Latest / AI for Founders with Ryan Estes / Built a $50M+ AI Support Empire by Owning the Edges

Transcript

Built a $50M+ AI Support Empire by Owning the Edges · AI for Founders with Ryan Estes

0:00
Your customer support team is a liability, and the founders who
0:03
figure this out first are going to eat everyone else's lunch.
0:06
Every founder I talked to is hemorrhaging money on a
0:08
patchwork of tools. One for chat, one for tickets,
0:11
one for voice, one for workflows.
0:13
None of them talking to each other, and your customers feel
0:16
every single seam. The play that most founders miss
0:19
is that support is not a department, it's a data
0:21
flywheel. When you integrate every touch
0:23
point into one platform, something wild happens.
0:26
The AI handling 94% of your inbound questions starts feeding
0:30
every failure back into the system.
0:33
Every escalation makes it smarter.
0:35
Every bad interaction becomes training data, your support gets
0:38
better while you sleep, and your human team stops drowning in the
0:42
level 0 nonsense and starts doing work that actually
0:45
matters. So don't build 1 product.
0:47
Build the integration between products that nobody else wants
0:50
to build. That's where the value lives.
0:52
Not on the nodes on the edges connecting them.
0:55
The founders who own the edges own the platform.
0:58
So where in your business are you paying 5 vendors to do one
1:01
job and what would it feel like to own that integration
1:03
yourself? Welcome to AI for founders.
1:06
I'm Ryan Estes. My guest is David Kurandish, CEO
1:10
and founderofcapacity.com. This one's for founders who are
1:14
done with the tool graveyard and ready to understand how compound
1:18
startup strategy actually scales.
1:20
You'll learn how David thinks about AI support as a growth
1:23
lever, why integration is the product, and how we built a $50
1:27
million plus business by solving the problem behind the problem.
1:31
And the newsletter is at AI for founders.co.
1:34
This is where the real stuff lives.
1:36
Frameworks, case studies, and the AI moves founders are making
1:39
right now that nobody's talking about publicly yet.
1:42
Come get smarter every week at AI for founders.co.
1:45
And if this hit go, leave us a five star review.
1:48
One sentence from you is worth 1000 algorithm prayers for me.
1:52
And if you're ready to get your story on the podcast that
1:54
actually matter, go to kickcaster.com.
1:57
Kick Caster is a premier podcast booking agency that places
2:01
founders on top shows so you can stop pitching into the void and
2:04
start showing up in the right rooms.
2:06
They handle everything, so you just show up and be brilliant.
2:09
Go to kickcaster.com right now. That's Kick Caster by MO Burst.
2:16
All right, David, welcome to the show.
2:17
Thank you so much for joining me.
2:19
Thanks for having me here, Ryan. We betcha, man, let's dive right
2:23
into it. Your company capacity looks
2:25
extraordinary. Your history as an entrepreneur
2:28
is, from what I can tell, nearly flawless.
2:31
You know, maybe we could. I wanted to hear some crash and
2:36
burns doing research on you. I'm like, God Dang, this guy
2:39
does not lose. But before we get there, you
2:43
know, let's let's kind of open with with some of the news
2:45
that's kind of cutting right now, which is the world of claw.
2:49
You know, it's it's what, March 11th.
2:52
And so we're, we're kind of past the, the original open claw, I
2:57
mean explosion and now at all the major models are all kind of
3:01
grasping at different elements that they can incorporate into
3:04
their platform. So I'd love to hear your kind of
3:06
take on this whole kind of open claw agentic kind of explosion
3:10
over the last couple of weeks. Yeah, I think in terms of Open
3:15
Claw, the thing that they figured out pretty quickly was
3:19
that being able to have access to the full computer as opposed
3:24
to being relegated to an amp goes a long way.
3:29
I I built my first Open Claw app a few weeks ago where I am very
3:36
worried about my wife clicking on a phishing attack link.
3:40
And so I built a fish fry app where it's a little Chrome plug
3:45
in and anytime a link shows up that looks like it's a bad
3:49
hombre it'll either filter it out or do a little pop up or
3:53
some kind of way to keep that from happening.
3:56
It's not a full-fledged antivirus, but it's just a very
3:59
lightweight a lot way of keeping keeping link safe.
4:04
And the best part was I texted with my open claw bot to go
4:10
build the entire thing. I didn't write a single line of
4:13
code, literally just said hey, here's what I want to build.
4:17
And it did it. And so I think we are moving
4:21
toward a phase where this technology was being used to
4:30
effectively do a slightly better version of an autocomplete to
4:35
now it's it's truly taking on autonomous tasks.
4:39
And I think the thing that I saw the big jump from when Open Claw
4:44
launched from what previous folks had had done was just the
4:49
ability to keep grinding it out and not ask you every single
4:54
prompt. Now there are people pushing
4:56
back on Open Claw using up a gajillion tokens.
5:00
So I don't know that it's the most efficient platform for
5:03
building things, but it is very effective.
5:07
Yeah, 100%. And that's actually a really
5:09
cool use case because kind of in the classic McAfee security, a
5:13
lot of times the malware protection is actually the
5:17
malware building it, you know, for yourself.
5:22
You're like, OK, I know I'm definitely not going to hold my
5:24
own files for ransom, so. The thing I hated on these,
5:26
these antivirus suites that just slow your computer down, they
5:31
take forever and like 90% plus of the intrusions come from some
5:38
bad e-mail that somebody clicks on.
5:40
And then next thing you know, they got access to your
5:42
computer. And so I'm like, OK, I could
5:46
probably reduce the likelihood of getting hacked by 90% if I
5:50
put this on my family's. Computers, yeah.
5:53
And that's also kind of tapping into some of the, the fears
5:55
maybe kind of what holds me back from really being so gung ho
5:58
with open claw is that bad actors out there are using it,
6:02
you know, in to, to do exactly that, you know, to, to basically
6:06
send ransomware or just attacks in hacks across the board, you
6:11
know, and, and the, the, the ability to spin up users by the
6:15
10s of thousands, you know, assuming that like your, your
6:19
token management doesn't go completely nuts.
6:22
You know, if the balance sheet still works, that it, it, it
6:25
makes me tiptoe into the arena a little bit, you know, however.
6:31
The big thing I've been thinking about is how long till the best
6:38
till the model that can run entirely locally on your device
6:43
is good enough to build whatever you want.
6:46
So platforms like Open claw do a good job of connecting to
6:51
multiple models and kind of pick what you want.
6:53
But if you want to build something today, most of that's
6:57
still going to go with a larger foundation model.
7:00
Pick, pick your favorite. But I do think we are going to
7:04
enter an era at some point soon where there are models good
7:12
enough that can run on a souped up MacBook that just as good at
7:19
writing software. And then actually storing it
7:23
locally too, which is interesting.
7:25
So it's almost like a reversion back from the big, you know,
7:28
cloud revolution of whatever 1520 years ago that now you
7:32
might be able to just like completely self host, publish,
7:35
distribute your app from your laptop.
7:39
Yeah, it's getting, it's getting wild.
7:41
You know, the death of on premise has been greatly
7:46
exaggerated as more and more people are saying, Hey, we don't
7:49
want to put stuff in the cloud for this or that.
7:51
And at the same time, it's like, OK, I could use pick your
7:57
favorite framework to go build up Microsoft Office or Google
8:02
Drive competitor. But I, I think, I think I heard
8:06
somebody mention this a few weeks ago.
8:07
It's like, how many bug tickets do you think have been filed on
8:12
an app like Microsoft Office historically or even a newer
8:16
one, Google Drive or Salesforce? There's, there's probably
8:20
millions of bugs that have been filed, addressed, tackled.
8:26
And so yeah, is the juice worth the squeeze for something like
8:30
that? Probably not.
8:34
If you had a point solution for doing employee reviews, that
8:40
might be something where you could you could cut up a 80% is
8:45
good product in 20% of the time and a fraction of the cost.
8:52
This episode is brought to you by a dangerous idea.
8:55
It's called reading, specifically the AI for Founders
8:58
newsletter. 47,000 Founders strong now.
9:01
And it's the good stuff. No top ten AI tools that will
9:04
change your life written by someone who doesn't ship
9:06
product. It's just real breakdowns of
9:09
what's actually working in AI, who's raised, who failed, who
9:13
pivoted, and which founders quietly built $1,000,000 in ARR
9:16
while everyone else was arguing on X.
9:19
It's written for people building with AI Every week.
9:22
You get tactical frameworks, unfair advantages, breakdowns of
9:25
real companies, and the occasional existential question
9:28
about whether your startup is intelligent or just, well,
9:31
branded. It's free, It's sharp.
9:32
It might mildly offend your current workflow.
9:35
Go to aiforfounders.co and subscribe before your competitor
9:38
does, because the only thing worse than being replaced by AI
9:42
is being replaced by a founder who reads.
9:44
That's the newsletter at AI for founders.co.
9:47
Now back to the show. Yeah, I think this is the real
9:49
front. You're here because, you know,
9:50
for example, I spent a day kind of tinkering around trying to
9:53
create a whisper flow for myself and it was all right.
9:57
But then I was like, you know what, whisper flow is free, So
10:01
what am I doing? But you know, the, the idea is,
10:04
you know, of course, if you have your bot, you're texting with it
10:07
and you know, you're offs kind of offloading some of the like
10:11
creative responsibility where it can be watching you 24/7 and
10:14
say, Hey, I notice you make these, you know, 5 tab opens and
10:18
these entries into the spreadsheet.
10:20
What if we did XYZ? Because a lot of times, you
10:23
know, as much as I cherish my creative ideas, like the, the
10:26
more I open up and say, like, Hey, here's what I'm struggling
10:29
with. What are some options?
10:31
It, it, it gives you some really good options or at least leads
10:34
you to, to a solution you weren't looking at for, which is
10:38
just I, I love the analogy of like being a sculptor, you know,
10:42
if you're a painter, you start with a, you know, you start, you
10:44
have to add everything. But sculpting is pulling stuff
10:46
off and AI is really good about just feeding you ideas and
10:50
iterations until you're like, no, but I like that plus this,
10:53
and then you start to kind of get somewhere.
10:55
So it seems like the next step with kind of some of these, this
10:58
agentic support is it's recognizing your patterns.
11:01
It's recognizing what's possible.
11:02
That's probably outside of like what you've even considered.
11:07
Yeah, I really like that analogy of the the painter verse, the
11:09
sculptor. It's like that favorite quote of
11:12
mine. It's like perfection is not when
11:15
there's nothing left to add, it's when there's nothing left
11:16
to remove. I think it's very, very similar.
11:21
I also think most of the work people are doing today, 99% is
11:28
still someone has an idea and they're going back to some
11:33
language model of some sort and say hey, go build this for me,
11:37
go do this, go fix this, go edit this, etcetera.
11:41
I think where it's going to start to get really interesting
11:46
is when the proactivity becomes a much bigger portion of the
11:52
equation. So it's not I'm telling it what
11:55
to do and then it goes and does it.
11:57
It's coming back to me and say hey, I took care of this for
12:00
you. Yeah.
12:01
That's that's when this is going to enter the next inning.
12:06
That's the fun part. You know where your your agent
12:09
knows noticed that you have a bit of a temperature about 2:00
12:13
AM and so it's going to it's going to push back your meetings
12:15
45 minutes and give you a little extra time and some
12:17
recommendations on some electrolytes.
12:20
Exactly. You know, then we're getting
12:22
into the Jetsons terrain and it's, it's so exciting and fun
12:25
because, you know, it really feels like it's right around the
12:29
corner if it's not here already. You know, that's the other part
12:32
is there's always a reaching right.
12:33
But when you start to sit down and see what we're doing now,
12:36
you know, the IT, it's pretty extraordinary.
12:38
I mean, whisper flow is a great one.
12:40
I just, I kind of have a little note to like use it as much as
12:43
possible. And the difference is really
12:45
interesting. You know, one, I, I realized how
12:48
much I was kind of dependent on typing for my thinking.
12:50
It slows down your thinking a little bit.
12:52
You can hold a thread. It's a little bit harder to
12:54
speak to it. But also my mood during the day,
12:57
if I'm really dictating and, and not using the keyboard, my
13:01
posture is better. I feel more lucid.
13:04
It's just, it's these subtle shifts that like really is is
13:07
kind of extraordinary. You're saying we didn't evolve
13:09
to hunch over at A at a laptop like a velociraptor all day,
13:13
that that wasn't that wasn't part of part of our design plan
13:18
here? Yeah, exactly.
13:20
You know, I had an idea for a product I was kind of working on
13:23
for a while that was a keyboard that was like a little ball that
13:27
you would hold in your hands in your lap for, you know, 'cause
13:29
it's like when you're doing Tai Chi, you're making the Tai chi
13:32
energy ball. Like really, that's how we did
13:34
evolve, right, Is to pick up a big rock and throw it over there
13:37
on the on the rabbit or whatever.
13:39
So like, if our interface was all like right here, I guess you
13:42
can't see it in the, in the, I feel like it'd be way more
13:46
ergonomic. So, you know, and we're getting
13:48
to that point, you know, in the in the beginning it was like,
13:50
oh, this has to be some kind of hardware component.
13:52
But surely enough, I mean, as soon as like people accept
13:56
wearing computers on our face, I mean, we're not there yet that
14:00
you could easily, you know, have some kind of projection that's
14:02
watching your hands in space in a much more ergonomic kind of
14:05
place. So, you know, we are moving
14:08
towards like a beautiful experience.
14:13
I think with with technology, at least I for the sake of this
14:16
conversation, I love talking about that because there's also
14:19
the flip side. It's it's all too slippery to go
14:22
down the scary stuff. But you know, really kind of
14:25
envisioning the beauty and and kind of the elegance that can
14:29
evolve from all of this is is pretty extraordinary.
14:32
Yeah. I mean, I think about it in
14:34
terms of, if you look at the history of computing, right?
14:39
We started out with physical machines that then got
14:44
represented in terms of zeros and ones.
14:47
And then it was like, wow, we've got this digital representation
14:51
of what used to be physical switches.
14:53
And then you had assembly language and it was like, wow,
14:56
this is a major breakthrough because we're not doing zeros
14:59
and ones. It's a higher order kind of
15:00
thing. And then you went to programming
15:03
languages like C and C++ and it was like, wow, I don't need to
15:06
know like memory registers for most of this code is, you know,
15:10
then you get in higher order, you're doing like HTML and
15:13
JavaScript and we're building apps and these types of things.
15:17
And now, I mean, it's getting to the point where people are using
15:21
natural language to go build what they want and it's
15:25
completely abstracted away from the underlying language even in
15:31
many cases. And So what I've seen is that
15:35
almost every jump in language interface has opened up an order
15:41
of magnitude number of new people who can build software.
15:48
And so now going toward a natural language interface,
15:52
we're at the point where you don't have to know all the
15:54
insurance and outs and intricacies, as long as you can
15:56
explain your idea credibly and tell it to go build the test
16:00
cases, which is very important, Anyone, anyone's going to be
16:04
able to build software soon. Yeah, and it it democratizes
16:09
like software in extraordinary way, which means we get to
16:12
capitalize on the the billion people in India or take wherever
16:17
that they've got cell phones. I mean, like once we can do
16:19
everything locally on our Mac mini, it's only a couple of
16:23
steps before we can do everything on our phone, you
16:25
know, and now you have kids that never had access to, to
16:28
opportunity that, that are brilliant, You know, that, that
16:32
we can really all kind of capitalize on on, you know,
16:35
their mental capacity. I mean, it's, it's pretty
16:38
extraordinary. It's wild because you think
16:40
about there is there's simultaneously more
16:47
democratization of this technology than anything else
16:50
we've ever seen. Wasn't like when the industrial
16:52
revolution happened. It was easy to go start a
16:55
factory if you were, if you, if you didn't have means right now,
17:00
your point, You can pull up an iPhone open clause, some credits
17:04
and you you can go build something.
17:06
And there's also the potential for a lot of wealth
17:09
concentration in the hands of some of the biggest corporations
17:13
in the world. So it's so it's going to be
17:14
really interesting to see what plays out economically.
17:17
Does the democratization win? Do the large corporations win?
17:21
Or is it a little bit of both? Yeah, that's, that's a big
17:26
question, you know, because, you know, that we talked about the
17:28
that Gilded Age and all those giants and those guys are
17:33
nothing compared to the guys we got now.
17:36
You know, the the amount of wealth that Elon and Bezos and
17:39
these guys have accumulated is it is second to none
17:42
historically, you know, so it does kind of beg the question of
17:46
like, OK, well, how does that help everybody else?
17:50
You know, what are we going to do if we're democratizing
17:53
technology? But the foundation models,
17:55
there's only four or five guys that really are, you know,
17:58
getting the bottom line of all those tokens.
18:00
How do we kind of effect that change?
18:02
And it's I, I typically don't get into politics or current
18:05
events, but it's, it's germane to the conversation, which is
18:07
kind of the wealth tax that's going on in California where
18:10
people are starting to propose kind of these one time wealth
18:13
taxes. Now Elon's out there stoking
18:16
this fire saying, hey, you're planning for the future.
18:20
Don't you're investing in your retirement.
18:23
You don't need to do that. You know, there's plenty for
18:25
everybody with the assumption there's going to be some kind of
18:27
Ubi or there's going to be some kind of wealth dis disbursement
18:31
or, or something, which means somebody's got the money and
18:34
someone's given the checks. And what I know about people is
18:37
they're going to be looking over their shoulder at your check and
18:39
be like, why is my check different than your check?
18:41
You know, so when you think about the future of all this,
18:43
particularly, you know, with software and I, I have a kind of
18:46
way to dovetail the capacity because I want to ask you about
18:49
that too. But what's your take?
18:51
You know what, what's the, what's the end game here?
18:53
10 years. OK, so in in no particular
18:58
order, Elon is obviously welcome to his own opinions and every,
19:05
every every can have their own opinion on these things.
19:08
Personally, I wouldn't bank on Ubi producing the results that
19:14
people are are talking about just yet.
19:18
I think there's a lot societally that we have to work through
19:21
before that becomes a reality. Number one.
19:24
I will say though, you know, hold hold any political
19:32
discussion of this aside, somebody like a Michael Dell put
19:36
in the other billions of dollars of call it Ubi light programs.
19:43
Let's see how let's see what happens there.
19:45
Like that's it. That's an interesting test case
19:48
test pancake to go study and see what happens to that cohort of
19:56
folks. So I think that's a really
19:57
interesting experiment. He put he put his, you know,
20:01
again, I'm not like the biggest Michael Dell fan or anti Michael
20:04
Dell. I don't I don't know him.
20:05
I don't know him personally, but he put his money where his mouth
20:08
is in terms of, Hey, you want to do Ubi?
20:10
OK, let's let's go try. Let's go try it and see.
20:15
So I'll put that aside. I would also say politically
20:23
there are, I think we want to be a country that rewards people
20:28
taking risks and building things.
20:33
I think if we get rid of that, that is going to be a huge dent
20:39
into one of the OR underlying principles of why the United
20:45
States has over many, many years done done very well.
20:51
So I think if you asked, you know, 10 people in the street,
20:55
most of them don't think you should be punished for being
20:59
successful. I I think the bigger flip side
21:02
is I don't know how many kids Bezos has.
21:07
I know Elon has a lot of them, right.
21:09
But if you're at a Musk junior or a Bezos junior and you didn't
21:15
build it, I'm not saying you can't leave an inheritance for
21:18
your kids, but to me, that's what I would be going after is
21:23
inherited wealth for people who didn't build it.
21:27
You want to leave billions of dollars to A next generation who
21:29
didn't, who didn't earn it, didn't build it.
21:32
Personally, I think that should be taxed heavier.
21:35
And then entrepreneurs who take risks and build companies and do
21:40
things that require a lot of, a lot of stepping out of their
21:48
comfort zone. I, I think they should be taxed
21:51
lower. And I think unfortunately for
21:53
our political climate, that policy or that, that that
21:57
concept would not fit well in either party.
22:01
My my friends on the left would be appalled at the idea that
22:05
Bezos or Elon Musk's tax rate should be low.
22:11
And then conversely, my friends on the right, who would be
22:14
appalled at the idea that their kids should get a, a very large
22:19
inheritance tax if they inherit, you know, 10s of millions,
22:23
hundreds of millions of dollars. Yeah.
22:25
Oh, you could see the headlines be like, oh, there's going to be
22:28
there's going to be tribe of people looking to assassinate
22:31
the the billionaire guy so he can collect, you know.
22:35
Yeah. I mean, it is an interesting
22:37
idea because then you could also borrow against that.
22:39
But of course you're like looking at valuation over time.
22:41
It's hard to consider, you know, and maybe the, the, the tough
22:45
thing is, is that there's not a lot of conversation around this,
22:48
largely because everything is so politicized and so polarize that
22:52
everyone if, if somebody has an initiative, they're, they don't
22:54
like it because it's blue or they don't like it because it's
22:56
red. And there, there's, there's so
22:58
much nuance here. You know it's it's.
23:01
Ryan it's like taking a pros like I'm saying here, if I I
23:05
said if I pulled 100 people on the street and said, hey, let's
23:08
we can figure out what the numbers are.
23:10
But conceptually, people who take risks should have lower tax
23:14
rates. People inherit, inherit most of
23:18
their wealth should have high, pretty high tax rates.
23:21
What do you say about that? I think it's pretty good. 70% of
23:26
people agree to that, 75% of people who agree to that.
23:29
And you'd have the the rest of your 15 on hard left would say
23:33
this and 15 hard right would say no to that.
23:36
But this isn't, this is like one of those middle of the ground,
23:39
middle of the road proposals that I think most people would
23:43
be like, yeah, that that kind of makes sense.
23:46
But we're we're in a climate where that's difficult too, to
23:50
push something like that for it. Yeah, or even to even talk about
23:53
it. You know, no one's really
23:54
talking about it. If the end game is, you know,
23:56
you ask all the guys, you know what, what is it?
23:58
Well, we'll probably have to have some Ubi, but it's like,
24:00
it's like, boy, that's that event Horizon is quickly
24:04
approaching, you know, and that creates a lot of problems.
24:07
But there's a lot of nuance within that, you know.
24:09
So like, to your point, you know, you can't de incentivize
24:13
people because some of us, we don't want lopped ears.
24:16
You know, we want the the ears that stick up.
24:18
We want to stay wild, you know what I mean?
24:21
But but but also, you know, maybe that that's not the first
24:25
layer of the conversation that OK, we're replacing income and
24:28
jobs with some kind of Ubi check.
24:32
You know, that that's probably not the first one.
24:33
But to your point, like I love those two examples, People that
24:36
take risk pay lower, people that inherit pay higher.
24:39
And then, you know, what are some other things that we can
24:41
agree on? Well, there, there shouldn't be
24:43
children that are suffering for lack of healthcare or there
24:45
shouldn't be people that lose their entire life savings
24:48
because they have to pay it on health expenses.
24:51
Where we do have, you know, Medicare and Medicaid for, you
24:54
know, it's downside being so big, it's pretty darn good
24:57
considering how big it is. You know, So there's ways that
25:01
like they could recapitalize some of this wealth in existing
25:05
systems. It's not perfect, of course,
25:07
Doge didn't clean all the houses.
25:11
But you know, even approaching that feels like divided in like
25:15
in contentious, you know? Yeah.
25:17
I mean, you look at Social Security as another example.
25:20
I mean, we can argue about what the benefits should be, what age
25:26
is it should be, but the math is the math like it's, it's not
25:31
going to last forever with the current population demographic,
25:35
you know, timeline. So you got to either lower
25:38
benefits, raise revenue, raise taxes, or raise the age in which
25:45
the payouts can have. Like, it's just a math.
25:48
We can argue about what the solution should be, but there
25:50
are people that don't even want to accept the fact that there's
25:56
a problem. Yeah, 100%.
26:00
It'll be interesting. You know, I, I, I think that
26:03
there's, there, it's a little bit slower when you're getting
26:06
to the Senate floor, particularly because the folks
26:09
in the Senate just probably aren't on X everyday pulling
26:13
their hair out about what they can do with open claw.
26:16
It'll get there. This is unsubstantiated, so
26:19
don't vote me on this, but I had someone tell me that one third
26:25
of our senators today don't use e-mail there either.
26:31
Staffers will print out the emails for them and hand them to
26:35
them. And again, not substantiated
26:38
that. Don't, don't put that as the
26:40
headline here. But do you want those people
26:44
making AI policy? Really.
26:47
Like just thinking about that for a minute.
26:48
That just seems, I'm not saying we don't need guard rails and we
26:52
don't need AI policies. I'm not saying that at all, but
26:57
the, the gerontocracy has to go either side.
27:01
I, I don't care. This is the non, non partisan
27:03
issue. Like if you're a pilot, you have
27:06
to retire at a certain age. We should, we shouldn't have
27:09
people in their 80s staying on and being senators for 50 years.
27:14
I think that's either party. There's not this, you know, I'm
27:17
not ageist like great elder statesman, awesome stateswoman.
27:22
But there are many things in our system that need to change.
27:26
Yeah, no, I couldn't agree more. It's it's funny too to see like
27:29
the European Parliament's when they're convening, it looks like
27:32
a frat party and they're all in their nice suits.
27:35
Everyone's like 20-30, you know, they're, they're younger so that
27:39
you know, you can probably the vitality is, is there.
27:43
Not that they don't have their own problems, but I, I totally
27:45
agree there. It is quite a morass of a
27:48
situation we have legislatively, particularly when we really need
27:51
some like pretty sophisticated folks.
27:54
You know, I I think, you know, Trump has Trump is Trump, but
27:59
you know, he did bring in the second the second term, a lot of
28:02
like field experts into his cabinet that I thought was like
28:05
really cool. You know, it seems like kind of
28:07
the spirit of like the the original constitution that you
28:12
have people that have found success or a found network and
28:16
and found influence. And then they dedicate some of
28:19
their time for public service. And you know that they're in
28:22
there four years, six years, eight years maybe.
28:24
But then they're back into the the private sector and there
28:28
it's not kind of like this, you know, career politician
28:32
situation. We've got now on both sides
28:34
where they really do benefit by having on being either, you
28:37
know, the incumbent or, you know, the opposition and it just
28:41
switches roles. But they're always the same
28:43
people there. You know, it can get a little
28:45
frustrating, but you know, you who knows, man?
28:47
The, the, the, everything's changing so fast.
28:51
You know how midterms will go and and everything else we'll
28:54
kind of see. Yeah, I think, I think I've hit
29:00
my my political limit for one one friendly podcast here, but.
29:05
Amen. I do think what what's been
29:10
interesting is, you know, my wife is a physician by by
29:16
background. She's now a practicing
29:18
therapist. And the the vaccine skepticism I
29:22
I think is really dangerous in many ways.
29:25
Yeah, but the food pyramid, yeah, telling people to eat real
29:29
food that you like, forget your healthy food versus unhealthy
29:34
food. Like, even before we get to
29:35
that, let's start with real Food First, not real food.
29:39
Like, like that's a huge win. And so I just, you could go
29:43
through all these and you got some policies that have done
29:46
well, some that haven't. There's a lot of character
29:48
issues, I think, in the current administration, but I do think
29:52
there are some wins there as well.
29:53
Yeah, 100% the food pyramid. Remember those commercials for
29:57
like, Frosted Flakes would be like a healthy part of a
29:59
balanced breakfast. And it shows the biggest feast
30:02
you've ever seen. It's incredibly like, what am I
30:06
drinking at? The low fat craze, the Alestra
30:11
or whatever it was. Cool man, I want to I want to I
30:16
love that rabbit hole or that was really fun to go down.
30:20
But I want to talk a little bit about capacity and, and
30:22
particularly I think it plays into what we were talking about,
30:25
how we're starting to figure out what the new UI of technology in
30:29
humans is, You know, as it becomes smaller and kind of it
30:32
gets littler. I was watching Zoolander the
30:35
other day and just the favorite joke on there is like how small
30:38
their phones are. But we're kind of like shrinking
30:40
again. We got really big and now it's
30:42
kind of coming back and we're trying to rediscover like how
30:44
we're going to communicate each other with each other with these
30:46
tools. So I'd love a a brief
30:48
introduction to capacity and kind of what you guys are
30:51
working on over there. Yeah.
30:53
So let me start out a little bit of background.
30:54
So serial entrepreneur, been in the tech space my whole career
30:59
and prior to this, I had started a company called announced media
31:02
vertical search portfolio websites.
31:05
And we began by adding Q&A to the individual sites, which led
31:10
us to go acquire a company called answers.com.
31:13
We took that private, merged it with the company I started, grew
31:16
that, got it to a great exit. And then I had this existential
31:20
crisis of what in the world I'm going to do next.
31:23
So I was reflecting back on my, that was, but most of my 20s
31:27
basically we're putting that company.
31:29
So reflecting back on my time at answers.com and I thought about
31:32
how there was a small segment of questions that were very, very
31:40
commercial in nature. They were very like
31:43
transactional at high value. And there was a whole long tail
31:47
of like, should I break up with my girlfriend?
31:50
What's the best animal at the zoo and stuff and stuff like
31:54
that. And so I, I, I liked this idea
31:57
of whatever I was going to do next.
31:59
I wonder if I could focus on some of these most commercial,
32:03
most important high dollar value questions that people have.
32:08
I also like the software as a service business model toward
32:11
the tail end of my time. But at that company, we had
32:13
gotten to about almost 5050 between advertising revenue and
32:17
SAS revenue. So I like, I like subscription
32:19
revenue models. And then I also have this
32:21
passion to see people do their best work.
32:24
Like there's a reason people pay you to work like it.
32:30
It's you're going to have bad days.
32:32
You're going to have everybody's got days they don't like.
32:35
But when you see people that are working out of their gifts, out
32:37
of their talents, out of their longings, out of the things they
32:40
desire and the things they love, it doesn't feel like work or it
32:44
doesn't feel as much like work. And it feels like you're in a
32:46
flow state or you're in a, you're in a good place.
32:49
And so I had all this kind of bobbing around deck of my head.
32:52
And then that Christmas time of 2016, the top selling on Amazon,
32:57
top selling item on Amazon, it wasn't a book, it wasn't a video
33:01
game, wasn't a toy. It was Amazon's Alexa.
33:06
And so that was kind of a watershed moment for me where I
33:08
was like, OK, the top selling product in the world right now,
33:13
consumer product is an AI that's, albeit a very
33:17
rudimentary one at that time. I think this technology is going
33:21
to have an even bigger impact on the world of work.
33:25
So we started capacity in the beginning of 2017 and it kind of
33:28
pulled, it's like a like that movie Slumdog Millionaire, where
33:30
the whole game show is the sum of all of his life experiences
33:34
into one place. When I started it, I pulled from
33:37
my experience at answers.com, the SAS background, his desire
33:41
to see people do their best work, this rise of AI, all of
33:46
that coming together. And we, we started the, the
33:49
company initially with the idea that you could have an AI
33:53
internally to help support your team.
33:55
So think HRIT, password resets, onboarding questions, all that
34:01
sort of stuff. And when we got into it, it was
34:05
one of those things where right idea, but we didn't have the
34:09
right product market fit initially.
34:12
And the big thing that we had to pivot on was that mid market to
34:17
small companies had this problem, but they didn't have
34:20
much of a budget. And then mid market to
34:23
enterprise companies had the budget, but needed a bunch of
34:26
enterprise Y things that we didn't have at the time as a
34:29
small start up. I didn't have SoC 2, I didn't
34:31
have HIPAA compliance, I didn't have role based access controls
34:35
and all the all the goodies. And so we got those now and we
34:38
built that now and got a big portion of that.
34:40
But the just getting in the game in AI, enterprise AI allowed us
34:48
to see that there was a huge need not only to support your
34:50
team, but also to support your customers.
34:52
And that started digitally first with chat bots and virtual
34:56
agents. It's since gone Omni channel.
34:59
And so if I Fast forward to today, capacity is about 20,000
35:03
customers, roughly 450 enterprise, big names for folks
35:08
like that you'd recognize like a T-Mobile or a Verizon or a Nike
35:13
or American Express. And then at the same time, we
35:16
worked with a lot of mom and pops, life coaches and spas and
35:20
salons and car dealers and folks who might want to automate, you
35:23
know, one part portion of their flip.
35:26
We're building this company with the idea that automating support
35:30
is one of the best things you can do, both for the customer
35:33
experience, but also to free people up where you can either
35:37
get people working on higher order tasks or companies can
35:40
pocket the savings. And that's what we provide back
35:43
to these companies to go decide. Incredible.
35:46
And when you're talking about Omni channel, is this like video
35:48
chat, voice chat, regular chat, text messaging, like audio
35:54
calls? What?
35:55
What is Omni? Yeah.
35:57
So you could be on a website chatting with our virtual agent,
36:00
You could be texting. We've got plug insurance with
36:03
WhatsApp. You could be on a call.
36:06
We can sit in front of your contact center software.
36:09
We can be plugged into your contact center software like
36:11
RingCentral or Nice and Contact that we can be on your intranet
36:16
internally, we can plug into Slack and teams for your
36:20
internal side, really anywhere where you get an e-mail, a
36:24
ticket, a phone call, an inquiry, we can be there.
36:27
But I'll, I'll say Ryan, one of the things that we, well, OK, so
36:32
I've been, I started this in 2017.
36:35
It's been a great run, had a ton of fun.
36:39
And I can say we, we definitely predicted that AI was going to
36:45
be big. And you got to remember, it's
36:46
hard, it's hard to remember now. But in 2016 when we were coming
36:49
up with the idea, people were like, blockchain is going to
36:52
dominate and AR is going to dominate and VR is going to
36:55
dominate. AI was, was maybe 4th on the
36:57
list of what people were talking about.
37:00
And I had a strongly held conviction that this technology
37:05
would be the defining thing probably until quantum really
37:12
goes full, you know, full deployment of quantum right
37:18
between Internet and quantum AI will be the will be the next,
37:21
next big thing. The other thing though, that I
37:25
think people get, people get a little tripped up on is that I
37:29
think about AI and for the practitioners on this call, I
37:33
think about AI in terms of two different types of problems you
37:35
can solve. You've got what I would call
37:39
salad problems. So I make a great Caesar.
37:45
I call it a Cyrus salad because my dad's from Iran, but it's a
37:48
Caesar style salad. It's got anchovies and it's got
37:53
fresh kale and I make my own croutons.
37:56
It's got garlic, all the things. It's so good.
37:59
So it's an amazing, if you ever come to Saint Louis, Ryan, come
38:02
to my house. I will make you my Cyrus salad,
38:04
right? So change your life, right?
38:07
It's so good. You could be out of one of the
38:11
ingredients and it's still going to be pretty good.
38:13
Maybe with the exception of the, you can't miss the lettuce,
38:16
right? But with the exception of the
38:18
lettuce, you could remove any one ingredient.
38:20
It's still going to be a really good salad, right?
38:22
80 or 90% of the benefit. But then there are other
38:26
problems that are more like what I would call brownie problems.
38:28
So I've got 4 kids at home and if they make brownies and they
38:34
miss one ingredient just slightly, they're not like 80%
38:39
is good or 90% is good. That entire batch of brownies is
38:43
in the trash. And so the thing I liked about
38:46
using AI to automate support is when we were starting out, I
38:50
felt like this is very much going to be a salad problem
38:53
where even if we can't automate all of support, if you can take
38:57
out half of the calls or 60% of the tickets or 80% of the live
39:04
jets, you can still deliver a lot of value in the interim as
39:09
you're working through those last stages first.
39:11
Other problems are more brownie problems, or if you don't solve
39:16
the entire problem, that batch is going to be the trash.
39:21
Love that analogy too, largely because if you don't a salad
39:27
without croutons it's not as delicious but you still you
39:31
still have the salad. But like to your point, like if
39:34
you don't have all the greens for a brownie, not only is it
39:37
not as good but like it's almost not even a brownie anymore.
39:40
It would like pudding or something, you know, it like
39:43
becomes something else so that there's maybe a, a, a point
39:47
where the AI products are also like that because you know
39:50
what's beginning as support, you know, online support, which is a
39:55
little text thing or it's got video or it's, it's basically
39:58
meeting where people where they are like, is it really support
40:02
or is it something new? You know, and, and to me, the,
40:06
the, the more sophisticated that these, you know, chat support
40:10
has gotten. It's fantastic, you know,
40:12
because in 2017 it must have been talk about like support
40:15
tickets. People get really upset because
40:17
you get in phone trees and you're like, talk to the
40:19
manager, talk to the manager. But but now, you know, and
40:23
that's credit to you of seeing where the puck was going
40:25
because, you know, back then it was like, man, Alex can't do
40:30
anything for me or, you know, but but now they really can.
40:34
So when I'm experimenting and I find like a good one, you know,
40:38
you can test the parameters and it's amazing.
40:40
You find yourself actually more engaged with with chat support.
40:44
You, you have all the patience you want.
40:46
You don't have to have all of the the politeness that we have
40:51
in in conversation. You know, like if I'm drolling
40:54
on for too long and be like, I better shut up and let this guy
40:56
get get on with this day. But you don't have that.
40:59
So it it almost, it becomes a new thing where it's like, OK,
41:02
this is like a portal into the entire knowledge base of the
41:05
company. Like, how complicated can I get
41:08
the questions? And you know, my favorite
41:10
question is like, hey, what should I be asking you right
41:12
now? You know, you know what my
41:14
intentions are. Like what else?
41:16
Tell me what I should know. Well, right, to jump back into
41:21
2017, we were just getting this idea off the ground.
41:24
You know, the technology wasn't there yet.
41:27
It was urging and it was improving and there was a lot of
41:32
good stuff going on, but we recognized pretty early.
41:36
So like some stats today, when we started in 2017, if you asked
41:43
our AIA 100 questions, we would respond to 55 of them and ten of
41:48
those would get thumbed down in the feedback loop.
41:51
And so my joke was 45% of the time it works every time sort of
41:54
thing. But Fast forward to now, I mean
41:58
REI is answering between 92 to 94% the questions you throw at
42:01
it, less than 1% get thumbed down in feedback if it's
42:05
improving every single day. And So what I what I would say
42:08
is we recognized early on that the linkage between the virtual
42:17
agent and the human escalation. This shouldn't be two separate
42:23
systems. The tight coupling of the
42:27
virtual agent handling the level 0 support what used to be level
42:30
1. Now it's automated.
42:32
But a clear escalation path and an escalation path to the right
42:37
human agent at the right time with the right expertise.
42:40
And critically being able to use AI to help that agent spend less
42:44
time. And then critically, making sure
42:46
that that human agent, what they do isn't in a silo that works
42:51
its way all the way back into the virtual agent.
42:55
If you can make that happen and close that loop and this thing's
42:59
going to thing's going to learn really quickly, even beyond
43:05
improving the underlying model or adding more training data,
43:07
etcetera. Incredible.
43:10
And you know, starting in 2017, flash forward what, we're 7-8
43:15
years into the company and you have 20,000 customers.
43:19
So talk about your approach to go to market and what kind of
43:22
what you know accounted for like that rapid growth?
43:26
Yeah. So a couple years ago I kept
43:32
hearing some version of the same thing over and over again, which
43:36
is David, we are so tired of trying to duct tape together
43:42
these different solutions. We don't want to go to 1 vendor
43:45
for a help desk and a different one for our work flows and one
43:48
for a virtual agents and totally different one for a call QA and
43:50
one for our outbound campaigns and completely different one for
43:54
our translation and support. Like could, could you just put
43:57
all these things together into one integrated play?
44:03
Because we as a customer, I don't, we don't have a team of
44:05
integration specialists go make this happen.
44:08
So I heard that once, I heard it twice, I heard it about a dozen
44:11
times. And I'm like, OK, I think we
44:13
need to go double click on this and kind of map this out.
44:17
So what we did is we mapped out 24 steps to the customer
44:21
experience journey, starting with the inbound agent
44:24
deflection, going to the agent assist escalation and cutting
44:29
down handle time, going to post interaction work flows and then
44:33
finally outbound campaigns. So those are the kind of the
44:35
four product lines or groupings. And I think at the time we were
44:41
I'm going to see maybe in two or three of those areas.
44:45
And so I went back to my executive team and I'm like,
44:47
guys, this is the map. I'm so excited.
44:51
This is what we need to go do. Here's our strategy.
44:55
And they looked at me like I had three heads.
44:56
They're like, David, how the are we going to pull this off?
45:01
And then number one, and then #2 is anybody else doing this?
45:04
Is there anywhere else we can draft off of?
45:06
So on the second part, I came across some of the work from
45:12
Parker Conrad Parker's the CEO and Co founder of Zenithets, now
45:16
CEO and Co founder of Riplin. We're a Riplin customer, happy
45:20
Rippling customer on. He's got this concept he calls
45:24
compound startup. And the basic idea of a compound
45:27
startup is that you take what used to be a bunch of disparate
45:31
point solutions. You put them together, you lower
45:37
the cost of the skew, but you maximize the price and the take
45:41
rate of the bundle. And then crucially with AI, you
45:47
become integration as the product and that becomes the
45:50
core of what you do. And so he's doing that in the HR
45:54
space. We looked at that and said, OK,
45:56
I think we can go do that in the customer support space.
45:58
I think we can do that in CX space.
46:01
And so we, yeah, go over that again.
46:04
Let me see if I can follow you again.
46:05
So basically just explain that to me I guess.
46:08
Yeah, so super high level, there are all these point solutions.
46:11
So 2026 there's a point solution for for everything.
46:15
Companies are tired of putting them all together themselves.
46:18
They want to go to fewer vendors that are more integrated, that
46:22
can do more things. And so this compound start up
46:25
strategy, you take discreet point solutions, you put them
46:27
together, the individual skew, the individual product cost
46:31
actually goes down the take rate of the bundle, you have people
46:35
using more of your products overall.
46:37
So your margins should actually go up over time.
46:40
And then crucially, you be integration as the core thing
46:46
that you do and what I mean by that.
46:48
All the other points together into one unified place.
46:52
Yeah. So think of it like this, the
46:55
difference between a product suite and a platform.
46:59
In a product suite, you have multiple products, but they're
47:02
very loosely coupled, like using that we were talking about
47:05
Microsoft Office earlier. If I use Office or I use Google
47:09
Drive, yeah, I might copy a spreadsheet from Excel into
47:15
PowerPoint or from G Slides into or sorry, Google Sheets into
47:22
Google Slides. But they're pretty they're,
47:25
they're, I would not consider that to be a platform.
47:27
That's it. They're, it's a product suite
47:29
where there's some light integration, but in a, in a
47:33
platform. So in a in a product suite, the
47:37
value primarily lives on the node of what the individual
47:42
products do. In a platform, the value
47:46
primarily lives on the edge that connects these different pieces
47:50
together. So in our world, if we have a
47:53
virtual agent that can go handle 90% of your calls, but then it
47:56
needs to be able to escalate to a support person, that support
48:00
person should be able to have AI to help them.
48:03
That might kick off a downstream workflow that happens
48:05
asynchronously. That workflow might send an
48:08
outbound message back out to the customer who may come in and
48:11
leave a survey or are reading the review or, or some kind of
48:14
feedback that then triggers some additional loyalty point like in
48:19
that case, the the solution or the, the, the, the, the, the
48:25
customer facing interface. It's not about the note, It's
48:28
about the, the intersection of how all these things tie
48:30
together into one customer experience.
48:32
Got it. And so I know there's a very
48:34
long way to answer your question, but we have been
48:36
running this compound start up strategy for the customer
48:39
experience that has allowed us to grow dramatically.
48:43
Some of that we've built directly and we've got great
48:46
engineers and we use AI for that.
48:49
Some of it we've partnered. So like take payments for
48:52
example, we have a payments portion of what we do.
48:53
We didn't have to build a payment processor in 2026.
48:56
I've got a partnership with Stripe for that.
48:59
And then some of this we've we've gone out and acquired
49:01
companies that are really good in their respective area.
49:05
We'll take on all the work of doing the integrating and the
49:07
duct taping and the Elmer's gluing so our customers don't
49:10
have to do that. And then we go cross sell up,
49:13
sell like crazy, back out to our respective markets.
49:15
So that's been the, the strategy we're running From a organic
49:20
growth perspective, we're roughly 5050 between 5050 sixty
49:26
40 between direct outbound sales and then partner LED growth.
49:31
But the the core at the end of the day, the problem behind the
49:34
problem, why haven't more companies automated their
49:37
support? It's because stitching all these
49:40
things together, even in 2026, if you don't eat and breathe and
49:44
sleep that it's very painful. And that's what we solve.
49:47
That's the problem behind the problem that we solve and allow
49:50
our customers to do awesome things.
49:52
Incredible. Cool.
49:54
Totally get it. Yeah, and it makes sense
49:55
because, you know, there there is more and more platforms that
49:59
everyone's kind of native to, you know, if you know, the
50:03
Senate isn't comfortable on, on emails.
50:06
So it's like they'll be able to opt easily into text messages if
50:10
you prefer that, you know. So the, the there is always
50:13
going to be need to be like a, a downstream e-mail confirming
50:17
everything that went out. It's, you're going to need to
50:19
have certain things in place, but you know, something like
50:22
capacity as you know, if it's my kids, I'm sure they would rather
50:26
have everything in Snapchat, you know, 'cause that's where they
50:30
are. And so like eventually you'll be
50:31
able to have service agents everywhere, you know, where
50:34
you're most comfortable and most native to those platforms. 100%.
50:39
Yeah, all right. It's time in the show to go to
50:42
the Quantum round. Are you ready?
50:44
Let's do it. This is the next horizon of
50:46
interviews. OK, First things first, when
50:50
you're talking about the acquisition, when you sold
50:53
answers.com and you were looking towards the next project, that
50:59
exit was $900 million. So how long did you take off and
51:06
why did you decide to jump back into a start up?
51:08
Why not just boat drinks into the sunset?
51:12
A couple things. One I took five months off, I
51:18
realized I can't just go drink Mai Tais on a beach for very
51:21
long. It's great for a couple days.
51:22
Not a great lifestyle. I played sports in high school.
51:26
I wasn't a good athlete, but I realized now in retrospect, I I
51:32
need to be on a team competing for something.
51:36
We have a black lab mix and we had a trainer come in to try to
51:39
help us because she was a little bit of a mess.
51:41
We got her and they're like Labradors need a job, They need
51:46
a they need a purpose, they need a a reason for they don't they
51:50
don't just hang around the house dogs right.
51:52
They they need to be doing something.
51:54
I'm chasing blush. Exactly.
51:56
I'm I'm the Labrador of of this analogy here.
52:00
And then I also looked at, I made my long laundry list like
52:03
here's 50 different things I could do.
52:05
And there were some probably 3 or 4 good ideas in there.
52:12
Another, I don't know, 6 or so OK ideas and then 40 terrible
52:20
ones. But this was the one where I
52:22
looked at and I said, if I don't start this company, I'm going to
52:26
regret it the rest of my life. And I don't want to look back
52:30
with regret. I, I, there's another Bezos
52:34
interview where he talked about, you know, why he started Amazon?
52:37
He went through a very similar type of framework.
52:38
He's like, if e-commerce took off and he didn't start Amazon,
52:42
he'd regret it the rest of his life because he knew or he
52:45
thought we had this just internal belief that this was
52:53
going to take off. And so very similarly, I mean,
52:58
I'm not Jeff Bezos, but I looked at this and said, hey, this AI
53:01
thing takes off like I think it will.
53:02
And I didn't start this company, man, I'm going to regret not
53:06
doing that. Yeah, so it was in five months.
53:11
That's that's a good amount of time.
53:13
You know, that's enough time to really do some soul searching
53:15
and and decide like what what what do I want to do next?
53:19
So, you know, like almost everybody I talked to, I kind of
53:22
asked them that question after an acquisition and it's just cut
53:25
boils down to you love the game. You know what I mean?
53:28
It's like some people get into this.
53:30
I do love. Yeah.
53:32
And it's interesting because I just enjoy talking to you so
53:35
much. You have a really like pleasant
53:37
affectation. You see them, you seem rested
53:40
and like healthy. It's not true of everybody, but
53:46
also like just incredible success.
53:48
You know, like the, it looks like your, your first year, you
53:50
did 5 million in, in annual revenue and like you're up
53:53
soaring past 50 million now. You know what, what keeps you
53:58
kind of base and, and calm? You know, I, I definitely talked
54:02
to a lot of founders that don't have your Labrador energy.
54:08
Yeah, I mean that the Labrador energy is definitely a
54:10
superpower for sure. Yeah, I'm a practicing
54:14
Christian. So I try to take a 24 hour
54:17
Sabbath from Saturday afternoon to Sunday afternoon and like
54:20
unplug. You can call me if there's a
54:23
major fire. But yeah, I I try, I try to have
54:27
rhythms of, of rest throughout the week and I work my tail off.
54:31
One thing I learned, this is something we've some of your
54:33
founders who are listening here. My last company I worked crazy
54:37
hours, but it was very erratic. In this company, I have a very,
54:43
very fixed schedule. So I work late every Monday
54:45
night, every Wednesday night, most Thursday nights, often
54:49
Saturday mornings. But I do a Tuesday night date
54:52
night every night with my every week with my wife.
54:54
And it might not be some big thing.
54:55
It might be we're watching some show and grabbing take out or
54:59
something. But we I try to have like fixed
55:03
time, sacred time where I'm not fully plugged in the rest of the
55:08
matrix. And what I found is my I do a
55:11
Thursday morning breakfast. I have 4 kids.
55:13
I do a rotating Thursday morning breakfast, one-on-one with each
55:16
of my kids. And that was a pretty easy way
55:21
of adding one-on-one time into a crazy busy schedule when they
55:26
know dad's worked a lot. And so one thing I would
55:30
recommend to any of the founders on this call as don't give up on
55:34
having to work hard. Like anything with everything
55:37
going on today, you're probably going to have to work harder.
55:39
But if you can order your your time where you have kind of an
55:46
ideal calendar that you're plugging everything into as
55:49
opposed to just reacting to what gets thrown at you, yeah, you
55:52
can have a lot more consistency, which allows you to run fast and
55:56
harder, faster and harder for a longer period of time.
55:59
Love it. And that's been instrumental for
56:02
me as well, you know, like having time blocks for certain
56:06
tasks so you can kind of prioritize and and do deep work.
56:10
What's what's hard work for you right now?
56:12
What's what's the stuff that's really challenging?
56:16
The most challenging part of my world is that running this
56:19
compound startup strategy, I could go spend the entire next
56:23
year, I'll take all of our developers and all the, all of
56:26
our clod code and everything we're doing and say just go work
56:30
on helpdesk or just go work on voice virtual agents or just go
56:34
work on identifying our survey like I could, I could point at
56:41
any one area and it would be well spent.
56:45
The hard part if you're running this strategy is you have to be
56:48
in constant prioritization mode and you're not prioritizing good
56:53
versus bad or good versus OK. You're you're constantly
56:57
prioritizing good versus great. And that's the that's the
57:01
hardest thing to do because inherent to the model is you get
57:06
value with having everything integrated and shared
57:09
components. But how often do you work on the
57:12
integration piece versus the individual nodes?
57:15
That's where I spent a lot of time thinking.
57:20
Love it. David, this has been awesome.
57:22
I kept you a little extra and I don't want to go overtime, but
57:25
if I have two minutes, I would love for you to talk about
57:29
Create a Loop, which is a really cool nonprofit organization that
57:32
you're on the board of, I believe.
57:34
Yeah. So a few years back we helped
57:38
start this organization called Create a Loop.
57:41
We helped teach computer science and AI to kids who don't have
57:46
access to to that sort of thing. 90% of our kids are on the
57:49
school lunch program. That about 1700 kids go through
57:52
the program we started out doing after school.
57:57
We've also done some summer program work.
58:00
We're now starting to teach teachers how to teach C, basic
58:04
CS, and basic AI usage into these schools.
58:06
It's one of the most impactful things I've I've been a part of.
58:12
For people that want to check that out, it looks really,
58:14
really cool. It's createaloop.org.
58:17
The website iscapacity.com. Definitely check it out.
58:21
Bunch of great information there.
58:22
And David, where's the best place for people to connect with
58:25
you personally? Just e-mail me
58:27
david@capacity.com. David@capacity.com Thank you so
58:30
much. I really appreciate this
58:31
conversation. Hey, thanks, Ryan.
58:33
Take care, man. You are about to spend anywhere
58:36
from 100,000 to $300,000 on your kids college education and
58:41
that's not a small bet. And right now there's a very
58:43
real chance that they graduate and can't find a job.
58:46
Not because they weren't smart enough, not because they didn't
58:49
work hard enough, but because the job market changed and
58:52
nobody told them. Employers are hiring AI native
58:55
candidates, people who don't just know their field but know
58:58
how to work with AI at a professional level.
59:00
And most colleges are not teaching this, not even close.
59:04
AI Native student.com is the program that fills that gap.
59:07
It takes high school and college students and teaches them how to
59:10
actually think and work with AI the way top companies expect on
59:14
day one. So your investment in their
59:16
degree actually pays off. You're already paying for
59:19
tuition, housing, textbooks. This is the one thing that
59:22
actually makes it all worth it. Go to ainativestudent.com
59:26
because a diploma is a starting point.
59:27
Being AI native is what gets them hired.
59:30
That's AI native student.com.