Innovating Video Search: Insights from Tech Pioneer Soyoung Lee

You are listening to the Business
Leadership Podcast with Edwin.

Paul: I'm Paul Newton, creative producer
of this future narrator mini series,

and I'm joined by Edwin Fzo, the host
of the Business Leadership Podcast.

We're recording live at Web
Summit Vancouver 2025, and we're

exploring how today's leadership.

Leaders shape the future, not just
through strategy, but through story.

We believe that strong points of
view are what inspires community

builds movements and cuts through
the noise in uncertain times.

So let's dive in.

We got a special guest today and today's
guest is building AI that doesn't

just watch video, it understands it.

So young.

Lee is co-founder of 12 Labs, a pioneering
research and product company developing

multimodal foundation models for video
understanding with a mission to make

video as searchable and accessible as
Text 12 Labs is helping developers embed

deep visual context into real world
applications from media and entertainment

to security education and beyond.

Born from a team of deep learning,
scientists and veterans of cybersecurity

and retrieval systems, 12 Labs has quickly
become a rising force in the AI space,

earning top honors in international
benchmarks, and securing major investment

to scale their vision at the intersection
of machine learning, creativity, and

real world impact so young is shaping
how we interact with the world's

most information Rich, medium video.

Edwin: Welcome to the Business
Leadership Podcast, so young.

Soyoung: Yeah.

Thanks so much for having me.

Yeah.

Edwin: How are you doing?

How you, how are you experiencing,
how are you feeling about,

uh, web Summit Vancouver?

Soyoung: It's awesome.

It's actually very inspiring.

Um, like the trends of what's being built
is so fast changing and so creative,

so it's actually very cool to see
companies that are building with the

latest, newest things that we're seeing.

Um, in the AI landscape today,

Edwin: are you, are you getting new ideas?

Soyoung: Always.

Edwin: Always.

That's amazing.

So, I guess where we wanna go is
if you could just, let's just start

by sharing, you know, 12 labs.

If you could share what is the
problem you are solving in the

multimodal AI and video understanding?

Soyoung: So, the real problem
we're solving for is that video is.

Really difficult to search for.

So, uh, I think we're all very familiar
with YouTube search, but YouTube search

is really, you're searching across
keywords and tags and descriptions

that the original uploader has.

You know, so if you're trying to search
across your own library of footage,

which I'm sure each of us holds onto
so many memories of like family, pets,

personal communication, um, it could
be, you know, forms of like learning

and education that we've stored.

You have to actually watch it
to an tag it to be able to find

moments in your, uh, in your lives.

Yeah.

And so that's really the problem that
we set out to solve, which is how

do we make video more accessible?

And video is very special because
it's the most natural form and the

most resonating form in which we
communicate with each other as humans.

Um, and tell stories and capture,
you know, the real world.

Um.

That's a, like, fundamental
problem that we help solve.

And then there, you know, there's
a plethora of like business

problems if we can get into that.

Yeah.

But, um, yeah,

Edwin: I mean, you, you know what
came to mind so young as, as you

were talking, I, I have a 2-year-old
son and my daughter's almost nine.

Mm-hmm.

But I have a lot of video from them.

Like if I don't remember where exactly,
like how to, like I, and tell me if I'm.

Off track here, but if I was, if
I remembered a moment, I could

probably, working with you in 12
Labs is like, oh, I wanna find

that moment where Kobe did this.

And then maybe through your contextual
or understanding of video, I wouldn't

have to search for, I mean, I
wouldn't have to spend like skimming

through all this video traditionally
just to find that one moment.

Is it, is that the problem?

Because that, that's a big
problem for me right now.

Soyoung: That, that that is exactly right.

Right.

Yeah.

Um, and if you think about like
content organizations or like.

Larger enterprise organizations, then
you can translate that into, you know,

metrics of how, what's the efficiency
and opportunity gains that come from it.

But deep down it's really
a personal problem of

Edwin: like, yeah,

Soyoung: I remember like two years ago we
were at SIGGRAPH and I like tried eating.

Fried cucumbers for the first time.

Oh wow.

Edwin: Those are good though.

Soyoung: They're good.

Uh, but somebody on my
team took a video of that.

Yeah.

And so we actually as a team,
uh, we store our team memories

in our own, uh, like service.

Yeah.

And so we, like, we almost use our own
service to be able to search through

our memories together as a team.

Edwin: Mm-hmm.

Soyoung: Um, but if you can
do that beyond that and just.

You know, all the personal experiences
and memories we're experiencing, how do

you make that accessible for yourself,
you know, for your loved ones and so on.

Um, you know, that's, it's actually
essentially the same problem.

Edwin: Yeah.

I mean, it, I mean, just as you, as
you go in, I mean, like, I'm, I'm just

geeking out right now, so, uh, stop me.

But it's just like, yeah,
we're getting into that world

where everything, everyone, and
everything is recording, right.

And mm-hmm.

And when.

Glasses are recording everything.

Like there's moments that you may
not even notice, right, that are

happening anymore or even doing this.

And I could even just imagine some
type of like contextual things that you

could probably find, hey, what happened
today and where, where did it happen?

Where did I see something?

Where did it do that type of stuff.

So.

I'm sure you thought about all this stuff.

It's just as I, as I think
about it and do that.

So I'm curious, Soen, when it comes to
that, you know, you've been deep into

this space now, you've been working
with it, you've been working with, uh,

different in within sports and, and,
and I don't want, I don't want to, I

and other industries, I'm sure, but
what is the unique point of view, uh,

that you bring to this problem now?

Soyoung: So traditionally be ev,
every organization that has a lot of

video, whether it's security, whether
it's manufacturing, automotive,

and also also of course media
organizations, they've done build,

tried everything to analyze video.

And it's not that the technologies
were like so severely lacking.

It's just a such a complex problem.

Mm-hmm.

Because.

You know, real world is messy, but
that's really well captured in video.

And so if you need to do video
understanding, well, you need to

understand sound background sounds.

You need to understand dialogue
and semantics of language,

um, visual information.

But most importantly, you also
have to connect all the dots across

how things are changing over time.

So that requires an understanding
of temporal context and narrative.

Um, so that was really our unique, I
guess, perspective when we first started.

Mm-hmm.

Uh, four and a half years ago.

What's a unique model architecture
and a unique way in which we can, the

model, the AI itself can almost create
this human-like understanding of video.

And be able to scale that across vast
amounts of data and then enable other

product builders that are building
really cool applications, they're

building cool, like in incredible
workflows internally within their orgs.

Um, you know, to power their next,
their set of creative activ like

creative activities or experiences.

Paul: Okay.

Well, I mean this is really, um, it's
really fascinating 'cause I mean,

you come from like this background
of uh, you know, just tech and, um,

uh, you know, just you're a learn, a
deep learning scientist and security.

But now we're talking about all
these different things about like

the semantics and the narrative and
what, what interests you in this?

What are, yeah, what are you motivated

Soyoung: by here?

Um, you know, we have five co-founders
and I'm sure each of us are

going to have a different answer.

Mm-hmm.

For me personally.

Um, I just watch a lot of video.

I'm a perpetual scroller.

I, uh, I, I don't think my English
is terrible, but I actually learned

English by, uh, I grew up in Korea.

Okay.

So I learned English by watching a
lot of American television and movies.

So that's how I absorbed
culture and language.

Uh, and so it's very, uh, like.

Dear to me, and I understand how
significantly impactful it is, um,

to consume content that has been
created by incredible storytellers.

So I'm not coming at this from a,
like creator's perspective, but it

really is from a consumer's pers
like viewer's perspective of wanting

to, you know, have access always
to stories that resonate with me.

Um.

Yeah,

Paul: well, I'm, I'm seeing
something unique here.

I think, you know, just, and, and in
knowing that this wasn't your first

language, but you've, you know, you've
really mastered it and you've, you

know, you're mastering the culture
and interactions and that really does

come down to tone and timing and the
different things that you're analyzing

for or studying in these videos.

Soyoung: Yeah, absolutely.

Um, and so, you know, we were
talking about this a little

bit before the podcast, but.

There are some videos where actually
a lot of videos where it's, you don't

need the full multimodal context.

Maybe all you need is speech to text.

Yeah.

And you take the transcript and you
just summarize what's being spoken.

Um, in many cases though, that
we, you know, we deal with in our

customers are dealing with, there's an
importance of intonation, of nuance.

That's.

In the, in the sound of like how we speak.

Edwin: Mm-hmm.

Soyoung: Um, it could also be
background sounds like things that

are happening in the background
and sound as we're speaking.

It could also be visual.

What facial expression am I, you
know, showing, um, you know, what

am like what's the body language,
what's happening in the background.

Um, and so connecting all those
dots, I'd say is like the essence

of like video understanding.

Paul: Yeah.

But I, I think like.

A really unique perspective you bring
is just like how much you gained

from being able to watch these videos
and learn what it really means to

be like North American or American
and just like get into this culture.

So, you know, how does that impact
what you're bringing to the company?

Soyoung: Yeah.

Um, so I think it's,
it's a very human thing.

Mm-hmm.

It's a very human thing
to want to tell stories.

It's a very human thing, um, to want
to hear and, you know, continue to.

Consume great stories.

Mm-hmm.

Um, and it gives you humility, I
would say, uh, as a technology company

that's working with companies, you
know, that are, may not necessarily

be based in Silicon Valley.

Yeah.

Because the customers that we work with
are, you know, organizations that are

building something entirely different.

And so it requires a lot of humility,
even, you know, to be able to say, okay,

I know nothing about the workflows of
creating content, or you know, to search

through large archives of evidence.

I'm going to, this is what I have.

We have the ability to understand video.

Edwin: Yeah,

Soyoung: but I'm gonna put that down.

Mm-hmm.

And just teach me, please teach me and let
me know what you do on a day-to-day basis.

What do you do on the weekend?

How do you actually get footage from
camera and actually log and like.

Maintain your spreadsheets of, you know,
logs of what's happening in your footage,

how do you actually search for it?

How do your creative teams, uh,
access that content later on?

What's the process to
actually publish that?

Um, and so on.

So that's like just one example, but I,
it is just, I think it's the curiosity

and the understanding and like.

With respect that kind of helped,
I guess, us, uh, work really well

with the key, I guess, lighthouse
customers we had in the beginning.

Yeah.

Paul: And I mean, it seems like you
really, uh, you know, you really kind

of connect to the, the purpose and
the, like the, the real underlying

reason people are using their video
in their, in their business mm-hmm.

And how it does connect with people.

Soyoung: Yeah.

Yeah.

Absolutely.

Paul: Oh, very cool.

Edwin: Um.

We went through here, right?

Yeah.

Okay.

Just double checking 'cause I know.

Soyoung: Sorry, I was
like all over the place.

No, no, no.

Edwin: It's good.

It's good.

Um, so, so there's two there.

There's.

Two last things mm-hmm.

That we want to go over before we let you
go soya in terms of like yourself as a,

as a co-founder, one of five, I guess,
um, I'd love it to share for those who

are listening, who are watching, you
know, the purpose for you as a, as, as an

entrepreneur, as a founders of, you know,
someone who is really telling the story.

Um, what was the importance of coming
to a web summit Vancouver, and what are

you like, what is your ultimate goal?

Here, like, you know.

Yeah.

It would be like a home run.

Soyoung: Yeah.

So my role as a co-founder
is on the go-to market side.

So I lead the go-to market organization.

Mm-hmm.

Which is everything from, um, growth
and marketing, our developer relations

work, um, to of course sales.

Uh, so my goal number one is
always let's find customers.

Yeah, of course.

Yeah.

Um, but I think the bigger,
uh, like broader purpose is.

Um, thinking about how do we create more
value for our users and customers and.

You can't think about that as a, here's
a limited product and I'm gonna go create

value with this thing I have right now.

Mm-hmm.

Because your product and the technology
have has to evolve very quickly,

um, to be able to expand the value
you're creating, and then you can

perhaps, then we can come back and
talk about how do you capture that,

Edwin: right?

Mm-hmm.

Soyoung: So what, so I, it's
been incredibly inspiring to

meet all these founder, other
founder, fellow founders, um, and.

Not just founders, but
also prospective customers.

Uh, but to hear like, what's being
built, what's new, what's something

that we hadn't thought about.

Uh, so, oh, like that's
been very exciting for me.

Edwin: Okay.

That's amazing.

Um, there are any final thoughts,
advice, uh, or recommendations that you

could share with, you know, the other
founders, the CEOs, or even business

leaders who may be listening today?

And you could take this, whatever
first thing comes to mind.

Soyoung: Um.

I'd say like, spend a lot of time with
like partners and customers you work with.

Build like the relationship and trust that
allows you to learn as much as possible.

Mm-hmm.

Um, because that's, you
know, will help you.

Uh.

Create more value, like
as you're building.

Mm-hmm.

Um, and try a lot of stuff
as in like a lot of products.

'cause there's, this is such
an exciting space right now.

Yeah.

New, you know, new things
being born every day.

Um, so the more you have hands-on
experience trying new tools and

applications, uh, the more you're,
you know, perspective will expand.

Um, and it's like very humbling every day.

Yeah.

No, I mean, I mean,

Edwin: that's good advice in terms
of really, I, I guess my key takeaway

from, from your share is like, really.

Stay present with your cl, with
your partners, your clients,

be in conversation with them.

And as you and I talked before, uh, I,
at the beginning of the recording is

like, you're getting lots of ideas over
here, but that's, that's brilliant.

But if your customers and your
partners aren't aligned to it or

asking for it, then at least you will.

I.

Stay on track, right.

In terms of your focus and your goals
as personal goals and business goals.

Absolutely.

Um, before we let you go, we just want to
share, so this is just a quick project,

VARs, we went to collision last year.

We did the same thing.

Mm-hmm.

We interviewed future narratives,
people who were coming here,

sharing their vision, and we
created a book podcast, minie.

So we just wanted to give that
to you, future narrator, and, uh,

excited to, to share that with you.

Soyoung: Thank you so much.

You are listening to the Business
Leadership Podcast with Edwin.

Innovating Video Search: Insights from Tech Pioneer Soyoung Lee
Broadcast by