Latest / SEA of Startups / 🎙EP3 Deep, Narrow, and Agentic: The New Stack That’s Quietly Taking Over

Transcript

🎙EP3 Deep, Narrow, and Agentic: The New Stack That’s Quietly Taking Over · SEA of Startups

• 0:01
  Welcome to the Sea of Startups,
• 0:04
  where Southeast Asia's startup game gets real.
• 0:08
  SPEAKER_00: No tech growth love,
• 0:10
  SPEAKER_00: no Silicon Valley cosplay,
• 0:12
  SPEAKER_00: just bold takes,
• 0:14
  SPEAKER_00: sharp turns,
• 0:15
  and unfiltered stories from founders and funders building the
• 0:19
  future of this
• 0:21
  SPEAKER_00: region.
• 0:22
  SPEAKER_00: I'm Kim,
• 0:23
  SPEAKER_00: the producer and co-founder of this relaunch,
• 0:26
  SPEAKER_00: and alongside Kevin Brocklin of Indelible Ventures,
• 0:29
  SPEAKER_00: we're pulling back the curtain on what it actually takes to raise,
• 0:34
  SPEAKER_00: build,
• 0:35
  SPEAKER_00: and win in the fastest growing markets in the world.
• 0:39
  SPEAKER_00: If you're in a game or want in, this is where the real talk happens.
• 0:51
  SPEAKER_00: If last week was about playing smart with your capital,
• 0:56
  SPEAKER_00: this week's about playing smart with your stack.
• 1:01
  SPEAKER_00: These days, we're seeing a quiet reckoning in how software gets built.
• 1:07
  Not just the AI that automates,
• 1:09
  but a genetic AI that adapts,
• 1:13
  decides,
• 1:14
  and acts like a teammate.
• 1:17
  It's showing up in smarter APIs,
• 1:20
  the connectors that act and not just connect.
• 1:24
  SPEAKER_00: And paired with focused tools, it's quietly transforming how teams work.
• 1:31
  SPEAKER_00: And this isn't the flashy side of tech, but it's a useful side.
• 1:37
  SPEAKER_00: And it's where the next wave is being built.
• 1:40
  SPEAKER_00: To unpack it,
• 1:42
  SPEAKER_00: Kevin sat down with Chris Burrell,
• 1:44
  deep tech operator and multi-exit founder to break down what
• 1:48
  agentic software
• 1:50
  SPEAKER_00: really means,
• 1:52
  SPEAKER_00: why the deep and narrow wins,
• 1:55
  SPEAKER_00: and how C founders can build ahead of the curve.
• 2:00
  SPEAKER_00: So let's get into it.
• 2:09
  SPEAKER_01: In this episode,
• 2:10
  SPEAKER_01: we're going to be diving into something that's on everyone's radar AI,
• 2:14
  SPEAKER_01: but not just any.
• 2:15
  SPEAKER_01: We're going to get a little bit of context in it.
• 2:17
  SPEAKER_01: I've got an expert here alongside me.
• 2:19
  SPEAKER_01: Our guest today is Chris Burrell.
• 2:22
  SPEAKER_01: He is the founder of Funk Ventures.
• 2:25
  And a former senior leader across a number of organizations has
• 2:28
  gone through
• 2:29
  SPEAKER_01: multiple exits and is based in Kuala Lumpur,
• 2:32
  SPEAKER_01: good friend of mine,
• 2:33
  SPEAKER_01: investing across APAC.
• 2:35
  Chris,
• 2:35
  maybe you can give a little bit more flavor to your background
• 2:37
  than I just did.
• 2:38
  SPEAKER_01: And thank you for being with us.
• 2:41
  SPEAKER_02: Thanks very much, Kevin, for the intro.
• 2:44
  And I must say,
• 2:44
  I'm a long-time listener,
• 2:46
  first-time caller of the podcast.
• 2:48
  SPEAKER_02: I'm very, very excited to be here.
• 2:50
  SPEAKER_02: I've been fortunate enough in my career.
• 2:52
  I moved out to Asia initially to Hong Kong 30 years ago and
• 2:55
  from Australia
• 2:56
  SPEAKER_02: originally and really found my legs out here as a technical leader.
• 3:00
  I was fortunate enough to work for four different software
• 3:04
  companies out here as
• 3:05
  SPEAKER_02: their head of Asia Pacific and
• 3:07
  SPEAKER_02: Those companies either IPO'd or were acquired.
• 3:10
  SPEAKER_02: So I went through four exits.
• 3:11
  SPEAKER_02: The last one,
• 3:12
  SPEAKER_02: I was the CIO of a company called Bluecoats,
• 3:15
  SPEAKER_02: acquired by Symantec for $4.5 billion.
• 3:18
  SPEAKER_02: So these companies became quite big through my career.
• 3:21
  And what I didn't like about the big companies was they were
• 3:24
  slow and hard and not
• 3:25
  SPEAKER_02: very innovative.
• 3:27
  SPEAKER_02: We actually did a lot of M&A and I was part of the M&A team.
• 3:30
  SPEAKER_02: And that's really where I saw the innovation happening.
• 3:35
  SPEAKER_02: And I really wanted to get back to that.
• 3:38
  SPEAKER_02: early stage where you're actually really disrupting and innovating.
• 3:43
  SPEAKER_02: So 2019,
• 3:44
  SPEAKER_02: I left the corporate world,
• 3:46
  started doing a bit of angel investing and sort of accidentally
• 3:51
  became a VC in
• 3:52
  SPEAKER_02: 2022.
• 3:53
  SPEAKER_01: That's how it started.
• 3:55
  SPEAKER_01: An accidental journey into becoming a VC.
• 3:58
  So tell me,
• 3:59
  because I know your thesis on Funk Ventures because you and I
• 4:02
  know each other well.
• 4:03
  SPEAKER_01: But for those people listening,
• 4:05
  tell a little bit about what your focus area is and why you
• 4:09
  accidentally became a
• 4:10
  SPEAKER_01: VC,
• 4:11
  SPEAKER_01: but I'm assuming intentionally chose that piece.
• 4:13
  SPEAKER_01: And I have a question.
• 4:14
  When I was the CIO at Blueco,
• 4:17
  we grew the company from 400 million revenue to a billion
• 4:20
  revenue.
• 4:22
  SPEAKER_02: across a period of four years.
• 4:23
  SPEAKER_02: And I was the economic buyer of software.
• 4:26
  I was the guy signing the check to Salesforce.com or the Oracle
• 4:31
  ERP suite and everything else.
• 4:34
  And you would sit down with a finance executive once a week for
• 4:37
  a few hours,
• 4:38
  SPEAKER_02: go through this massive spreadsheet and go,
• 4:41
  SPEAKER_02: wow,
• 4:41
  SPEAKER_02: we spent a lot on software.
• 4:43
  SPEAKER_02: It's really expensive.
• 4:44
  My boss,
• 4:45
  because we were owned by private equity,
• 4:47
  was saying,
• 4:47
  how do we cost,
• 4:48
  not spend more?
• 4:49
  So as an economic buyer of software,
• 4:52
  I knew I can't just go and buy another SaaS product.
• 4:57
  SPEAKER_02: If I want to, I've got to turn something else off.
• 4:59
  SPEAKER_02: So there's a very high bar in selling into me,
• 5:04
  a very high bar for these sales executives to walk into my
• 5:07
  office and sell me
• 5:08
  SPEAKER_02: something new.
• 5:09
  But what I did notice is that my engineers were signing up to
• 5:12
  smaller services,
• 5:14
  SPEAKER_02: things like Jira and Twilio.
• 5:17
  and adding it into their stack almost under the radar by using
• 5:21
  a corporate credit
• 5:22
  SPEAKER_02: card or what have you.
• 5:23
  They were kind of bypassing my cost-cutting mechanism and they
• 5:27
  were prototyping
• 5:27
  SPEAKER_02: these products and they would bring it to me and say,
• 5:29
  SPEAKER_02: hey,
• 5:29
  SPEAKER_02: Chris,
• 5:29
  we're using this API here to do SMSs and it's a lot cheaper
• 5:33
  than our existing
• 5:34
  SPEAKER_02: provider.
• 5:35
  So we're going to turn that off and we're going to save you
• 5:37
  this much money.
• 5:38
  SPEAKER_02: I was like, wow, these APIs are pretty cool.
• 5:41
  They allow people to sell quickly without having to go through
• 5:45
  a bureaucratic buying process.
• 5:47
  SPEAKER_02: So that initially attracted these companies to me as an entrepreneur.
• 5:52
  But then what happened is these companies became quite
• 5:56
  important and bigger.
• 5:57
  So you could do this deep and narrow strategy to get some
• 6:01
  initial traction,
• 6:02
  SPEAKER_02: but then you could land and expand.
• 6:03
  SPEAKER_02: Think of companies like Stripe where they've come to.
• 6:06
  SPEAKER_02: They were originally a very simple payment gateway.
• 6:09
  SPEAKER_02: They now do full stack transactions.
• 6:11
  SPEAKER_02: POS, treasury function, payout and pay in.
• 6:15
  SPEAKER_02: And they're a massive company now.
• 6:16
  SPEAKER_02: So that's what attracted me.
• 6:18
  It's like the business model because it was a way for an
• 6:21
  entrepreneur to build a
• 6:22
  successful company and not having to go through that painful
• 6:27
  process of selling a
• 6:28
  full stack SaaS product with a user interface that required
• 6:31
  change management into
• 6:33
  SPEAKER_02: a company.
• 6:34
  SPEAKER_02: So that's what attracted me to me initially.
• 6:37
  SPEAKER_01: Let me ask you one thing just definitionally.
• 6:40
  SPEAKER_01: Most of the people listening in are tech founders.
• 6:42
  SPEAKER_01: There's likely a wide understanding already,
• 6:45
  SPEAKER_01: but in API,
• 6:47
  can you explain like super simple layman terms in regards to
• 6:50
  the examples of the
• 6:51
  SPEAKER_01: Stripe,
• 6:52
  SPEAKER_01: Twilio,
• 6:52
  SPEAKER_01: et cetera,
• 6:53
  SPEAKER_01: how they operate and fit into TechStack as you referenced?
• 6:57
  SPEAKER_02: Sure.
• 6:58
  SPEAKER_02: Put very simply, APIs are how software talks to software.
• 7:03
  SPEAKER_02: Typically, you don't have a keyboard and a mouse.
• 7:07
  SPEAKER_02: You're using code, some kind of software.
• 7:12
  SPEAKER_02: For example, an iPhone has the Grab app inside the iPhone.
• 7:19
  The people that built that app,
• 7:21
  they don't want to go build their whole own mapping technology.
• 7:25
  SPEAKER_02: So there's an API into the Google Maps software they use to,
• 7:30
  SPEAKER_02: you know,
• 7:30
  draw the map and allow you to search an address and figure out
• 7:34
  the route between
• 7:35
  SPEAKER_02: address A and address B.
• 7:36
  SPEAKER_02: So that's an example.
• 7:38
  So the Google Map itself is embedded in that application and
• 7:44
  Google makes money
• 7:45
  SPEAKER_02: every time somebody uses...
• 7:47
  SPEAKER_02: The Grab app or the Uber app.
• 7:50
  SPEAKER_01: It's a way of not necessarily having to build each individual piece.
• 7:53
  You can use an API to call another software to get it kind of
• 7:56
  integrated in with it
• 7:58
  SPEAKER_01: without having to like build every single piece.
• 8:00
  SPEAKER_01: Although Grab got tired of paying the millions to Google.
• 8:03
  SPEAKER_01: So they started with Google.
• 8:04
  SPEAKER_01: Now they've since replaced.
• 8:08
  SPEAKER_01: They were doing it originally.
• 8:10
  Yeah,
• 8:11
  I'm sure they've got their own or they've probably partnered
• 8:13
  with somebody cheaper.
• 8:14
  And it happens in,
• 8:15
  you know,
• 8:15
  Twilio was kind of the gold standard in SMS API.
• 8:19
  SPEAKER_02: When you want to send an SMS for an OTP, use Twilio.
• 8:22
  SPEAKER_02: But there's lots of other alternatives now as well.
• 8:25
  SPEAKER_02: So that's the...
• 8:28
  SPEAKER_02: That's the example.
• 8:29
  SPEAKER_02: Carefully don't limit it to those examples.
• 8:31
  SPEAKER_02: There's things like plugins as well.
• 8:33
  Imagine a plugin to Shopify,
• 8:36
  like a widget that does a charity donation inside the checkout.
• 8:40
  I like those kind of companies as well because they're like not
• 8:43
  pure API plays,
• 8:44
  they're more platform plays where you're selling that widget
• 8:48
  across thousands and
• 8:50
  SPEAKER_02: thousands of different platforms as well.
• 8:56
  Super interesting how this has enabled a lot of software layers
• 8:59
  and the unbundling
• 9:00
  SPEAKER_01: of the monolith.
• 9:01
  SPEAKER_01: How do you see this evolving now?
• 9:03
  Your thesis is likely evolving in the era of more AI as we're
• 9:09
  talking about agents
• 9:10
  SPEAKER_01: that can talk between different software.
• 9:13
  SPEAKER_01: How are you seeing this evolution and how are you adjusting your mindset
• 9:17
  SPEAKER_01: considering this is your core niche?
• 9:21
  Yeah,
• 9:21
  just talking about it more philosophically,
• 9:25
  I like founders that are deep and narrow.
• 9:28
  SPEAKER_02: So they're very deep in a particular vertical and very focused.
• 9:33
  SPEAKER_02: So I invested in one company that does KYC.
• 9:39
  SPEAKER_02: They do KYC as a service and they broker multiple KYC providers.
• 9:44
  SPEAKER_02: They can do 400...
• 9:46
  SPEAKER_02: KYC providers across about 120 countries now,
• 9:50
  SPEAKER_02: I think,
• 9:51
  SPEAKER_02: companies called Frankie One,
• 9:53
  SPEAKER_02: they are very deep and they really just do KYC.
• 9:57
  SPEAKER_02: That kind of founder,
• 9:58
  they're very well positioned to now attack the new emerging
• 10:03
  market of agentic AI.
• 10:06
  I don't know,
• 10:07
  do you want me to give a quick example of what agentic AI
• 10:09
  means?
• 10:10
  SPEAKER_01: Yeah, let's do that.
• 10:12
  SPEAKER_01: I think people would benefit from a little bit of definitions here.
• 10:16
  SPEAKER_02: Yeah,
• 10:16
  so everyone understands how the typical kind of AI large
• 10:21
  language model products
• 10:22
  SPEAKER_02: like ChatGPT or Gemini works.
• 10:26
  You ask it a question and using its immense understanding of
• 10:30
  everything that's ever
• 10:31
  been written comes up with an answer that's usually interesting
• 10:36
  and accurate and
• 10:38
  SPEAKER_02: getting better every day.
• 10:39
  SPEAKER_02: And I think a lot of us use that for things like
• 10:41
  SPEAKER_02: creating content for English language or foreign language content,
• 10:45
  SPEAKER_02: images,
• 10:46
  SPEAKER_02: even movies now,
• 10:47
  SPEAKER_02: right?
• 10:47
  SPEAKER_02: If you think about the particular very deep and narrow specific topic,
• 10:53
  SPEAKER_02: like KYC,
• 10:54
  SPEAKER_02: an example,
• 10:54
  SPEAKER_02: the large language model itself cannot...
• 10:58
  SPEAKER_02: onboard my driver's license.
• 11:00
  SPEAKER_02: It didn't program my driver's license and address into its model,
• 11:04
  but it may want to be able to embed something into an answer
• 11:11
  that verified who I am
• 11:12
  SPEAKER_02: and using my driver's license.
• 11:14
  SPEAKER_02: So there's multiple ways you can do that.
• 11:17
  SPEAKER_02: One of the ways is Agentic AI.
• 11:18
  SPEAKER_02: Agentic AI is a
• 11:20
  SPEAKER_02: long running process that has the ability to do multiple tasks.
• 11:27
  SPEAKER_02: So the task might be, I want to give Chris a membership to my gym.
• 11:35
  and that's going to want to know where I live when I want to
• 11:38
  validate my driver's
• 11:39
  SPEAKER_02: license and some of my health data.
• 11:41
  SPEAKER_02: The agentic AI can talk to the KYC provider, do that.
• 11:45
  SPEAKER_02: It may also go talk to my,
• 11:47
  I might give it access to my health data and then issue me with
• 11:51
  a gym pass and then
• 11:53
  SPEAKER_02: send me a wallet card for my phone that's got my gym pass in it.
• 11:58
  SPEAKER_02: Normally as a programmer,
• 11:59
  SPEAKER_02: I program all those steps myself,
• 12:01
  SPEAKER_02: but agentic AI can actually go and do the whole thing
• 12:05
  SPEAKER_02: and come back and tell me.
• 12:06
  And it might have some humans in between that process that
• 12:09
  interact with it,
• 12:11
  SPEAKER_02: but essentially it's doing that whole task as a bundle.
• 12:13
  SPEAKER_02: And then as a user of that Agente AI service,
• 12:18
  SPEAKER_02: let's say I'm the gym owner and I employed that Agente AI process,
• 12:21
  SPEAKER_02: I might pay 10 bucks instead of 50 cents to use the API.
• 12:25
  That's more expensive,
• 12:27
  but it did the whole thing and I didn't have to worry about
• 12:31
  SPEAKER_02: going through those steps myself and almost think for itself,
• 12:34
  SPEAKER_02: which is kind of freaky and figure out,
• 12:36
  SPEAKER_02: okay,
• 12:37
  SPEAKER_02: this guy's this certain age.
• 12:39
  SPEAKER_02: I'm really not interested in what sports he plays.
• 12:41
  SPEAKER_02: I'm more interested in his health and the condition of his knees.
• 12:45
  SPEAKER_02: I can customize my gym package around that.
• 12:50
  SPEAKER_02: And maybe I want to offer him some stretch classes and physiotherapy.
• 12:55
  But if it was a younger person,
• 12:56
  maybe I would be offering them HIIT classes and
• 13:01
  maybe some invitation to our latest,
• 13:03
  you know,
• 13:04
  daytime disco dancing class or something.
• 13:06
  SPEAKER_02: So you could do very specific things.
• 13:08
  The AI agent would be making those decisions for itself and
• 13:13
  also taking feedback
• 13:15
  SPEAKER_02: and maybe refining it.
• 13:17
  SPEAKER_02: And when the answer came back to the gym owner using a keyword or a voice
• 13:21
  SPEAKER_02: interface,
• 13:21
  SPEAKER_02: not really sure yet how these things are going to evolve,
• 13:24
  SPEAKER_02: could say,
• 13:24
  SPEAKER_02: look,
• 13:24
  SPEAKER_02: nice try.
• 13:25
  SPEAKER_02: Can you please go back and modify it in these ways?
• 13:31
  SPEAKER_02: And it would do that,
• 13:33
  SPEAKER_02: but then not only would it do that,
• 13:34
  it would actually become meta because it would have learned
• 13:37
  from its mistakes.
• 13:39
  Rather than talking to a computer,
• 13:41
  it's like human to pseudo-human to computer.
• 13:44
  SPEAKER_02: Anyway,
• 13:44
  my thesis is that the same deep and narrow founder is in the
• 13:49
  best position to
• 13:50
  create AI agent itself or the correct interfaces to be able to
• 13:54
  do what I just
• 13:55
  SPEAKER_02: described.
• 13:56
  SPEAKER_02: I'm backing the founders who will have the
• 13:58
  the deep and narrow specialized skills within that vertical to
• 14:02
  transition their API
• 14:03
  SPEAKER_02: product to an authentic AI product.
• 14:06
  or at least become a service provider,
• 14:08
  which I think is going to become common.
• 14:09
  SPEAKER_02: I think Della said that the future of software is all agents.
• 14:14
  SPEAKER_02: I think he said that two weeks ago.
• 14:15
  SPEAKER_01: Yeah,
• 14:16
  SPEAKER_01: he said something in that any software that is basically a CRUD,
• 14:20
  SPEAKER_01: which is like the create,
• 14:22
  SPEAKER_01: read,
• 14:22
  SPEAKER_01: update,
• 14:23
  SPEAKER_01: delete sort of database interface,
• 14:26
  SPEAKER_01: that can be fully replaced by agents.
• 14:29
  SPEAKER_01: The logic makes sense because,
• 14:30
  SPEAKER_01: I mean,
• 14:30
  SPEAKER_01: historically we have had robotic process automation,
• 14:34
  SPEAKER_01: which is very simple rule-based.
• 14:36
  And it can be something as simple as if an email is received to
• 14:40
  this particular
• 14:41
  SPEAKER_01: email account,
• 14:42
  SPEAKER_01: automatically push it into my CRM.
• 14:44
  SPEAKER_01: It's an if-then sort of statement.
• 14:47
  SPEAKER_01: The majority of it has been very straightforward, rules-based.
• 14:51
  SPEAKER_01: Even on my email,
• 14:53
  SPEAKER_01: if I haven't received a response within five days,
• 14:56
  SPEAKER_01: send a follow-up email with the following text.
• 14:59
  SPEAKER_01: You can formulate that robotic process.
• 15:03
  SPEAKER_01: But when we start entering into this new era,
• 15:06
  SPEAKER_01: as you're describing it,
• 15:07
  there's a higher level cognitive order enabling it to actually
• 15:11
  think at multiple
• 15:13
  SPEAKER_01: levels along that process,
• 15:15
  SPEAKER_01: which may no longer be a linear flow.
• 15:18
  SPEAKER_01: It could end up being multifaceted, much more complex.
• 15:22
  SPEAKER_01: Yeah.
• 15:24
  There's an exciting time to be in the space,
• 15:25
  but it's still,
• 15:26
  as you say,
• 15:27
  deep and narrow.
• 15:28
  SPEAKER_01: So I'm imagining, and I'd like to hear your views.
• 15:32
  Does that mean we're entering the age of more verticalization
• 15:39
  and customization so
• 15:41
  SPEAKER_01: that we can enable that deeper,
• 15:43
  SPEAKER_01: narrow,
• 15:43
  SPEAKER_01: no longer one size fits all,
• 15:46
  SPEAKER_01: same CRM ERP for everyone,
• 15:48
  SPEAKER_01: no matter who you are.
• 15:49
  SPEAKER_01: Are we entering into that?
• 15:50
  SPEAKER_01: Or what is it going to look like?
• 15:54
  SPEAKER_02: Yeah.
• 15:55
  One of the things I always profess people with in terms of the
• 15:58
  era of AIs,
• 16:00
  I'm a pretty old guy.
• 16:02
  And I remember when the internet started,
• 16:04
  you know,
• 16:05
  the first browser came out in 1993.
• 16:08
  SPEAKER_02: Google wasn't even founded until 1998.
• 16:11
  SPEAKER_02: I always preface this with we're early stage people.
• 16:14
  SPEAKER_02: So in our portfolio, we try to invest in as many companies as we can.
• 16:21
  SPEAKER_02: You really don't know what's going to happen with the market,
• 16:23
  SPEAKER_02: with timing,
• 16:24
  SPEAKER_02: with luck,
• 16:24
  SPEAKER_02: with all kinds of things.
• 16:25
  But I think something like this is going to happen and I'm
• 16:28
  going to try to
• 16:30
  SPEAKER_02: make some guesses here.
• 16:32
  You're going to have the very large vertical players aren't
• 16:35
  going to go away.
• 16:35
  SPEAKER_02: So you're going to have the sales forces of the world,
• 16:38
  SPEAKER_02: the Google enterprise of the world,
• 16:40
  SPEAKER_02: even big behemoths like IBM and Oracle,
• 16:43
  SPEAKER_02: they aren't going to go away.
• 16:44
  But what you're going to have is a plethora of these deep and
• 16:48
  narrow vertical
• 16:49
  SPEAKER_02: plugins,
• 16:50
  SPEAKER_02: APIs,
• 16:51
  SPEAKER_02: authentic AI agents,
• 16:52
  SPEAKER_02: whatever you call them,
• 16:55
  SPEAKER_02: become the experts in their field.
• 16:58
  SPEAKER_02: And what they're really augmenting is services businesses.
• 17:02
  SPEAKER_02: So I see this as an equivalent of like a company that does maybe company
• 17:08
  SPEAKER_02: accounting,
• 17:08
  SPEAKER_02: like auditing a company account.
• 17:10
  SPEAKER_02: Right now it's done by human.
• 17:12
  SPEAKER_02: I think there'll be a lot of vertical AI agents that do most of the work.
• 17:16
  SPEAKER_02: There'll probably still be a guy at the end that signs off on it.
• 17:19
  SPEAKER_02: But it'll crawl through the balance sheet,
• 17:22
  the cash flow statement and profit and loss and come up with
• 17:26
  red flags and what
• 17:27
  SPEAKER_02: have you.
• 17:28
  SPEAKER_02: So the big guys are here to stay along with the big LLM.
• 17:32
  SPEAKER_02: They're obviously investing millions and billions of dollars into.
• 17:35
  But then you have these vertical players and there'll be
• 17:38
  thousands of them,
• 17:40
  SPEAKER_02: sort of like today's SaaS market.
• 17:42
  So today's vertical SaaS market will be replaced by vertical,
• 17:46
  authentic AI companies
• 17:48
  SPEAKER_02: that have their various specialities.
• 17:50
  SPEAKER_02: I think that's the way it's going to play out.
• 17:53
  That's an exciting time right now for founders who have
• 17:56
  specialist knowledge,
• 17:57
  SPEAKER_02: in particular vertical.
• 17:58
  I think the founders that I like the most are the ones that
• 18:01
  have got a real bee in
• 18:02
  SPEAKER_02: their bonnet about something.
• 18:03
  SPEAKER_02: They wake up in the morning and they have to fix it.
• 18:07
  SPEAKER_02: It's like, I can't believe this still works like this.
• 18:10
  SPEAKER_02: This is so stupid.
• 18:11
  SPEAKER_02: Let's go fix it.
• 18:13
  SPEAKER_02: So they will have a particular vertical special field that they work in.
• 18:18
  I invested in a company recently that does payment gateways for
• 18:21
  public transport.
• 18:22
  SPEAKER_02: Very esoteric field where you tap on tap off a bus.
• 18:26
  SPEAKER_02: But currently all that's done on premise hardware.
• 18:28
  SPEAKER_02: He's building the first world cloud-based system to do it.
• 18:30
  SPEAKER_02: He's in a great place to build agentic things like to tell your customers
• 18:34
  SPEAKER_02: LLM travel agent, please go buy me a train ticket to Penang.
• 18:40
  SPEAKER_02: And it will be able to just do that.
• 18:41
  SPEAKER_02: So I think things are going to play out.
• 18:44
  SPEAKER_02: You're going to divert the horizontal layer.
• 18:46
  Most of those guys are probably already either established or
• 18:51
  extremely well funded.
• 18:52
  SPEAKER_02: And then you're going to have thousands of these vertical players.
• 18:55
  SPEAKER_01: Do you see out of curiosity,
• 18:56
  SPEAKER_01: because I like what you're talking about,
• 18:58
  SPEAKER_01: where it's not necessarily the flashy side of things.
• 19:01
  SPEAKER_01: It's about solving real questions.
• 19:02
  SPEAKER_01: problems, getting down into the nuance of it.
• 19:06
  SPEAKER_01: But when we talk about horizontal versus vertical,
• 19:10
  SPEAKER_01: historically,
• 19:10
  SPEAKER_01: we've had a very thick horizontal layer.
• 19:13
  SPEAKER_01: Now,
• 19:14
  as that starts to thin out because of the availability of tools
• 19:19
  and resources that
• 19:20
  SPEAKER_01: we can now go deeper and more narrow,
• 19:22
  SPEAKER_01: the customization aspect of it.
• 19:26
  The couple of things that I question presently in regards to
• 19:31
  what does the future
• 19:32
  SPEAKER_01: hold is a little bit twofold.
• 19:34
  SPEAKER_01: One is those entrenched players are very good at doing generic use cases.
• 19:41
  SPEAKER_01: So we have that horizontal layer.
• 19:44
  But as we start having more prompt-based coding tools that can
• 19:50
  prompt to action,
• 19:51
  SPEAKER_01: prompt to build out agents,
• 19:55
  SPEAKER_01: So the twofold aspect is,
• 19:57
  SPEAKER_01: are more enterprises doing DYI solutions,
• 20:02
  SPEAKER_01: do-it-yourself solutions,
• 20:03
  SPEAKER_01: leveraging on that horizontal layer,
• 20:07
  SPEAKER_01: then do-it-yourself for the customization.
• 20:09
  SPEAKER_01: And then when the elephant walks the grasshopper's dance,
• 20:12
  SPEAKER_01: like the moment OpenAI or Google or Anthropic,
• 20:16
  SPEAKER_01: the moment their Dev Day conferences,
• 20:19
  how many AI companies all of a sudden are probably sitting in
• 20:23
  their offices saying,
• 20:24
  SPEAKER_01: ah,
• 20:24
  SPEAKER_01: well,
• 20:25
  SPEAKER_01: Maybe we're obsolete now.
• 20:26
  SPEAKER_01: How do you see the evolution?
• 20:27
  That's the point that I'm trying to get at is finding a
• 20:32
  defensible space when it
• 20:35
  SPEAKER_01: seems like barriers to entry are lowering.
• 20:38
  SPEAKER_01: Yeah.
• 20:38
  SPEAKER_02: I mean...
• 20:40
  I think about this a lot in terms of moats and a company like
• 20:44
  OpenAI can wipe out a
• 20:46
  SPEAKER_02: bunch of companies overnight with a new release.
• 20:48
  SPEAKER_02: Keeps me awake at night, obviously.
• 20:50
  SPEAKER_02: I think where the differentiation lies,
• 20:54
  SPEAKER_02: I think the barrier to entry,
• 20:56
  SPEAKER_02: is something that's unique in terms of its data.
• 20:59
  SPEAKER_02: So OpenAI can train itself on legal documents,
• 21:04
  SPEAKER_02: It has access to many, many legal documents.
• 21:07
  SPEAKER_02: So whenever I see a legal startup that's claiming to,
• 21:10
  SPEAKER_02: I can trade an NDA or whatever,
• 21:13
  SPEAKER_02: I think OpenAI will be able to do that very shortly.
• 21:16
  SPEAKER_02: That's not defensible.
• 21:17
  SPEAKER_02: But when someone is...
• 21:19
  SPEAKER_02: touching some real-world ecosystem,
• 21:23
  SPEAKER_02: whether that be data,
• 21:25
  like I'm working at a company right now that does electricity
• 21:28
  calculations based on
• 21:30
  SPEAKER_02: real-time information coming out of the grid.
• 21:32
  SPEAKER_02: OpenAI can't do that, right?
• 21:34
  SPEAKER_02: So something that's connected to the real world is a good example.
• 21:37
  SPEAKER_02: Something that's connected to IoT devices, like tapping onto a bus,
• 21:44
  SPEAKER_02: I look for some kind of ecosystem in the space of stable coin.
• 21:50
  SPEAKER_02: Tech stacks are very interesting to me right now.
• 21:53
  The things that open up or the incumbents aren't going to
• 21:55
  disrupt because there is
• 21:57
  SPEAKER_02: a defensible moat around either the data or the physical aspects or the
• 22:02
  SPEAKER_02: cryptographic aspects of that tech stack.
• 22:05
  SPEAKER_01: So I look at the ecosystem is very important.
• 22:08
  SPEAKER_01: Let me tie this back into the region where we're both living.
• 22:11
  SPEAKER_01: My geographic focus is a little bit more narrow than yours,
• 22:14
  SPEAKER_01: but speaking from this part of the world in APAC,
• 22:18
  Are you seeing any difference in how to drive success relative
• 22:24
  to Silicon Valley?
• 22:25
  SPEAKER_01: Obviously you lived in Silicon Valley,
• 22:27
  SPEAKER_01: you know,
• 22:27
  do you see a difference based upon the geography and what
• 22:30
  should founders here be
• 22:33
  SPEAKER_01: thinking about?
• 22:35
  SPEAKER_02: Yeah, it's definitely an interesting question.
• 22:39
  SPEAKER_02: Because I'm early stage, it's all about the founder.
• 22:42
  SPEAKER_02: It's all about their personality, their drive, their passion.
• 22:44
  SPEAKER_02: A lot of the fields I've invested in require deep knowledge.
• 22:49
  SPEAKER_02: I've tended to find companies in either Australia,
• 22:52
  SPEAKER_02: New Zealand,
• 22:53
  SPEAKER_02: or Singapore,
• 22:54
  SPEAKER_02: but I don't think that's necessarily the right way to think about it.
• 22:57
  SPEAKER_02: I think many founders in other markets in Southeast Asia,
• 23:01
  SPEAKER_02: Malaysia,
• 23:01
  SPEAKER_02: Thailand,
• 23:02
  SPEAKER_02: Philippines,
• 23:02
  SPEAKER_02: Indonesia,
• 23:03
  have very specialized skills in other areas like agrotech or
• 23:09
  even things like
• 23:09
  SPEAKER_02: health tech,
• 23:10
  SPEAKER_02: recycling.
• 23:11
  SPEAKER_02: I think there's a lot of potential here.
• 23:14
  SPEAKER_02: It's really a mindset.
• 23:16
  I'll give you an example about Australia and their breakout
• 23:19
  moment a few years ago.
• 23:21
  This company came out of Sydney called Atlassian,
• 23:23
  which most people probably identify with Jira.
• 23:26
  SPEAKER_02: Yep.
• 23:27
  That was a big deal because it created billions of dollars
• 23:29
  worth of wealth.
• 23:30
  And you had people coming out of that company and setting up
• 23:32
  other businesses.
• 23:33
  SPEAKER_02: Australia's now had big exits like Afterpay was bought by Square.
• 23:38
  SPEAKER_02: Canva, although it's still a private company.
• 23:41
  SPEAKER_02: A lot of wealth created there nonetheless.
• 23:42
  SPEAKER_02: A lot of wealth created there.
• 23:43
  SPEAKER_02: So there's quite a number of examples.
• 23:46
  So the ecosystem in Australia,
• 23:47
  the entrepreneur has crossed that mental threshold.
• 23:50
  SPEAKER_02: Yes, I can build a multi-billion dollar business.
• 23:53
  SPEAKER_02: I think Singapore,
• 23:54
  SPEAKER_02: to a lesser degree,
• 23:55
  SPEAKER_02: with Grab,
• 23:56
  SPEAKER_02: even though the Grab founders were from Malaysia.
• 23:59
  SPEAKER_02: But I think it's a mindset, frankly.
• 24:02
  SPEAKER_02: The world is a much smaller place now.
• 24:05
  SPEAKER_02: I mean, that's a bit of a cliche, but it's true.
• 24:08
  SPEAKER_02: But if you have a particular type of...
• 24:10
  SPEAKER_02: knowledge and access particular type of industry or data or ecosystem,
• 24:14
  SPEAKER_02: you can be living anywhere now.
• 24:16
  SPEAKER_02: I think that's not a barrier.
• 24:17
  SPEAKER_02: I think there's really a mindset and the support of the ecosystem.
• 24:21
  SPEAKER_02: Sydney,
• 24:21
  SPEAKER_02: Melbourne,
• 24:21
  SPEAKER_02: those cities in Australia have massive ecosystems,
• 24:25
  SPEAKER_02: a lot of great talent in terms of programmers and AI professionals.
• 24:31
  SPEAKER_02: Malaysia's lagging a little bit in that area.
• 24:35
  SPEAKER_02: Singapore relatively is doing okay.
• 24:38
  SPEAKER_02: The population of Singapore is about 6 million.
• 24:40
  SPEAKER_02: So, you know, Sydney, Melbourne's about the same, 6 million.
• 24:43
  SPEAKER_02: They also have that kind of mentality.
• 24:45
  SPEAKER_02: There's not a lot of examples coming out of Singapore either.
• 24:48
  I think a lot of successful Singaporean entrepreneurs perhaps
• 24:51
  moved to Silicon Valley.
• 24:54
  SPEAKER_01: But I don't think you have to.
• 24:55
  There's definitely a greater concentration of VC dollars
• 24:59
  sitting in Silicon Valley.
• 25:00
  SPEAKER_01: So oftentimes people see it as the pathway to raise rounds.
• 25:04
  SPEAKER_01: Yes.
• 25:05
  But certainly it's not a necessity in order to build a
• 25:09
  successful business.
• 25:11
  SPEAKER_01: I'm quite aligned with you on that.
• 25:13
  SPEAKER_01: Let's do a couple of wrap-up questions.
• 25:16
  SPEAKER_01: And I want to get one that's more like,
• 25:18
  SPEAKER_01: inspirational for the founders that may be listening.
• 25:22
  So if you had to do all the action for them,
• 25:25
  what should they be thinking about?
• 25:28
  SPEAKER_01: First and foremost,
• 25:29
  knowing what you know and based upon your vision for where the
• 25:33
  market is going.
• 25:37
  SPEAKER_02: If you're building a product, that's the pathway to success.
• 25:41
  That product has to be 10x better than anything else you're up
• 25:44
  against,
• 25:45
  not 2x better.
• 25:47
  SPEAKER_02: Think out of the box.
• 25:48
  SPEAKER_02: You've got to be very contrarian in a way.
• 25:51
  SPEAKER_02: Really,
• 25:51
  really got to search deep within your own knowledge and within
• 25:54
  your circle of
• 25:55
  SPEAKER_02: friends and finding that first initial customer or customers.
• 26:00
  SPEAKER_02: Your product has to be freaking amazing.
• 26:04
  SPEAKER_02: So amazing that people will go,
• 26:05
  SPEAKER_02: well,
• 26:05
  I'm just going to start using that because it's so much better
• 26:07
  than anything else.
• 26:08
  SPEAKER_02: That's the really call to action is don't think 2X,
• 26:11
  think 10X in terms of your mindset around the products that
• 26:15
  you're building.
• 26:16
  SPEAKER_02: You're not trying to create the next cardboard box.
• 26:19
  SPEAKER_02: You're trying to create the next way to do international bank transfer.
• 26:25
  SPEAKER_02: Look at a company like Wise, right?
• 26:28
  SPEAKER_02: Yeah.
• 26:29
  I mean,
• 26:30
  there's still some limitations,
• 26:34
  but better than before they existed.
• 26:36
  SPEAKER_01: International transfers remains a pain point, but certainly better.
• 26:40
  SPEAKER_02: Yeah, so my photo action is think...
• 26:45
  SPEAKER_01: 10x, not 2x.
• 26:46
  SPEAKER_01: Yeah.
• 26:47
  SPEAKER_01: Find real problems to solve.
• 26:49
  SPEAKER_01: It has to be 10x better.
• 26:51
  SPEAKER_01: And like the value proposition,
• 26:53
  SPEAKER_01: especially in the B2B space,
• 26:54
  SPEAKER_01: are you going to make somebody more revenue,
• 26:56
  SPEAKER_01: saving them more costs?
• 26:57
  SPEAKER_01: Ideally do both.
• 26:59
  SPEAKER_01: And I would lump efficiency gains into saving the cost of
• 27:02
  SPEAKER_01: an increasing revenue bucket as well.
• 27:04
  SPEAKER_01: And I love how you talked about starting niche, going deep and narrow.
• 27:09
  SPEAKER_01: You can,
• 27:10
  SPEAKER_01: you know,
• 27:10
  SPEAKER_01: land and expand,
• 27:11
  SPEAKER_01: find that wedge,
• 27:12
  SPEAKER_01: that deep narrow,
• 27:13
  SPEAKER_01: and then you end up building out from there.
• 27:15
  SPEAKER_01: This has been a fantastic conversation.
• 27:17
  SPEAKER_01: Look,
• 27:17
  SPEAKER_01: Chris,
• 27:18
  SPEAKER_01: where can people find you or Funk Ventures if they want to connect,
• 27:22
  SPEAKER_01: pitch you,
• 27:23
  SPEAKER_01: or just hear your thoughts further?
• 27:27
  Just hit me up on LinkedIn,
• 27:28
  Chris Burrell,
• 27:29
  LinkedIn,
• 27:29
  or email chris at funk.vc.
• 27:33
  SPEAKER_01: Fantastic.
• 27:33
  SPEAKER_01: Thank you very much again, Chris.
• 27:35
  A big thank you for joining us and making sense of the world of
• 27:39
  APIs,
• 27:40
  agentic AI.
• 27:42
  I've learned a lot in regards to deep and narrow where the
• 27:46
  trajectory is going and
• 27:47
  SPEAKER_01: look forward to hearing more and seeing how this journey continues.
• 27:53
  SPEAKER_01: Thanks, Kevin.
• 27:53
  SPEAKER_01: It was my pleasure to be on the show.
• 28:00
  SPEAKER_00: And that's a wrap for this week on Sea of Startups.
• 28:05
  If you are a founder placing your next bet,
• 28:08
  the answer might be deep,
• 28:10
  narrow,
• 28:11
  and agentic.
• 28:13
  SPEAKER_00: From a 2 billion funding rebound skewed by mega deals,
• 28:17
  SPEAKER_00: to Jakarta staking its claim as an AI compute hub,
• 28:22
  SPEAKER_00: to Malaysia's policy push to reboot its VC engine,
• 28:27
  SPEAKER_00: Southeast Asia's playing its next hand.
• 28:31
  SPEAKER_00: So next week,
• 28:32
  SPEAKER_00: we've flipped the mic again for our monthly roundup,
• 28:36
  SPEAKER_00: decoding what's been brewing across the region and
• 28:40
  the capital flows,
• 28:41
  the founder moves,
• 28:42
  and the infra-arms race reshaping the game.
• 28:47
  Follow us on Spotify,
• 28:50
  Substack,
• 28:51
  and of course,
• 28:52
  LinkedIn,
• 28:53
  where the real convos happen.
• 28:56
  SPEAKER_00: And hey, this message won't self-destruct, but your sass tech might.
• 29:03
  SPEAKER_00: See ya on the next one.