And later on, we realized this
name performed better because people think that what's L10?
And we start to explain this story saying that people should
be your first priority in your level 10 meeting.

(09:21):
And they start to get it, yes, that that's the way we should be
doing. And we realized it's a lot
better than our original name team culture in selling.
But our goal is basically still the same.
And that's why we haven't changed our legal name.
It's still team culture dot AI because our heart, we still
believe that we should keep the same mission, which is what
everyone said, right? You should never change the

(09:43):
mission from the moment you started your company.
So that's why we still keep up incorporate name team culture
dot AI that we sell our product we use in L10.
I love it. That's a great story behind it.
So level 10, it should be there at those, your culture should be
discussed at those, those sort of high level meetings.
So that's that's brilliant, I think for anybody.

(10:05):
And obviously we have a wide variety of people joining us on
our live streams. But for anybody working in a
company they're very familiar with, maybe twice a year, if
they're very lucky, maybe every quarter.
And so if they're not so lucky, just once a year, right, they
they have kind of these internalsurveys which are generally
linked to their own performance and growth or promotions or

(10:28):
salary increases et cetera as well too.
And I suppose the key thing there is that as we touched down
on the interim, you know that that system is kind of broken
and as you said as to as the experiences that you both had
working in other companies. Can you expand a little bit
maybe on on how that model doesn't work anymore for for

(10:48):
most companies? So that that's a really good
question because we didn't know before that as well.
So when you start a start up, the first thing to do is to
interview people to ask them what their problems are.
And because we're both come fromcomputer science background, so
we decided to go after our friends and they're all
engineers right now. They're more like engineering

(11:09):
lead or some some of them becomeACT or a big company as well.
So and to them saying that as a tech lead or as ACTO, what's
your biggest problem are when you manage your people and they
all say that HR actually give them a lot of data at tech leads
telling you how your team performed.
The problem is HR only tell themonce a year because that is the

(11:29):
only data point. Yeah, and I think a lot of you
are familiar as well, right. They call them either engagement
survey, satisfaction survey, some of them even call them
performance survey and so on. So a lot of companies will be
once a year, but that's already really lucky.
You know, in Hong Kong, Korea, which is more user located,
they, they some of them do it like only once every two to

(11:50):
three years. So you will only hear from your
leadership or HR saying that forlast two years we found these
problems and they would be like,oh, it's last two years, but our
company is already completely different.
What should we do right now? And the responsibility from the
hedgehog become from the team, team lead.

(12:11):
But for you, like you never really get the feedback from
your team apart from one-on-one.And for one-on-one, it's mainly
like, you know, identify feedback, we call them, which is
I can't give you enormous feedback as an employee or as a
as a I will report sometime. I just think will we right,
especially right now in Europe, a lot of people worry about
diversity issues affecting theirworkspace, especially in Ted,

(12:35):
like a lot of women engineer that were interviewed, they
actually we said about the situation because it's very,
very hard. One of our users like UK, they,
they're like having a race or 12:50, which is one female to 10
engineer in the same team. And it's really hard for them to
become really themselves to workvery hard.

(12:56):
But the company only collect a bit once a year and that's very
hard. And because of that very reason,
we decided to find a solution that maybe we can help them to
get them more feedback more frequently, but not only once a
year like most company in in theworld.
OK, OK. And I mentioned the word kind of
silent quitting on the introduction, which was born out

(13:17):
of the the conversation you and I, Azerian had, you know,
prepping for this, the silent quitting or the low engagement,
you know, and if they're only doing it once a year or, or even
worse, companies are completely missing out on that kind of
feedback. And and that's what's happening,
right? So this is something that we

(13:38):
realize in a lot of countries. I tell you something very
interesting that we learned fromthe last half a year we started.
We also incorporate in South Korea and some of it's like a
lot of European countries that you can't just fire your low
performance and a lot of time inAsia or in the US you may heard
about that as well. If you find your team have some

(13:59):
high performance, you try every way to keep them, you try to
promote them and so on, give them a bonus.
But a lot of time for low performers, they don't have a
lot of help. So a lot of company kind of
annoy them. But you know that you know that
if you're engineer, you know that there's no such thing as
low performer for your whole life.
You're low performer because your empowerment doesn't suit

(14:20):
you, because the task you're working on is not very exciting.
It's different from what you expected before you join a
company and so on. But in Korea and in Europe, you
can't. The company just can't fire this
low performer and this low performer.
They think that if my company ignore my feedback, what I'm
going to do is I'm just not going to work very hard.
And we all experienced that before.

(14:40):
It's like if I have a lot of feedback, my company don't
listen to me. Maybe I just don't work hard.
I just become so-called a silentcreator.
I just do my day-to-day work. I don't innovate, I don't do
extra miles. And that's not the employee
suffering because the employee can also find another job, spend
more time with the family and soon.
But the company losing a lot because you're supposed to have

(15:02):
a very hard work and working foryou or as a team manager, you're
supposed to work very, very collaboratively.
But now there's some people, they never respond to Slack
messages, they never learn more news from the company.
And that's become the biggest loss for the company.
And we realized from a lot of research, like McKinsey and so
on, they realized silence creator being the biggest factor

(15:23):
for the company to lose on profit and to become the biggest
problem for the team. Because if you have one quick
silence creator, you started to see that, oh, maybe I don't have
to respond to team messages anymore.
Maybe I shouldn't team members anymore and it's become a very
big thing. So this is the main problem we

(15:44):
want to solve. We want to not only help the
manager to find out who the low performance are, but our goal is
to help these people, the peoplewho don't like to to work in
this company to have better lifein that company.
Because we did. We believe that everyone should
kind of like contribute to the work and they shouldn't just

(16:04):
hate the jobs. Yeah, I think it's it's a great,
you know what I mean? It's a great goal to have and,
and how much like I know you found it in 2023 when we were
coming out the tail end of COVID, but obviously, you know,
that would have had an effect asI you read a lot on people's
morale and companies morale and you know, remote working for all

(16:26):
its benefits as well too, also has its downsides of, you know,
lack of interaction, not building team culture, etcetera.
How did that play into this scenario of kind of not having
engaged employees or, or did it or did that help things in terms
of kind of companies understanding, you know, that
this was key and crucial becauseemployees are the company's best

(16:48):
and biggest resource, right? It's probably their biggest
expense maybe as well too. And, and you need to have the
right engaged employees on boardand, and kind of do everything
you can to make sure that they're comfortable and happy,
right and performing obviously so.
This is some this is another thing that Keith and I debate
quite heavily these days as well, because for the last 10

(17:09):
years I've been in consultanciesand for consultant, a lot of
them are very extroverted. We're trained to be a bit more
awful spoken. So when I when I see a problem,
I have to say it, it's like the company trained us to do that.
There are a lot training programto do that.
But remind me, when I was an engineer myself, it's not the
case. A lot of time engineer.

(17:31):
When I was an engineer, I didn'treally spoke out a lot.
When I see some problem, I wouldn't say it and when manager
come to me, I wouldn't do that as well.
I was keep it to myself and realize in our research when we
talk to engineers, designers andall project managers as well,
they are not used to just in a town hall with 1000 people tell
their leadership, what's the problem now, even in one-on-one

(17:53):
they don't do that as well. And we realise they appreciate a
lot of other method to provide feedback.
Like one company we interviewed,they have like a like a little
sue spot and you can put like enormous feedback in it and
realize the energy, the energy is a lot higher when they have
that and a lot, I think that if I have that, a lot of people are

(18:13):
going to complain a lot, but actually not I, we believe that
maybe we are Chinese as well. We believe that people are nice
when they're born. So a lot of time a lot of
people, even though they want tocomplain, but the intentions are
normally good because if you don't want to complain anymore,
likely that you can't save the person anymore.

(18:34):
So when they still want to complain that they still have a
passion, they want to change thecompany.
So that's how they have a suit spot.
They receive a lot of negative feedback, but they realized this
negative feedback actually solvable.
A lot of simple as well. Like I don't mind to work over
time a lot when deadline hits, but when it's not when there's

(18:55):
no deadline, maybe allow us to do more innovation to really
help the company to become to change it, like using more new
technology to save for time and so on.
It will keep you, you observe that as well.
So that's why I realized one of very good way to to address the
silence creator problem is just to create a space for them to

(19:15):
share the feedback. OK, that's for the introverted
ones. Yeah.
No, I hear you. I mean, I think that a lot of
people have, you know, who seem to be outspoken, maybe just have
the best interests of their colleagues and the company at
heart and they just want to makethings better.
And they're very kind of invested in, you know, their

(19:37):
career and their role and their company.
So I do like that. I know we're going to have a
demo later on of the L10 platform where a lot of this
will make a ton of sense. But let's talk a little bit
about L10 and how this system works to capture that sort of
insight and to surface those insights from employee feedback.

(19:59):
And tell me a little bit about that mechanism.
We've talked about how. Infrequent normal companies do
that. How frequently does L10 get
employee feedback? So right now what we do, we do
it. We collect 5 feedback every
single week. We call them the weekly check

(20:19):
insurance. So normally the employee or the
team member will spend 5 minutesto answer 5 questions.
Most of them are open-ended questions because that's allow
them to say whatever they want. And with weekly 5 feedback, we
collect around 150 to 200 checkpoints of every single
person. And these checkpoints are either

(20:40):
anonymous or identified. And enormous feedback mostly is
about diversity problems, psych,psychological safety.
A lot of that is about leadership feedback because
these supposed to be protected. So we don't let our identity
information and they can just click a link, everyone will have

(21:00):
the same link to answer these enormous feedback.
Obviously in order for the manager to help individual
person, we also have some identified feedback where you
have to share identity with yourown manager.
For example, like what have you achieved this week, which who
who in your team helped you before?
So it's a bit like a mix of cellreflection and 360 review in a

(21:24):
lot of company and you imagine we our goal is to try to replace
three annual things that a lot of company doing. 1 is the
annual survey is a bit about engagement or diversity and this
not the company would do 50 to 100 question every year while 30
minutes a year. Second is 360 review in order to

(21:47):
save time, a lot of company would only ask you to give
feedback to five people, which is very biased and kind of picky
about who to give feedback as well.
But for us, we're trying to encourage people to give
feedback to their colleagues whenever they feel like.
And lastly is the weekly check in on cell reflection.
That's something we only realized from last year.
We realized a lot of manager, especially when we were a new

(22:09):
manager like 10 years ago. What we do is, oh, when I walk
in the one-on-one, tell me aboutwhat you've done this week.
And the whole thing is becoming a reporting one-on-one instead
of like A and a lot of they don't know how to fix it.
A lot of them, they're new managers.
They were like, isn't one-on-onesupposed to be my, my diary
report, tell me everything they've done this week.

(22:31):
But that's really not, not the not the good way.
Because in a one-on-one, the time is value available.
The reporting can be done advance in advance.
So what we do is basically want to move this reporting system to
like the weekly five question check in.
So in the one, it's really all about coaching, It's really all
about as an employee, I can be emotional with, I can be

(22:53):
vulnerable with my manager. They can share whatever they
feel like. So this is what we do.
So we actually collect a lot more feedback than only the
survey, the traditional survey, but also we collect other types
of feedback in a weekly setting.OK, I'm smiling because I'm I'm
very familiar with this. I'm very familiar with all the

(23:14):
the negative things that you say, you know, it's not frequent
enough. It's the one on ones is
reporting all of that. I'm not saying current company
accepted and all of that good stuff.
I think all all companies can dothis better.
And you know, I really love the approach that yourself and Keith
have taken towards this. So if how do you make the

(23:35):
employee fill this in? Is it just part of their daily
routine? Is it integrated with other
products like Slack, etcetera? What prompts them to do this as
frequently as you need it? So maybe I asked you to make a
guess. So do you know what's the
biggest reason an employee don'tanswer survey?
Or what's the biggest reason whythey just randomly answer a

(23:56):
survey if they? Put me on the spot, Syrian, I
would say, because they don't see, they've done it a few times
and they don't see any follow upor outcomes from doing it and
then they just get fed up doing them.
Would that be on the right track?
Exactly the reason I think you've done some research or you
just you just knew on your heart.
So yeah, it's not on me now, I'mjust given a general answer.

(24:22):
Yeah, but this is like the biggest answer in the last 20
years. The reason never changed for the
last 20 years. The biggest reason why employee
don't want to answer surveys because they think that it's
useless for them to answer because company not going to
change it anyways. And if you are, if you I think
so and you now you manage peopleand so on, you realise the
company actually want to change it.

(24:43):
A lot of leadership, they put that in there, even shareholder
report for the listed company. But the problem is they don't
know how to do it. And imagine that if you only
collect feedback once a year andthen you want to change it, what
you're going to fix is actually yesterday problem.
It's not today problem. And on point of view, they don't
see that. They don't know the company's

(25:04):
actually changing it. But one year later, maybe you
ought to quit the company and soon.
But what we do is basically we change it to deliver the
feedback to the leadership with deliver to the middle
management. And as a manager like myself, I
can change the the employee experience on my team.
Like when I hear that today, I can do something today and

(25:26):
feedback in real time like just now reset weekly and some user
using our platform even on a daily basis just now later I can
show you the demo, which is we set it as a daily basis.
We ask one question a day and sofive question every week that
you can really make changes every single day.
And we realize when the employeesee changes, when they see that
the company or the manager actually adopt what they

(25:49):
complain about, what they reallywant the company to change, they
will answer more feedback. Does it make sense when respond
weight go the way respond weightgoes up as they see the problem
they propose solved? This is what we find out why the
respond weight of our user usingour platform is higher than a

(26:10):
lot of annual survey platform because they actually see the
problem solved. Yeah, I think that's great.
And then look at yeah. And I think it will probably
engender even more trust in the platform.
If the feedback you're giving isbeing actions and stuff is
happening and that's improving, then you kind of go and look,
they're listening, it's working.I'm going to participate even

(26:32):
more, I suppose. Is there any concern from
employees about what they're being asked to share might be
used against them or anything like that if they're trying to
be super honest, understand the platform?
So this is a problem we observedbefore.
So people trying to be like protecting themselves, everyone

(26:54):
would do that. So let's let's say if someone
find that it's for me then does it affect my promotion or does
it of? Course, that's a natural
concern, right? Yeah, yeah.
But we realized a lot of problems that company facing is
actually every employee would complain a lot.
Like I told you about the diversity example, sometimes

(27:15):
about burnout and so on. We realized the employee
wouldn't tell 100% of the truth.So they wouldn't say that, oh,
this week I work until like 11 AM, 11:00 PM every every night.
But they would say that, oh, there would be a bit more mild,
like, oh, we experienced a lot of burnout and overwork, but our
AI can recognize these themes, You know, right now I, I heard,

(27:38):
I, I knew that a lot of people using MIMO DBS are very familiar
with AI. So it just, we're using our AI
agent to find out the themes andrealise that their 100% feedback
and they just have to share somelike general observation.
That's the first thing. But the second thing is what we
try to be transparent. So as an employee, they can go

(28:00):
into our platform, see exactly what the manager is seeing and.
OK, OK. This feature we see two effects.
One is the employees are a lot happier because they know that
01 share is actually protected. Plus another problem is a lot of
enterprises, they don't like it.So they'll be like, oh, but we

(28:22):
still want to secretly find out who they are.
And yeah, I think you know that as well.
Like a lot of e-mail, our colleagues, a lot of people
think that e-mail completely protected because they actually
know a lot of culprit. They would go into e-mail to
find out like what kind of e-mail he's sent before or even
go into Slack so they can find out your messages as well.

(28:43):
But because of that, they would try to be like, oh, probably I
should use something else. So that's why in Asia like Hong
Kong, Japan and Korea that a lotof people, they don't use e-mail
to communicate on workplaces. They will use WhatsApp, they
will use, OK, all work, all right, that's guaranteed.
The company cannot find it out because they use it as a number.

(29:05):
So that's in our case, because alot of user base in Asia.
So we were like, what if we justmake it like 100% transfer?
So as an employee you can even log into a platform and so on.
They can see whatever. So you can see what the manager
sees, So what the tool is surfacing up on their behalf,
but anonymized and things like that.
Exactly, this is always trying to attempt to solve this

(29:25):
problem, but even though they don't, we think that and also
another thing we do is we don't do one to five.
So right now most employee employee survey they ask you a
question like from 1:00 to 5:00.How much do you think that you,
the leadership helping you to doyour job?
And because one to five, a lot of employee would be like, oh, I

(29:47):
can do like four or five or two.But from a leadership point of
view, because when, when I was aconsultant, we realized a lot of
company, they don't know the reason behind.
So let's say if you have a scoresaying that your company
psychological safety is only from 2 out of 10, but you don't
know the reason behind, you don't know what to do with it
and you don't know how to solve the problem With our platform,

(30:07):
we, we really want to embrace open-ended question so the
employee can actually answer like very long answer.
And we have a feature that I'm going to show you later as well.
We can help them to change or weface the sentences so that the
manager wouldn't know like who wrote this feedback and so on,

(30:27):
which is very small feature thatinspired by one of our user, but
we realize. It makes sense though, because
if you're if you know the individuals and you're presented
with some content from each of them on your team, yeah, you
know them well, you can usually guess who wrote what you know.
So even if it is a anonymous. Story how we do that in the
demo. OK, perfect, perfect.

(30:50):
And I suppose, look, obviously I've done a lot of the live
streams when we're talking aboutAI and the common question that
comes up is AI is going to replace jobs, etcetera,
etcetera. But it's really good to see L10
using AI for employee and empowerment etcetera.
I'd like to get a little technical if I could.

(31:11):
And perhaps Keith for for you iskind of how, what's the
architecture underneath L10? Where does Mongo DB fit in?
How does AI fit in, what models are you using etcetera.
Any give us a high level of that.
I know we'll see the product in action in a while, but how does
that work? OK, basically architecture based
in Pool for for ISP basically isAAI agent system that you can

(31:36):
imagine there's a back end serving some API for the front
end core. And then in the back end there's
a lot of thing happens like theywill trigger some agentic
workflow that process the data and also answer the users
inquiry. And basically in the past we are

(31:57):
trying different kind of database like the rational
database honestly, the RDS for me or PWS.
And you know, that's the difference between the RDS and
the Mongo DB because Mongo DB isa flexible data model database.
And we found that basically Mongo DB it suits us more

(32:20):
because as a start up we need toreact fast to maybe the user
request or the feature changes. With a flexible data model
provided by Mongo DB, it's very easy for us to support the new
features comparing to the past. We need a very rigid database

(32:45):
migration process if you're using the direction of database.
And another things I think the mongo DB is good for us is they
basically can store every type of data in one data model.
Like for data agenting model, webasically need some better

(33:08):
embeddings and Mongo DB basically can store the normal
data and as well as the red database into the same model.
It's helped really help us to enhance the user experience.
Like we can build a search function on top of that.
And Mongo DB basically also offer a very good search

(33:33):
functionality out-of-the-box. Like I can give you an example.
Because for the AI application, sometimes I, the user will type
into the AI check board and thenask us the maybe getting the
user information or getting someinformation from the database.
And sometimes it will have some typo in the works.

(33:54):
And mostly basically there is a fussy search function
out-of-the-box. And it basically how it's very
easy to identify which record inthe database and you can
retrieve it very easily. And also the hybrid search and
the vector search from mongo DB basically help us to not only

(34:19):
search with the factor, we can also search with some semantics
or some keyword search. Yeah, I think I have to mention
about the out of the boss feature because as a startup, I
probably not only startup in bigcorporate as well.
A lot of time we don't want to do innovation and we don't want

(34:40):
to spend a lot of resources justto build a feature out, want to
test it out. But you know that when I used to
be a designer, so I know that a lot of time we just trying to
test with like fake my prototypeand so on.
But it's actually not letting the actual user feedback.
So what we like about Mongo DB is there are a lot of out of the
boss solution. So in the beginning we want to
realize, oh, this feature could be very useful for our users.

(35:03):
Let's just instead of just fake my prototype for like 2 months,
why don't we just build it out with the out-of-the-box Mongo DB
feature that adopt AI model. And that's what we want to build
a feature in a week or two. And then we can get like real
time feedback, actual feedback from the user before we make it
more complicated. And the second part is actually
less about engineering, but maybe about that as well.

(35:26):
A lot of user are from engineering teams because ours
tool is really, really useful for engineers to share the
feedback. A lot of them did not.
I was looking as to say before and before we before you use
Mongo DB, A lot of people don't know how architecture like they
don't trust that we store data very safely.
And we're trying to be scale scalable as well.

(35:47):
So when when we tell them now, Oh, remember, we share with a
listed company in Taiwan, they started use of service and it's
the AI team using a product and the reason they trust because
they know that, oh, we're using MongoDB now and we share about
technical stuff and so on. And that's really not only
helping us to develop, but also help us with the branding that

(36:11):
they trust that you're serious startup.
You're not a startup that couldn't close in two years, you
know, and you know, that's on aswell.
Mow DVD is not the cheapest software in the world.
So when you use it, a lot of people will just, oh, you're
actually being serious about your product.
You're not for the future. Yeah.
So that's that's obviously not entirely about engineering, but

(36:33):
also about branding and business.
Excellent. Well, listen, I, I think that
those answers particularly from you Keith over and and you know
the what you added at the end Siri and over the last 6 or 7
minutes I think are going to have to cut that piece of audio
and send it upwards inside MongoDB.
Great pitch for Mongo DB. You covered everything like the
flexible data, all the data types, the fact that we have

(36:56):
your data and your vectors stored together, the fact that
you've got search and fuzzy search built in, you know,
hybrid search as well, too. That's a great pitch.
And I don't have to do anything.You did it all for me.
Yeah, we. We've been a big fan of Mongo DB
for basically it really help us,yeah.
Had you used Mongo DB before youformed LL 10?

(37:18):
Had you used it in other companies?
No, but I I used mongo DB for maybe for some prototyping.
OK, OK. Yeah.
Excellent. Well, listen, no, it's, it's
great to hear, you know that it really helped you and you talked
a little bit about scale Keith there when you were chatting.

(37:38):
How does the you know what? How does Mongo DB help support
that in terms of the scale that you might need as you roll out
with new clients? Oh, basically I like the two
links from mongo DB. It's very easy for me to attract
the performance of the query or they provide a kind of, I think

(38:03):
quite good monitoring features provided by the Mongo DB.
And I, I use most is basically generate Gen.
AI features to generate the aggregation pipeline and
basically help us to scale because sometimes we we need to

(38:26):
have a very performance base or we concern the performance of
how we carry data out actually kind of features help us a lot.
OK, excellent. And look, I know that people are
coming in on the on the chat as well too.
It's great to see everybody saying hello.
Please, if you missed the beginning, if you've any

(38:47):
questions for Xurian or Keith, just drop them in there.
You probably will once we get tohave a quick look at the demo
before we do that. Just in terms of the AI,
obviously MongoDB is a vector store, but you know, we don't do
the embeddings or anything on our side.
We don't have LLM. So whose products are you using
on that side of things then? You mean the embedding part?

(39:11):
Basically we send LM like code here like we we use that API to
get embedding and put it into the mobile DB.
Yeah, I think we we use quite, quite a variety of things.
Unfortunately, our head of AI cannot join the call today.
But also we also try with like quite a lot of like, yeah, the

(39:34):
model, which a lot of different model.
And also the reason I want to say that is because we change
quite frequently, you know, thatwe used to be considering fine
tuning. We tried that as well.
It's very expensive. We see the how about we just
build our own model based on theexisting model and so on.
It's also very expensive. And based on the development for

(39:55):
AI, we realized that we can't be.
Just make sure that like yesterday, we may use code here
we realized. Another LLM has become a lot
better right now. So that's why we're trying to be
a lot more flexible that we don't want to say that oh, today
we're using these four models and tomorrow we won't using
others. We realize we have to be
catching up with the trend and people keep saying that, right?

(40:16):
Right now I think everyone say that as well.
Maybe two years ago is really good for fine tuning.
At the very moment right now fine tuning may not be that
efficient because by the moment you spend all the money and
resources on fine tuning, there will be way better solution out
there that we can leverage and asmall startup and we're not big
enough to have like 2000 fine tuning and all the hardware

(40:37):
right now. So I think we so that's why
we're trying to be humble and flexible at the moment.
Yeah, no, I think that's a really good approach.
I think. Yeah, I'm still constantly
amazed at how quickly things move in this space.
And look, we've been involved inprojects and fine tuning mostly
at the kind of code level in helping I suppose some of the

(40:58):
major LLM providers out there bebetter at Mongo DB.
So to understand it, we're giving it some, you know, data
sets for training we're giving then running them through
evaluation sets ourselves to make sure it's having an effect.
But yeah, I fully agree that thepay, things that are moving, if
you were to go down the fine tuning route and the time and
effort and resources that that takes, then, you know, the next

(41:22):
week a new model comes out that does your job for you.
It's just money down the drain basically.
Brilliant. I think, I think we've got a
really good overview of the problems that you're trying to
solve and how you're going aboutit.
But picture speaks 1000 words. So I think it's probably very
timely maybe to jump into L10 and show our viewers, you know,

(41:44):
how the system works in the background as well too.
So maybe let's get the demo up and you can talk us through
that. OK, let so we would prepare for
a small demo and we share this one because we've realized a lot
of people joining the podcast are tech manager.
And the reason we say that because we're trying to create a

(42:08):
situation like is for your tech manager, you're not the
leadership of a team. Or if you're leadership of a
team, you can use it like this way as well.
But let me explain to you. So for a platform, if let's say
if you're a manager, what you want to know is how your team
performance and how individual persons performing.
So I'm going to talk about thesetwo-part first.

(42:28):
So on the left hand side, all the feature we provide is a bit
overwhelming as a new manager. But if you has been a manager
for more than 10 years, I think you're familiar with a lot of
terms here already. But let's go into the right
part, which is very easy to understand.
We this what this part is about your team analytics.
We share with you how your team performing in terms of

(42:50):
performance and engagement. And the most important part we
realize is respond way. And a lot of manager, you may
know as well that if the people are very involved in a company
culture, they tend to answer more question.
So a lot of pitch health field leaders, they know that the
respond way and the sick leave, which is another topic we can

(43:12):
share later as well, is basically indicating the future
retention rate. So more people perform, the more
likely the people are very engaged and performing very
well. So that's why we think that this
feedback is very, very useful for a lot of people to
understand. But obviously apart from this,
we have another two parts. Let's go to engagement for for

(43:33):
engagement, basically it tell you how your team are, are they
happy, are they engaged and it'svery much influencing how well
they perform. So the the majority of people
checking engagement is with one question.
We call them the EMPS. Let me scroll to the bottom and
show you. So PS is the biggest index small

(43:55):
company using right now to calculate the loyalty of
employee. Basically ask them one question
from one to 10. How likely are you could
recommend this company to another friends or colleagues?
OK, OK, just a recommendation question.
So I'm very familiar with this question because they were
invented by Bing more than 50 years ago and still the most

(44:18):
commonly used metrics in the world.
And there's good and bad about this Question 1 is very easy to
understand, very easy to implement because you just have
to ask one question. The problem is even everyone
say, yes, I will, I will recommend this company.
You don't know why and 2nd is. Yeah, there's no right.
You're right. My 9 and Keith 9 is different.

(44:40):
I used to live in Germany so I'mI got trained to be a bit harsh.
I would never answer 9 on a question there.
To be fair. It's very, very hard for me to
do that things. Can always be better, yeah.
Because Keith, he's been workingin Hong Kong for the last, last
like decade For him, he's very genuine.

(45:00):
He was like 09 is actually OK. I will always put 9 and 10.
We talked about this topic before and this question create
dynamic that if you're if you'refrom different culture and
background, you have to actuallysee the number differently.
But in EMPS, 9 to 10 mean you'reyou love the company, 7 to 8 is
like you don't really love the company, you don't really like

(45:22):
the company and zero to 6 means you hate the company.
So we believe that this questiongood thing is very easy to
implement and understand, so easy to calculate in the old
days, but it's actually not veryfair for a lot of company and
teams. So that's why we still keep it
because a lot of companies stilluse it.
You can still OK weekly feedback, but what's more

(45:43):
important is everything about. So that's why it's a bit
overwhelming. We try because with AI with a
lot of categorization and you know, with Gen.
AI, it's very smart right now wecan do a lot more.
So the first one we call engagement score, which is how
AI will look into every single feedback from every single
person in your team to tell you how how happy they are with the

(46:04):
current situation. And you may ask, how do we do
that? We combine quite a lot of
different things like sentiment analysis, how happy, how the
emotion in the sentences, the intensity of the emotion, and
also how do we respond to different question.
So let's say if I ask you how much you don't like like your
leadership, and if you say no, I, it seems very negative, but

(46:27):
it's actually a yes in in English, right?
So yes, analyze the answer. You have to adopt to the context
of the question as well. So basically, so that's why I
won't say that it's just a simple sentiment analysis
because you can't just do that. But you may ask, right, with
this score, how can I solve the problem?
What should I do with it? So This is why we do

(46:49):
classification. Keith talked about embedding
before with the current Mongo DBstructure, is that actually a
lot easier for us to do the classification?
So we provide categories the company want us to do, but for
most, they just use our default 9 category, which is the most
recommended method in the world for culture analysis that

(47:09):
separate into does this employeealignment, company purpose, do
they love about the leadership and so on.
And we'll have different part how much they like about the
company, how much they are happywith the company.
And obviously, I'm not going to show you all the details, but
basically if you click into individual category, we show you
even more analytics on what's going on within each categories

(47:30):
so that as a team manager they would know how to solve the
problem, to address the problem.OK, OK.
And the second part is about performance.
So the basically this is like how many cell refraction we
receive and how many peer recognition we receive.
And obviously the more feedback you receive, the more
interaction between people. The bad idea is, but I don't I

(47:51):
don't want to go too much into into only 10 minute it takes
because as a team manager, your number one goal is to get every
individual person. But that's why I realise this is
actually the third important thing in our platform based on
our usage. So you've got your dashboards
and that can surface up the highlevel info quite quickly, but

(48:11):
it's really lower than that where you're going into the team
members, OK. Exactly.
The main part is the team member.
As a manager, your number one goal is to make sure every
single person is happy with onlythe genetic overview.
That's very useful for leadership.
So that's why our user, if the leadership like Hedgehog, they
really like to go into this tab.But for today, as I mentioned
before, is mainly for the manager.

(48:31):
So I really want to sort out a feature.
The first one, we call them the attrition risk.
So this one is basically to understand if this employee has
an intention to leave the company.
OK, OK, Yeah. So again, we calculate the score
and the reason based on a lot ofdifferent methods.
So basically we tell you the reason behind as well.
So but not employee. Actually, you may ask the

(48:54):
question, why do the employee want to let you know I'm going
to quit? So actually a lot of employee,
they don't want to tell you, they don't want to tell you like
I want to quit like tomorrow, but they don't mind to tell you
about the intention. They want to leave it because
OK. And we believe that if the
employee already looking for jobs, they're already
interviewing of 10 jobs, it's actually quite late for you to

(49:16):
solve the problem. You should solve the problem
when you have the intention because you, when they start to,
oh, maybe there's some jobs online.
Maybe I should go to the jobs tab on, on LinkedIn to see that
any, anything happening. So what we do here is basically
that we tell you if they have the intention to start looking
for jobs. So is that actually?
So that's why a lot of managers use it as when they see the

(49:39):
severity middle or high, we recommend them to just start
using this as a one-on-one template to start to talk to the
employee what's going on and. Is that built out of the using
the AI for the sentiment over all of the answers that that
employee has submitted over the?Time basically we extract the

(49:59):
inside the sentiment from let's say you're, you're the employee,
get a cell reflection and you will get the feedback from maybe
your manager, maybe your diary port and your peers.
And also we lack some appreciation data and all the
data we can extract the insightsand we extract sentiment and
then we can know basically how possibility that they will leave

(50:26):
the company. OK.
And it's good to see Zirion showing up as low there because
it'd be terrible if you were going to.
This is the interest of the. Company that you're just seeing.
Yeah, but this is like basicallywe see that it's very, very
useful. But you may ask, right, what if
I read this? I still want to know more like
why they're not happy or why they're very happy.

(50:48):
Then you can go into individual person, which is what we call
one-on-one evaluation. So for example, maybe I just go
into here for one-on-one evaluation.
A lot of people say that, oh, maybe they are not my direct
report. What if they're like my reports
of reports? A lot of people in the line,
they may manage teams of teams. So you can basically do that as

(51:08):
well. And let's say for this example,
so we go into like Daniel or James, we go to James.
So let's go to James. You will see that they're more
like reasonings and analytics about this person.
And as I mentioned, this is transparent between James and
the manager and we'll see that the performance score based on

(51:29):
monthly evaluation that I'm going to share later as well.
But they also respond way James can put on objectives that you
know, OK and so on. And our AI will analyze it as
well. We saw that some people use this
feature, but not not like everyone love OK in the world,
but we're not going to cover that right now.
But we have this feature like James feedback in the last 90

(51:51):
days. And this as as I mentioned is
all identified data. So enormous feedback.
You wouldn't see it here becausewe don't saw it in our database
as well. And also our 360 feedback, we
call them basically peer feedback.
The feedback that James receive from the subordinates, from the
leadership or from the peers, basically everyone can give
feedback to James. And because for that reason, we

(52:15):
can go into the analytics of James because this feature is
mainly for manager. So as a manager, I'll be like,
oh, I want to do a one-on-one with James right now.
And you're going to click this button and once you click it,
you will see a one-on-one evaluation based on the feedback
James received and provided thisperiod of time.

(52:35):
And also based James previous one-on-one report and the
one-on-one note as a manager, you can put it here as well.
Then based on community feedback, we, we wouldn't
provide at the manager the themeyou should cover with your
subordinates. Like for James, we, we believe
that you should cover these topics.
There's some gaps that you should cater right now, what

(52:58):
kind of things that they want todevelop like aerial improvement
and also the overall evaluation as well.
And this is basically our summarization from the technical
point of view based on all the feedback we provided, all the
feedback we summarize. But we realise this is the most
important feature that for especially new manager, they
really need that to guide them to do one-on-one because they

(53:19):
basically create this basic created topics for one-on-one
based on all these synthesis that we provide.
OK, very powerful insights thereAnd I suppose very, very helpful
to I suppose get the most out ofyour interactions with the
people who are reporting to you and to help the manager kind of
understand the key concerns. Because as we said in the

(53:40):
introduction, those one on ones can, you know, all too often
just be a reporting function as opposed to an enablement
function like this. Yeah.
So this is basically what we realized.
It's very useful. But the last thing I want to
tell you is about the question. So the question we basically

(54:00):
customized for individual peopleour our head of people used to
be the head of a child in AWS, Hong Kong and Taiwan.
So. To work with the industry expert
to help us to create some question.
We believe that for small and medium company is really, really
useful to ask. But also we work with as as I
mentioned before, some listed company as well.

(54:22):
We tailor a question for them. But there's a question we asked
the first time we called them engagement anomalous question
that we cover in the, in the conversation before 2nd
reflection, mainly around the goals, the skill habits, Roblox
and so on. Put some alternative answer.
The very last bit is the 360 review.

(54:42):
Our company call them 360 feedback, which is you can
gather appreciation feedback, peer feedback, upward feedback
and so on. And we have one more thing,
which is what we, we realized a lot of people start to use it.
We call them the monthly evaluation, which is a lot of
managers that we actually forgotto evaluate our employee

(55:04):
regularly. We only do once a year and
that's not fair because a lot ofpeople, they may not perform
very well in Q4, but they're actually very well in Q1 to
three. But because all the managers,
all the leadership, they evaluate the employee only on
Q4. You have a habit that people
study, become hard working only on Q4.
I think you've seen that before as well and and realise not only

(55:25):
the employee don't like it, the manager don't like it too.
So everyone don't like it and wewe believe that monthly or even
frequent evaluation is a lot more helpful.
So a lot of manager using our platform to draft that yearly
performance review report to theHR saying that you shouldn't
just look into the annual reviewby the end of the year.

(55:48):
So like very regularly. So we asked a question like
this. So it's like basically 1:00 to
5:00 weight your employees and also you can provide open-ended
question and feedback as well. Because of this reason you will
that's why you saw that we have a lot of performance goal in our
platform, but it's entirely optional.
So every feature here optional. Maybe you only you only do money

(56:09):
video evaluation, you only do weekly check in.
Basically, our goal is to provide as many data sources as
possible to collect feedback andthen to find like comprehensive
analysis for the manager to do abetter job.
But I mean like, because at the time I probably don't have time
to show you how company HR and how company leadership would use
our platform. But as we mentioned, our main

(56:31):
goal is to empower the manager. So we spend majority of our time
to build a better manager tool instead of like a hedgehog tool.
And most people using our platform, they don't think it's
HR tool, They think that it's like a manager empowering tool.
Yeah, it's a nice way, a nice way to phrase it.
It's very clear as you walk through all the sections as
well, I suppose. Is there not naming any names or

(56:52):
anything, but can you give an example of like where you did
roll out L10 within a company and within a short space of time
they surfaced up feedback that they otherwise would never have
seen or come across with their employees?
I think maybe which company we should we should use example
maybe like this startup in Korea, we really like to work

(57:14):
with them because I think they they we realize a lot of people
they provide feedback in the company and we saw that.
Let me let me go to the feature and show you and make it more
visual as we mentioned before, Sure.
So let me go into the analytics and engagement.
So I think what's I can say a lot of your example like or

(57:38):
running better one-on-one, reducing retention, reducing all
this number, but a lot of peoplethey are actually not interested
as number. What they really want to know is
the actual use case. And what we realise is apart
from the number, our most important part is reminding the
manager what they should be aware of.
And like when we go into Hemat, I remember one person we we

(58:00):
talked to, it's like, I didn't know that as a manager, I have
to take care of all these thingsabout psychological safety.
And they realised this is a really key problem for the
company. The people don't psychologically
save to share the feedback. And I feel like this is
basically our biggest learning is that our platform is not only

(58:21):
a number. They call them, how to call them
HR software, a magic dashboard. Like it's not all numbers.
A lot of time for us about the qualitative feedback and
qualitative results. So that's why I think that our
biggest learning is that we're trying to be not only be a
platform to give them number, but also kind of like educate

(58:42):
them or providing them more tipsto become a better manager.
So that's why, why now we start to have some feature like, let
me go into, I think it's in one-on-one and here like as a,
as a manager, you will see that how you can better review the
objective as a, as a, as a teaching materials.

(59:04):
And we generated these with AI and we personalized for
individual person as well as an employee.
As you write down your objective, the AI will guide you
more about how to wipe that objectives.
We believe that these actually bigger feedback, bigger learning
from this company. But obviously, if you want to
learn about the number, we help some people to increase our

(59:25):
retention quite quickly, not only not only in year, but in
weeks, OK, like we see increase,we spend with one way and so on.
But what I think our biggest learning is that we actually
educated manager to become better manager.
It's not only about goal but also all the other aspect as a
as a better manager. So the managers are becoming

(59:45):
better managers. Is there in, in the companies
that you've rolled this out in so far, is there any trends or
patterns emerging that surprisedyou perhaps in the in the sort
of feedback the system is surfacing?
Is there anything you kind of orI think as you might have said,
some things of the feedback havemade you build new features to
be able to accommodate that as well too, right?

(01:00:07):
I want to share a mistake that surprised me a lot initially.
We believe that the best way to ask question is the AI to
understand individual person, understanding what you need or
AI design who to ask, what to ask at what time.
All right, you can AI be very smart and do everything for you,
which we do like which we believe that everyone.
'S. We realized people are very

(01:00:28):
scared of uncertainty. I don't, I don't want to be like
for tomorrow. I don't even know what my
employee, what my team is answering.
So later on we like, we can't just let AI to do every single
work, which is like maybe we talk on this topic as well.
AI is really hard to replace 100% of human being work because

(01:00:48):
of this reason. So later on we realize we
realize this problem when we create new feature called
question sequence. You can turn it on.
Once you turn it on, you can setthe question what question to
ask the visual person every single day, for example, or
tomorrow you're going to ask this question because our AI
believe that you ask more 360 feet, but you don't receive it

(01:01:10):
enough. But maybe you would be like, oh,
I don't liked it. I want to ask another question
like like this one. OK, so you can see that now
instead of just let AI doing every single thing what we
believe that we should do, but we realize we should let human
being to do some choices. And this is a realise helping a
lot of people understand what kind of question they ask.
And also basically we will find more feedback like this question

(01:01:32):
only for people respond. But this question actually a lot
of people respond. It's a good question or they
actually have a lot of feedback on that.
Maybe you can have them from there as well.
So this is like what we learned is like is we shouldn't just
allow AI to make all decisions. Definitely not, definitely not.
And and I suppose look, yeah, perhaps maybe falling into the

(01:01:54):
you don't want AI to reshape management altogether and
automate it all. It is a personal thing after
all. That's the whole reason the one
on ones exist. But obviously AI is as through
your example there, the sentiment pulling everything
together, etcetera. It's really, you know, being the
backbone of L10, it's really helping build the platform.

(01:02:18):
What do you think? Will you know what's the most
exciting thing in AI? I ask all our guests this
question that you know, you think is coming down the tracks.
You mentioned Digentic at the beginning as well too.
I know a lot of I guests will answer that at the moment.
Anything else from both of you for that do?
You want to answer that or like I can share as well.

(01:02:38):
So like last week I was in London and I was in the
conference and a lot of people saying that, oh, agenda AI is
over. It's been a few months.
Yeah, it's been a big topic. Everyone talk about it.
And the reason we still talk about agenda AI because I think
it's easier to explain the line of work and because we

(01:03:00):
basically, it's basically what we use is most AI company do we
do multimodal AI and we're trying to use different AI to
solve different problems. And for our case, some of this
about one-on-one, some of them about sentiment process and so
on. But what I see the AI trend is
moving is actually helping people to do another area of

(01:03:20):
work easier. People know that as well.
I also teaching at university. I teach entrepreneurship, but
it's not like startup style. It's more like because when I
was in being, I do venture building.
So it's basically and then I understand how big company to do
innovation as well. And what we see that is a lot of
people I teach are designers andthey don't know how to do coding

(01:03:41):
in the in the past life. But now it's all the AI tool
like VC world, like lovable, like web plate, they can now
start to do some development, make it a lot easier.
So I feel like AI is really helpful for people who can do
something that they they know that they can never do before.
And what we believe that is basically here, the reason why

(01:04:03):
we build manager tool is that a lot of manager they actually not
familiar with AIA, lot of peoplewould help out.
They they actually don't know how to do that.
And we want to empower these people to have a new ability and
they don't know how to do beforeand our heart, I don't think we
don't think we want to replace human being and I don't think AI

(01:04:24):
can do it, at least for now. So that's why we we're not
branding ourselves as O and AI to replace managers.
We don't brand them that AI is still assistant to helping human
being. And the last thing about the AI
trend is in the conference, I think someone says something
very exciting is like, actually,actually AI can always replace

(01:04:45):
all my jobs. And and we, we think that maybe
there's, I don't know, like thisis some, some feeling that I
feel, I feel, we feel that as well.
I feel that AI start to do a lotof work we can't do before, but
now they can do it. And yeah, I don't, I don't know.
But this is like really hard to answer.
But we, we still believe that human being is very important.

(01:05:05):
This whole reason why we do thisstartup because of that, we
think that AI can help people. So yeah, but keep on.
Yeah, I mean. Yeah, I think fundamentally
maybe the agent AI is the way people solve the problem with a
smarter way. Maybe we can say it is a smarter
automation of work solving it has.

(01:05:29):
So if there's problem day by day, and then I think AI
actually cannot replace people, really replace human because
there's a lot of problem we needto solve.
AI is just a tool for us to leverage to.
Solve the problems. Yeah, I love that it is just a

(01:05:49):
tool and I think, you know, there's, there's a lot of talk
about it replacing jobs and I, you know, and some of the big
companies don't help, they're saying they're getting rid of
junior developers, etcetera, etcetera because the the tools
can do it. But I fundamentally believe
whilst, you know, I think exactly what you both were
saying, AI is an enabler, It's not at all a replacement.

(01:06:09):
I think, you know, the only thing to be worried about is AI
won't replace your job, but somebody using AI properly will
be better at your job than you if you don't use it.
So I don't think you can get left behind, I think, but it's a
tool. It's purely just a tool like so
many other tools we've had before to help.
And I hope, because all too often the demos that we see in

(01:06:31):
the AI space are like, you know,writing songs and creating
images and doing all this lovelystuff that in fact, I don't want
AI to do, you know, I want AI todo the boring stuff, right?
You know, so that I can create graphics and pictures and videos
and things and, and let it do the, the mundane stuff.
But I, I think that's really good.
The one thing I always love to do with?

(01:06:52):
Sorry, no, I just you remind me something that's very
interesting. I remember when I when I was in
the conference, the people were asked like what's your favorite
technology? And a lot of you would say AI,
right? And when a person asked me like,
what's your favorite technology,I asked them, is it for today or
for yesterday or for years before.
And because I told you that I used to be a designer.

(01:07:12):
So my favorite technology is always pen and paper.
And that's what we believe as well.
Like a lot of people in the veryold days, they believe that pen
is going to replace a lot of people because before, before
that, we don't think they don't,they don't think that pen
exists, right? But for us, we think that now
pen is basically part of our daily life.

(01:07:33):
We use pen to write things and we, some of them we don't use it
anymore because we use computer,we don't use it as well.
So I think that for AI is basically something similar like
like you mentioned, like we should start to embrace it
instead of just worry that AI isgoing to take my job away.
Maybe you just use it like a pen, like a like a technology
that embrace it, use it in our daily life, which we're trying

(01:07:55):
to use it ourselves as well. So that's why.
But this is the way that we don't want to say we don't
normally want to save it, because when we save it, the
people think that we're going tosell them our software.
Yeah, Yeah. It's like basing a life.
And then we're like, are you going to convince me to use our
ten? Yeah.
I think, yeah, I think it is just a tool.

(01:08:16):
I know Steve Jobs years ago usedto say computer is like a
bicycle for the mind. It's the same, you know, it's a
tool, right? And I think it's how we use it
and, and what we use it for and,and how we get the good out of
it, as it were before we go. And I I conscious of time for
yourselves. Having two founders on the show
with me is always really good. What lessons did you learn

(01:08:39):
building L10? You were both in big companies
and you jumped out of the security of that for all the
reasons that you mentioned at the beginning.
But as startups and as founders,you know what you know what's
the key lesson building L10 thatyou might have wished you'd have
known earlier or you know what kind of any key learnings.

(01:09:00):
We have a lot of kind of startupand developers join the stream
and they'd love to hear from youand.
Nothing. As a technical founder, maybe a
lot of technical guys will focuson how to implement it, how to
make the the software awesome orperformance.

(01:09:20):
But I think the most important thing is to understand the user.
Like if you build a software that is perfect, but there's no
people or no no one want to use it is I think it's waste.
It's just waste your effort to doing that.
So I think you're focusing on solving the solving the problem

(01:09:46):
of your customer is the key thing for me.
OK. Yeah, I think we're going to say
whatever in the books because wereally experience it ourselves.
Like we, we pivoted, I don't know, more than 10 times.
Yeah. And we always think that this
product is going to work, but it's not.
So I think, but you heard that as well.
Like I think every founder knew that like they said that don't

(01:10:07):
focus on the first product trying to test with user.
We, we made a lot of mistakes like before, remember this one
product, we built it for like 2 months.
It didn't work and we shouldn't do that.
But we're like, no way this is not working.
And we talked to a lot of user, you know, interview, this is the
problem we're solving for them. But I feel like for us, the

(01:10:28):
biggest learning probably is we prepare ourselves to pivot a lot
and we still pivoting today. There's no way that the final
product, we all know that it's not, We hope that it's not.
We hope that the product will bea lot better.
So we're prepared and. I love that.
I love that Prepare, prepare to pivot.
Yeah, I, I, I used to before Mongo DB, I used to work with a
lot of startups as well too. And, and all too often, you

(01:10:52):
know, they get really stuck in their idea and they, they don't
want to launch it. They don't want to put it out
there publicly because they don't want any real world
feedback because that would maybe make them pivot or derail
their, their plans. And it's like, just get it out
there. See what people give you the
feedback on, see what they use it.
I think the key question I used to, you know, tell our clients

(01:11:13):
at the time was ask somebody, would they pay for that?
You know, because that's really the key thing is like, yeah,
everyone loves this. It's a great platform.
Or would you would you put your hand in your wallet and pay for
it? You know that changes the
conversation quite quickly as. Well, I think for us is was
someone just pay it without talking to you.
We have a user just when them they go into a website and just

(01:11:35):
pay it without contacting us. I think that's a really big
thing for us, like suddenly there's a payment going to a
swipe. Account that I can imagine yeah,
that's a massive buzz that you have no previous contact with
that they came on during account.
Brilliant. Yeah, I love that.
I love that. So listen, I think we we covered
an awful lot. It's been great to hear your

(01:11:56):
journey, the inception of L10, that level 10.
These are the things that you should be concerned about at
these meetings and this is the tool that's built to help grab
those and why you keep pivoting as well too and add more
features and you know, as you do.
And maybe in eight months or a year, I get you both back on the
show again. Let's see what L10 has become

(01:12:17):
now as well too. And Keith, you, you encapsulated
all the good stuff about why choose Mongo DB.
So I'm not going to go back overany of that.
Where can people learn more about L10?
How did they get started? You mentioned kind of somebody,
you know, the buzz of somebody finding it and signing up of the
service. So get the plug out there.
Where do they go? Yeah, so we created one page let

(01:12:42):
let, but you can go to that likeI think for today we prepared an
author page. Our website is called L10 dot
tag but if you go to L10 dot tagslash mongo DB you can get free
credit trial software S. Oh, excellent.
Do you want to bring that up there?
And I'll put it into the comments.
So yeah, as well too. Maybe, maybe on I put it on our

(01:13:05):
private chat Perfect, the L10 dot TECH and mongo DB.
We still have launched it beforethe call.
If not, we'll launch the websiteright after the call.
Yeah, we still have done that. But you know as a startup.
In the startup basically we have1000 things on reminder list.

(01:13:28):
You know, you see, you see this is proper live stuff going on.
This is there so so folks who are joining us and seeing the
comments, I'll post that into LinkedIn now shortly as well
too. Not live yet.
I'm getting a 404 then. We we're going to check it right
after the call. Perfect, perfect.
No, that. That's brilliant too.

(01:13:50):
If if they go there, do they get?
Yeah, if they go there, they'll get some credits to get started.
Yeah, we would prepare like 1000US credits.
So I think it's enough to use itfor quite a while.
Excellent well, listen, that's abrilliant call to action towards
the end of the stream as well too.
Hopefully everybody everybody goes there.
So it's for those joining and watching and maybe not looking

(01:14:14):
at the comments L10 dot tech forward slash mongo DB.
So please go there and try it out.
Hopefully you'll have great success.
Any last words then Ziri and or Keith for our audience before we
go for today? I think I always, I always say
that in a, in a, in a talk that just be a better leader, not a

(01:14:36):
manager trying to care about every single person in your
team. There's no low performance.
It just you we you haven't helped them enough in your team.
Yeah, Yeah, I think, I think that's key, yeah.
Maybe I can give some works to maybe the stock of founder
attending this webinar on this this live stream.

(01:14:59):
I want to say that emotionally there's a lot of up and downs as
a star founder and yeah, believe, I think all of us need
to believe your idea. And then, yeah, don't give up.
Believe the team. Exactly.
We say that all the time. Well, I love all the content of

(01:15:22):
what we chatted about today. I love the fact that you two
were in school together as youngboys and all the way up.
And then I founded this as well too, you know, went through two
separate ways and back together as well too.
But above all, thank you so muchfor joining me and sharing how
L10 helps organizations and managers in particular, listen
better and and lead smarter as well too.

(01:15:44):
So that that's brilliant. Thank you for everybody who
joined us on the live stream as well too.
It is being recorded so you'll be able to catch up and the link
is there as well. L10 dot tech forward slash mongo
DB when you turn it on to make it some credits available to
you. But for everybody else, it's
been a pleasure, Zurian and Keith to have you on the show.

(01:16:06):
Thank you so much. I really appreciate your time
and your effort and the insightsand the the knowledge that you
shared. So thank you so much for joining
the the MongoDB podcast. And for everyone else, we do
this every Tuesday or so, or I do it most Tuesdays anyway.
So just like and subscribe on YouTube and follow us on

(01:16:27):
LinkedIn and you'll get alerts for more shows like this coming
in your way. But for now, from me in in
Ireland and Zurian and Keith in Hong Kong, it's been a pleasure
to have you both. Thank you so much.
Thank you so much for the day. Thank you.
Excellent. Take care everyone.
Good luck. Have a good day, bye.
Bye.