Up next in 10
Sell Your Expertise as an AI Twin: David Riha on BuddyPro and the $5M Expert Economy
Jul 7, 2026
What if you could package your entire expertise into an AI that coaches your audience 24/7 — without you delivering anything? In this episode of Build With What You Know, I sit down with David Riha, founder of BuddyPro, to explore one of the most interesting new digital product models in the creator economy: selling your own AI twin as a subscription.
We cover why memory and long-term relationship are what separate real AI coaching from a simple chatbot, how 160 experts have already generated $4.8M in subscription revenue on the platform, what it actually takes to build and launch your AI twin, and why David believes every major expert will eventually sell their own AI as a digital product.
🔗 BuddyPro → https://buddypro.ai
🔗 David's blog → https://davidriha.com
💼 LinkedIn → https://www.linkedin.com/in/rihadavid/
📸 Instagram → @davidrihaofficial
📘 Facebook → https://www.facebook.com/rihadavid
🐦 Twitter/X → @rihadavid12
Subscribe for more conversations on online business, enterprise technology, and digital expertise.
#AICoaching #AITwin #ExpertEconomy #OnlineBusiness #BuildWithWhatYouKnow #DigitalProducts #CourseCreator #ArtificialIntelligence #PassiveIncome #KnowledgeBusiness
Show More Show Less View Video Transcript
0:02
Hello David. So you are the founder of
0:04
Buddy Pro
0:06
which is a second
0:09
uh well a kind of second brain. Can we
0:12
say that?
0:13
>> Yeah. It's like an AI version of uh an
0:16
expert or a coach where they upload all
0:20
their knowhow and create their own AI
0:22
that people can chat with or a
0:25
subscription.
0:27
I see actually a lot of such initiatives
0:31
different products uh they kind of
0:34
market themselves differently. So what
0:36
is what makes body pro unique?
0:39
>> Mhm. So so uh we basically built a body
0:42
pro because none of the none of the
0:45
platforms were good enough. like uh my
0:48
brother came to me in uh uh almost two
0:51
years ago now and uh he's been uh an
0:54
expert teaching people how to do online
0:56
businesses and he's been doing online
0:58
courses and stuff like that. uh but
1:01
basically he want wanted to come up with
1:03
an AI version of himself that can teach
1:06
people what he is teaching and we tried
1:09
the other platforms like uh deli or
1:11
coachbox but uh the problem is that like
1:15
they don't they don't retain uh memory
1:19
they don't it's it's more like a Q&A
1:21
like you you come you ask the question u
1:24
and you get the answer but it doesn't
1:25
form a coaching relationship with with
1:28
the user
1:29
So, so we wanted to build something that
1:32
actually coaches uh over a long uh long
1:36
time remembers everything about the
1:37
person and guides them uh on a
1:40
transformation. So uh we decided to
1:43
build it ourselves and uh uh now now we
1:47
we see that it's a it's a really good uh
1:49
business model and and it like this uh
1:52
long-term relationship is what's really
1:53
important because people come again and
1:56
they talk with it every day and this is
1:58
what uh enables the expert to sell
2:01
subscriptions for like one or $2,000 a
2:03
year if it's just a Qi chat Q&A chatbot
2:06
on the website uh people come ask
2:09
question and they never come back again.
2:12
So, it's it's uh it's hard to sell this
2:14
kind of uh product for lots of money to
2:17
your audience. But if it's a an AI that
2:20
actually builds mentoring relationship
2:22
with the users, that's that's a totally
2:24
different game.
2:26
>> So, actually, what's the difference
2:28
between your product uh or similar
2:31
products and a standard um generative
2:35
artificial intelligence? Is it that it
2:37
already has context and it keeps context
2:39
and it grows on this context? Yeah, it's
2:42
a great question. Like uh the tools like
2:45
chpt or clo they do have some kind of
2:48
memory and they they can use it if you
2:50
if you ask uh ask some questions that
2:53
like if you go for it if you ask like
2:55
hey based on what what I said yesterday
2:57
do this and that uh but if you open a
3:00
new window and start chatting in chat
3:02
GPT it doesn't automatically start based
3:05
on your full uh conversation history and
3:08
profile. So, so uh we've built it more
3:11
like an AI companion that's always
3:14
always referring to what you were
3:16
talking about in the past and always uh
3:19
answering uh questions in the context
3:21
and building on the relationship and why
3:25
this makes uh so much sense for the
3:27
experts and coaches uh selling their AI
3:29
like this and like people actually buy
3:31
it because like people people like
3:35
talking to AI they use JPT they use
3:37
cloud but
3:39
They actually follow some experts that
3:41
have some proven expertise. They know
3:43
that this guy is really good at selling.
3:45
This guy is really good at building
3:46
businesses. This guy is really good at
3:48
advising uh about relationships and uh
3:51
they can go and ask GPT but they will
3:54
get advice uh that the CH GPT gives to
3:57
everyone based on everything it knows.
3:59
So it's not proven advice. It's not
4:01
something that uh that would be based on
4:04
uh proven uh proven expertise. So when
4:08
the expert create his own AI, people are
4:11
actually super happy to talk to the AI
4:13
version of the experts because they know
4:16
they follow the expert, they trust him
4:17
and they wants his knowledge. They don't
4:19
want the general knowledge from CH GPT
4:21
because nobody knows if if it's going to
4:23
work or something. But if the expert
4:25
actually creates his AI based on his
4:27
proven methodologies, uh the people can
4:30
follow it easily through the AI advice
4:32
of his own AI and uh they they know that
4:35
this is going to work. this is because
4:37
this is how this expert is doing doing
4:39
it and it's working for him.
4:42
>> So basically your product allows experts
4:45
to resell their to package and resell
4:49
their own expertise as an AI. Correct.
4:52
>> Yes. Yes. Exactly.
4:54
>> Wow. And it's accessible to anyone.
4:57
>> Uh well the expert like any any any
5:01
expert any coach anybody who has
5:03
expertise can build it. uh uh it it's uh
5:07
they can just uh go uh go to our uh
5:10
website uh I will share share the screen
5:14
body.ai AI and there's a so like anybody
5:18
can build it. This is uh this is DIY. So
5:20
you just upload all your knowhow inside
5:23
uh it trains itself. It's it's pretty
5:25
simple. It can be built within a couple
5:27
of hours and uh then the experts can
5:31
start selling it to their audience and
5:34
uh it's it's usually not free for the
5:36
audience uh because the expert wants to
5:38
monetize this knowledge. So, uh, usually
5:41
people sell sell sell it to their
5:43
audience for$1 or $2,000 a year as a
5:46
standalone product and, uh, together
5:49
they already made $4.8 million in
5:53
subscription revenue, the experts
5:55
selling to their, uh, end users. So, uh,
5:59
it's it's a really nice uh, recurring
6:02
revenue stream and uh, a way to monetize
6:06
your knowledge uh, as an expert.
6:11
How long does it take to set it up? What
6:13
do you need to give to the system?
6:16
>> Yeah. So, so basically takes a couple of
6:19
hours if you already have all your
6:21
knowhow in some folder and you just
6:23
upload it in Google Drive folder. We we
6:25
uh take it from the Google Drive folder.
6:27
The AI trains itself on the AI uh on the
6:31
content that you put in the Google Drive
6:32
folder. So, if you already have the
6:34
content, you've done most of the work.
6:36
Like if you have book some books maybe
6:39
some uh YouTube videos or YouTube
6:40
channel uh could be even like Instagram
6:43
res or you have some online courses
6:45
videos so you can upload everything
6:47
inside and then it trains itself within
6:50
a couple of hours and uh then you adjust
6:53
uh the system prompt a little bit maybe
6:54
to describe the behavior give it some
6:56
name and uh then you connect to it uh
7:00
connect it with the stripe integration
7:02
and you can basically start selling it
7:04
like we we've had people that uh started
7:06
started selling their AI within 24
7:08
hours. But uh usually it takes people
7:12
couple of weeks because they need to
7:13
figure out the pricing. They want to
7:15
test it with some users first. They want
7:17
to get some reviews so so they can sell
7:19
it better and uh they need to figure out
7:21
how to position it in their like suite
7:24
of their other digital product.
7:26
>> Exactly. I guess actually most people
7:29
experts that you are targeting they are
7:31
probably expert in their fields but not
7:33
experts in marketing sales and marketing
7:35
which is totally different.
7:37
>> Yeah that that's a good point like uh
7:40
this this works best for people that
7:42
have audience
7:44
uh has have some expertise obviously and
7:47
know how to sell to this audience. So
7:50
like for people that have been doing
7:52
online courses before or some membership
7:54
sections or coaching calls, group
7:57
coaching, then it's uh it's really easy
7:59
to offer this to uh the existing
8:01
audience. If uh if you would be a let's
8:04
say a therapist that has 20 clients, it
8:08
can still make sense, but you need to
8:10
actually uh offer it to them and uh like
8:13
you you need you need to sell it
8:15
yourself. We are not going to be selling
8:17
your AI to your audience. uh that's
8:19
that's your goal. But like even if you
8:22
even if you would sell it to three
8:23
people, it's you you can still get uh
8:26
get to some nice passive income from
8:29
that. But like usually people sell it to
8:31
hundreds of people if if they go big
8:33
with this?
8:35
>> Yeah. So by the way, the numbers on your
8:37
website, are they real?
8:39
>> Yes. Yes. We have around 650 people that
8:44
uh that are building their AI on body
8:48
pro and uh around
8:52
160 of them already launched their AI
8:54
and those 160 people have uh uh
8:59
recurring revenue that's uh around $5
9:01
million. It's actually probably a little
9:03
bit higher because we are not tracking
9:05
uh the revenue of everybody because
9:06
somebody does some custom integration
9:09
with uh Zapier. So we don't see the
9:11
sales but uh this $4.8 8 million in
9:15
subscription revenue is calculated based
9:17
on 130 uh AI twins that have been
9:21
launched on body pro and uh there's uh
9:24
there's multiple of them that made
9:26
multiple six figures selling their AI
9:29
twin and there's even one that's uh very
9:32
close to seven figures. So, uh, it's
9:35
it's real numbers and and the average
9:37
revenue is around $35,000 once you
9:40
actually launch your AI to build on body
9:43
pro.
9:44
>> Actually, of your more or less 650
9:47
clients, only 100 something have started
9:51
selling their own. Uh,
9:53
>> yes. Yes. That's that's a that's a
9:55
common point like people people ask this
9:58
question a lot. Why why why is there 650
10:02
that bought the bought the access to
10:04
build their body pro and only 150 that
10:07
launched it? And uh the answer is like
10:11
lots of the people are not in a stage
10:14
where when they are ready to start
10:15
selling it. So like they don't have the
10:17
content together. So they need to write
10:20
it down or they need to collect it from
10:23
all the different sources or it's just
10:25
maybe not their priority at the moment
10:27
but they really like the idea. So they
10:29
didn't launch it but they are preparing
10:31
it like we we first launched this
10:34
product a year ago. So from these 650
10:38
people some like couple of hundreds of
10:40
them bought within like the past couple
10:42
of months. So so they are still in the
10:45
preparing phase and uh it's a it's like
10:49
the the build is the build is easy but
10:52
also you need to uh figure out how do
10:54
you want to sell it. You need to have
10:55
the content. So u it's not only about
10:59
building it, it's also about as you said
11:01
being good at marketing, being good at
11:03
sales and uh having the time to actually
11:06
launch this.
11:08
>> Yes. So you you mentioned it can
11:11
actually take existing videos as well.
11:13
So it can uh for example read a whole
11:15
YouTube channel.
11:17
>> Yes. And
11:19
>> yes, we we we have some people that uh
11:22
put inside you can just like put inside
11:23
the link to your YouTube channel and it
11:25
will process all the videos. Uh usually
11:29
we advise people to like check if
11:31
there's not any videos that would be
11:32
like uh interviews with other people
11:34
because then the knowhow would contain
11:36
uh information that should not be there.
11:38
But uh yeah, you can totally put inside
11:40
the whole YouTube channel and uh we have
11:42
uh some of the most successful experts
11:45
do have their YouTube channel inside as
11:47
the source and it's working really well.
11:50
>> Wow, that's very interesting.
11:54
>> And and typically for example, if you
11:57
have a look at your customers that are
11:59
making the the most sales, what are what
12:04
do they have in common? H I would say
12:08
they have in common that they they are
12:10
very experienced with selling digital
12:13
products. So they've been doing uh
12:15
online courses uh in the past or they
12:17
they uh have membership sections or they
12:19
do group coaching and they they offer it
12:22
to to already paying clients. So if you
12:26
are a paying client of some expert and
12:28
he tells you here's my AI now you can
12:30
talk to my AI 247 and uh it costs one or
12:33
$2,000 a year. lots of people are like,
12:36
"Wow, this is exactly what I want. I
12:38
want to talk to you 247. Uh your AI
12:41
version, it's it's a great product." So,
12:43
it's it's really easy to sell it if you
12:44
have the audience. And uh like about the
12:48
the topic of the expertise, it's
12:50
actually it can actually be anything and
12:53
it v varies. It's uh like we have uh I
12:56
would say the like three out of five of
12:59
the most uh most successful ones are
13:01
business coaching or some specific uh uh
13:05
topics around business like how to do
13:07
marketing, how to be successful at
13:09
online business, uh these kinds of
13:11
stuff. But also there's uh among among
13:14
the top 10 there's also like uh
13:17
relationship coach or or healthy
13:21
lifestyle coach. Uh there will be a
13:24
celebrity chef launching his uh his AI
13:27
uh really soon with us and it's going to
13:29
be one of the most successful ones. We
13:31
have also a real estate expert that's uh
13:35
that's focusing on u uh real estate ref
13:39
lips. So so it's it's really it can be
13:41
anything with any kind of expertise.
13:45
>> Wow. Okay. Um so any kind of expertise
13:49
but then once they have launched their
13:52
own um kind of second brain how does it
13:56
look like is it just another prompt to
13:59
to query or
14:01
>> yeah so it's uh it it lives inside
14:03
telegram uh we also have option to
14:07
launch your own mobile app on the app
14:09
store but uh the main the main uh
14:12
interface is telegram and there's two
14:14
reasons for that first of all It's the
14:16
it's the best uh platform for building
14:20
chat bots and uh and that's because it
14:23
can be proactive. So when people chat
14:25
with the AI, it will message them like
14:26
three days later, 5 days later and ask
14:29
them how how they are progressing which
14:31
uh really increases the retention and
14:33
also Telegram is super private like
14:35
nobody is uh reading the messages of the
14:37
users they so they go deep with their
14:39
personal topics and uh also it stands
14:42
out because uh if if it would be on
14:44
WhatsApp which doesn't doesn't allow the
14:47
retention messages couple of days later
14:49
but if it would be on WhatsApp it just
14:51
it would just get lost in so many
14:53
convers conversation threats but uh
14:55
actually not many people not that many
14:57
people use telegram so when they get a
14:59
not notification from telegram they know
15:01
it's from the AI that they are talking
15:03
to so it stands out a little bit in the
15:05
all the blood of notifications
15:08
and uh also yeah as I said we can people
15:11
can also release their own mobile app in
15:14
the app store
15:16
>> so it is text only right
15:18
>> yeah it's it's text and voice messages
15:21
uh and uh it's it's also a good point
15:24
like why don't we have for voice mode or
15:26
video mode. Uh we are actually
15:29
experimenting with it but the thing is
15:31
the an like the highest quality the the
15:34
biggest reason why people use this is
15:36
because they get really good answers and
15:38
that they get really good uh messages
15:40
and really good coaching and if you want
15:42
to build voice mode you need to make it
15:45
little less smart because there's just
15:47
no time to think. So when when I message
15:49
the AI, it's it's it's actually like
15:51
thinking what it should answer. And if
15:53
if it would be real time voice mode, we
15:57
need to make it a little little bit less
15:59
smart and uh then we are risking that it
16:02
could potentially give the wrong answer
16:04
which nobody wants. So so far we don't
16:06
have a voice mode but we are planning to
16:09
also build it within a couple of months.
16:12
And then when a user buys the product
16:15
from one of the expert, he can does he
16:19
have any limit?
16:21
>> Yeah, that's that's a that's a good
16:23
question. Uh there is uh the expert can
16:26
set his own limit. Uh typically they
16:29
they set a limit of 1,000 messages per
16:32
month and uh if people run out of the
16:36
messages, they can uh buy a package of
16:38
extra messages to continue chatting. But
16:40
it's like both the price and the message
16:43
limit is up to the expert how they set
16:45
the pricing and uh they also they need
16:48
to cover the AI usage of their users
16:51
because uh it's using very smart models
16:53
and a huge context because of the
16:55
unlimited memory and uh which is what
16:58
makes it smart but the expert has to
17:00
cover the AI usage of his users uh which
17:03
is uh on average around $22 per
17:06
subscriber per month. So they need to be
17:09
uh selling it for something like 100 or
17:12
200 a month to uh have some uh nice
17:15
profit there and we don't take anything
17:16
from the profit.
17:19
>> Okay. And how much actually cost your
17:21
product for the expert?
17:23
>> So so the the license for body pro is uh
17:26
currently
17:27
uh for uh $197
17:31
per month with annual payment. Uh we and
17:35
we don't actually offer monthly payment
17:37
at the moment. So, so it's uh $2364
17:42
per year to uh get the license and that
17:46
includes everything like updates and uh
17:49
uh support and uh just everything you
17:51
need to build it.
17:54
>> Okay. And then any expert that way can
17:56
build a new income stream provided that
17:59
they have a good following or they are
18:01
very good at marketing.
18:03
>> Yes. Exactly.
18:06
and they don't then they they don't have
18:07
to like deliver anything that's the
18:09
beautiful thing when when you build and
18:11
sell your AI twin people just chat with
18:13
it and like maybe before the expert was
18:16
doing online courses and it takes so
18:17
much time to prepare the online courses
18:19
and then like every every year you need
18:22
to launch a new online course because
18:24
things are changing and people want new
18:26
new content so like there's so much so
18:28
much work with delivering the expertise
18:31
but here with the AI you actually you
18:32
don't have to deliver it you just put it
18:34
in the Google drive folder the AI trains
18:36
itself and then it's chatting with
18:37
people. So the expert only needs to sell
18:40
it and then he doesn't have to do
18:41
anything for like one year people are
18:43
chatting with the AI. Obviously he can
18:45
update it. He can put more sources
18:47
inside but uh usually it's not necessary
18:50
if it's not some topic that keeps
18:52
changing every week. Uh people just chat
18:54
with the AI and the expert doesn't have
18:56
to do anything.
18:58
>> And how is the data staying on your
19:01
system?
19:02
>> Yeah. So, so uh the expert puts it in
19:05
Google Drive folder and from that we
19:08
extract the data. We put it in the in
19:10
vector databases uh in pine cone and uh
19:16
but but like the we don't we we always
19:18
make sure to not use any AI models that
19:20
will be training themselves on on the
19:23
content. So so it's it's protected. We
19:25
don't use it for anything else than
19:26
providing the service and we don't give
19:29
it uh to any third parties to use it for
19:32
anything. So uh it's it's protected and
19:36
the expert has control. If you if he
19:38
would want to remove it, we just uh
19:40
clean the database and uh it's always in
19:43
his control.
19:45
>> Exactly. because personally I am a
19:47
content creator, course creator and
19:49
that's what I would be afraid of that
19:51
actually all this knowledge would be
19:52
leaking um into the AIS and actually
19:55
accessible without this product.
19:58
>> Yeah.
20:00
Yeah. We understand that this is uh this
20:02
is really like it's it's your life's
20:05
work if you are the expert putting the
20:07
information inside. So we are protecting
20:09
it and uh it's not going to leak
20:11
anywhere.
20:13
But but but how can you ensure that the
20:15
AI models behind it will not train
20:17
themselves even if
20:20
>> Yeah. So so we use we use open router
20:22
for for our AI models providers and open
20:26
router has a set setting where you can
20:28
uh you can choose to only use the
20:31
providers that have in their terms of
20:33
service that they don't train on the
20:36
inputs. So we only use uh the providers
20:38
that uh that have uh in contract with
20:41
open router that uh they don't use uh
20:44
use it for training the models.
20:47
>> Wow. Wow. Okay. Um
20:51
yes. And does it still updates? I I mean
20:55
it's not only based on the knowledge
20:57
from the expert, it also updates based
20:59
on the general available data.
21:03
>> Yeah. Yeah. So, so we like as as uh the
21:07
AI updates over time, we also keep uh
21:10
updating the models that we use
21:13
uh and uh we are always trying to use
21:16
the best models and the newest models.
21:18
So it can totally also answer from uh
21:21
from the general knowledge of the model
21:22
that that depends on how the uh expert
21:26
configures the AI if it should or
21:28
shouldn't be saying things based on
21:30
general knowledge and uh it can also
21:32
search on the internet if the user asks
21:35
to get some uh latest information
21:39
>> and maybe do you have examples of how is
21:42
it better to query like one of these u
21:46
models created by your system or the
21:49
general available models.
21:52
>> Uh you are asking like if uh it would
21:55
make sense to build this yourself
21:57
instead of uh using a buddy pro.
22:00
>> Yes.
22:02
>> Yeah. So so it is totally possible to
22:04
build this yourself. It's just uh really
22:08
hard to build it in a way that's uh
22:10
production ready and uh that has uh like
22:14
a fine-tuned uh companion core because
22:17
there's there's like a couple of things
22:19
that need to happen for this to work
22:21
which is it needs to be proactive. It uh
22:24
needs to uh have the unlimited memory.
22:26
It needs to be able to answer very well
22:29
based on the knowhow and uh it uh it
22:32
needs to uh create like it needs to
22:35
write such a good messages that it
22:37
touches people deeply and it it delivers
22:40
transformation and this is really hard
22:41
to crack because even with chat GP or or
22:45
cloud if you chat with it it's it's kind
22:47
of plain like it doesn't doesn't really
22:49
do uh much with you but what we created
22:52
is AI that really like touches people
22:54
deeply and
22:56
So, so this is like how how this is the
22:59
things that it needs to be good at which
23:01
is not not really easy to build and also
23:04
AI is not really good at prompting
23:06
itself to be like that. So that's one
23:09
thing and then then you have like all
23:11
the integrations with payment providers
23:14
all the maintenance because the models
23:16
keep changing. Uh so if you would build
23:18
this yourself uh you could spend like uh
23:20
one weekend in a cloud code trying to
23:23
build it and you would have uh you would
23:25
have a chatbot that works kind of well.
23:28
But if you we would want want to scale
23:30
it to hundreds of people and keep it
23:31
updated and do it like worldclass
23:34
solution. That's just that's just really
23:36
hard hard to build because like we've
23:38
been building the software for almost
23:40
two years now and there's still ways to
23:43
improve and there's still lots of uh
23:45
things that need to be updated like
23:47
every week because uh the the market
23:49
keeps changing and the models keep
23:51
changing. So it's just like yes you
23:53
could build it yourself but it's really
23:55
hard to uh nail it down and to make it
23:59
really good and it's really hard to m
24:01
maintain it. So I would not recommend to
24:04
building it yourself.
24:06
>> How do you see the evolution of this
24:08
market, the second brain market?
24:10
>> Mhm. Yeah. I feel I feel that it's still
24:15
still super early and like 90% of
24:19
experts are probably not thinking about
24:22
creating their own AI twin that they can
24:25
sell for one or $2,000 a year. Uh and
24:29
like we are probably the first on the
24:32
market that are bringing this uh
24:34
category like the at this price one or
24:38
$2,000 a year. Uh we we do have people
24:41
that uh that are selling it for $2,000 a
24:43
year and uh it's and they have a huge
24:45
success with it. But like if you I I I
24:48
asked Cloud like two weeks ago. I was
24:52
play like role playing. I'm I'm a I'm a
24:53
famous coach and I want to build my own
24:55
AI based on my knowledge and sell it for
24:57
$2,000 a year. And Claude told me, I
25:00
need to stop you there. This is not a
25:02
good idea. Nobody has ever done this.
25:03
It's probably not going to work. And but
25:06
like we have it proven. We have uh we
25:07
have people that made lots of money
25:09
doing this. So we are bringing this uh
25:12
this new idea on the market and the
25:14
market still doesn't know this. So we
25:15
are like trying to spread the word
25:17
trying to bring it to more and more
25:19
experts and I believe in the future uh
25:22
every every major expert will have uh
25:25
his uh own AI uh that he's selling as a
25:29
digital product because it's just like
25:30
the best digital product you can have.
25:32
It's personalized, it's scalable and it
25:35
earns while you sleep and you don't have
25:37
to do anything.
25:39
>> Yes, it looks beautiful.
25:44
So to whom would you recommend your
25:47
product like if you would uh target your
25:51
typical client?
25:53
>> Yeah, I would say to everybody who uh
25:56
who has uh who has some coaching or
25:59
expertise based uh business and who is
26:02
already selling digital products. Huh?
26:05
>> How big should be the coaching business
26:07
to make sense?
26:10
I would say if you have uh if you have
26:14
like at least 10 or 20 people that
26:17
already bought some online courses from
26:20
you or something then they can also buy
26:24
uh your AI to from you. So like your
26:26
audience doesn't have to be uh very big
26:29
to make money of this. But if you if you
26:32
have hundreds of people that already
26:35
bought some uh expert uh expert based
26:39
products like online courses or coaching
26:41
then it's it's like the best best
26:43
digital product you can have and it's uh
26:45
absolutely the right way to do to do it
26:48
in 2026.
26:51
>> Okay. And probably one last question. So
26:53
you are mostly targeting single users,
26:56
right? Not corporations, companies.
26:59
>> Yes. Yes. Usually it's businesses that
27:02
are based on one person but uh very
27:06
often it's uh these experts are so
27:09
successful with their business like
27:11
they've been running the business for 10
27:13
years or something. So they already
27:14
might have a team of like five or 10
27:17
people maybe around them. So they might
27:18
be doing like media, they might be doing
27:21
YouTube, they might be uh they might
27:23
even have a sales team. So sometimes
27:25
it's businesses but uh it get businesses
27:29
based on one person uh but uh also uh
27:32
also uh it can uh it can be just really
27:36
one one man business uh and uh also it
27:40
it makes sense for businesses that are
27:43
based on expertise of multiple people.
27:45
So uh if it's like uh let's say it's uh
27:49
some coaching business targeted at
27:51
specific topic but there's like three
27:53
coaches doing the business then it also
27:55
makes sense uh to build the AI
27:58
>> also it can uh merge different experts
28:02
together kind of.
28:04
>> Yes. Yes. You can basically like give it
28:06
a name madeup name or just name it
28:10
however you want. Uh for example we had
28:12
one that's named online strategist. So
28:15
it's not based on on a name and you can
28:18
upload inside
28:20
sources of multiple people. You just
28:22
need to have the copyright for that
28:24
obviously but uh it can be based on
28:26
multiple people. Yeah.
28:28
>> And can it be on different topics?
28:32
>> Yes. Yes. It can totally be on different
28:34
topics because uh what we do is when you
28:37
upload the knowhow inside we create
28:39
expert roles uh based on the knowhow. So
28:42
if you upload know how about gardening
28:44
and about relationships we will create
28:48
hundreds of roles uh that are like
28:50
specific to all these topics. So it can
28:53
be uh like uh specialized roles for
28:57
gardening and like 20 different roles
28:59
based on gardening and then 20 different
29:01
roles targeted at relationship and it's
29:04
all generated automatically from from
29:06
the knowhow. So then when people chat
29:08
with the AI and they talk about
29:09
relationships, it's going to be using
29:11
those expert roles for the
29:13
relationships. And if there's another
29:15
user talking about gardening, it's going
29:17
to be using the expert roles about
29:18
gardening. And there can also be a user
29:20
that's using that's like interested in
29:23
both of those topics. So that's also
29:25
possible to build it like that.
29:28
>> Okay. Well, sounds like a great product.
29:30
Thanks a lot for answering my questions
29:32
and giving us all these insights. And
29:35
your website is uh bodyypro.com.
29:39
>> Bodypro.ai.
29:40
>> Oh, bodyypro.ai.
29:41
>> And And I also have my own personal
29:44
davidriha.com.
29:46
I'm writing a blog blog there about AI
29:49
digital twins. And there's also probably
29:52
a link to my Instagram somewhere down
29:54
there.
29:55
>> Okay. I leave the links in the
29:57
description.
29:59
Thanks, David.
30:01
>> Thanks a lot for inviting me. It was
30:03
nice chat. Thank you.
#Science