Latest / AI for Founders with Ryan Estes / AI Friends That Actually Call You (And They Remember Everything)

Transcript

AI Friends That Actually Call You (And They Remember Everything) · AI for Founders with Ryan Estes

0:00
Welcome back to AI for Founders. I'm Ryan Estes.
0:02
What happens when your most reliable teammate isn't an app,
0:05
it's a voice that calls you at exactly the right moment.
0:08
Today we're meeting the founders of Committify, an AI
0:10
accountability coach with personalities like life coach,
0:14
Zen master, sleigh bestie, hype beast, drill Sergeant and CEO
0:19
that literally ring your phone to snap you back into flow.
0:22
The tension. Real friends have boundaries and
0:25
biased therapists have calendars and co-pays and apps drown in
0:28
notifications. Committify bets to future is U
0:32
eyeless voice with long term memory that feels like a friend
0:35
and works like a coach. In this episode you'll learn why
0:37
they chose phone calls over screens, how they design agent
0:41
personas that motivate without being cringy, what memory should
0:45
and shouldn't remember the wild origin story 100 hour hack a
0:48
thon Sprint pricing by calls per week and how they're hunting for
0:52
true ICP before scaling. This one's for founders,
0:56
builders and product people who want big retention with small
1:00
and anyone exploring AI that actually changes behavior, not
1:04
just creates content. Let's get into it.
1:07
And if you like founders shipping weird, useful AI that
1:09
bends behavior, you'll love the field notes at AI for
1:13
founders.co. This is a newsletter back
1:16
channel for experiments, playbooks and the we tried it so
1:18
you don't have to kind of reports.
1:20
Join the newsletter and upgrade your lux surface area at AI for
1:24
founders.co. If this episode sparked an idea,
1:27
drop us a five star review like you're shipping it to production
1:30
on a Friday and it actually works.
1:33
This episode is brought to you by Kitcaster, the easiest way
1:35
for founders and execs to get booked on Top Podcast.
1:39
If you've got a story, a launch, or a category you're shaping,
1:42
Kitcaster turns that into steady drum beat of interviews on shows
1:45
your buyers are already binge. No cold outreach, no guessing,
1:49
just research targets, host ready briefs and interviews that
1:53
compound into SEO, social clips and sales conversations.
1:57
Our team preps you with tailored talking points, pulls the
2:00
receipts from your case studies, and lands you on the right shows
2:03
at the right time. So every appearance drives
2:06
awareness, pipeline or recruiting.
2:08
Whether you're precede or public, 1 tour with Kick Caster
2:11
beats 1000 cold emails ready to turn conversations into
2:14
conversions. Visit kickcaster.com and tell
2:17
them AI for founders sent you. That's Kick Caster by MO Burst.
2:37
All right. What's up, everybody?
2:39
Mehdi guys say it right. It's close enough.
2:44
It's nice enough. Welcome to the show.
2:47
I appreciate you guys coming on. Good to.
2:51
Be here. Thank you for.
2:52
Inviting us, you know, excited. Absolutely.
2:55
Now I'm really excited about your product 'cause it looks
2:57
really cool and I I've got a lot of takes on on this kind of
3:00
product. So before we get into the
3:02
conversation, introduce the audience here to committify.
3:10
Well, yes. You wanna go with me guys?
3:13
Yeah, go ahead, go ahead. So committify is like an AI
3:17
coach that's adjust to what you need and whatever it is like
3:22
does it, do you need some commitment help or you might
3:25
need some reflection help or just some want to talk to during
3:28
the day committed files there for you.
3:32
Love it. So essentially in what I like
3:35
about this is everything takes place over the phone.
3:37
So you have a series of different personas.
3:42
Some might call them agents, but it looks like friends.
3:44
It looks like these are kind of AI friends and depending on how
3:47
you're feeling, you can call one of them or they can call you and
3:51
and basically they're giving you, I don't know, pep talks
3:54
advice there. It seems like they're just like
3:56
that cool person that doesn't exist in your life, but now they
3:58
do. So they're the personas are life
4:01
coach. You got Zen master sleigh,
4:02
bestie, hype beast, drill Sergeant and CEO.
4:07
So, and it's a really interesting concept because
4:13
First off, I, I would be curious why you decided to create a
4:16
product like this. Yeah, that's.
4:19
Maybe for me, we were like, I was just one night I was laying
4:23
in bed, you know, I was thinking, So what do I need or
4:26
what, what helps me, you know? And then I realized like, OK,
4:31
whenever a friend of me calls me on the right time, I realized,
4:35
wait, I was doing this so I can be a person with a lot of things
4:40
if I'm interesting. So at some point when someone
4:43
calls me and they ask, what are you doing?
4:44
And I'm like, oh, wait, I was doing this instead of maybe
4:48
distracted or something or building a new product.
4:51
I was like, wait, focus. And then I thought, what if we
4:54
automate this, this activity of someone calling to you?
4:58
Because if you make it by AI, it's can be there for you 24/7.
5:05
And that's the power of this tool.
5:08
You adjust to what you need, and based on what you need, you get
5:13
it. I love it.
5:16
Is that to say there's something wrong with our friends?
5:20
You know, like, why wouldn't we just call a friend?
5:24
Well, of course a friend that calls you is amazing, but it's
5:31
it's a friend of you. So you have to respect their
5:35
boundaries. And if you ask them to call you
5:38
25 times in a day, whatever time you want them to call you, if
5:42
you need to call at 3:00 in the morning or at 7:00 in the
5:45
morning or one at 11 in the evening, it's it's friendship.
5:50
So you expect things, but it's also there needs to be a
5:55
balance. And with Comidify, it's a friend
5:58
who's always there for you. Doesn't matter what the time is
6:01
to adjust based on what you need.
6:03
Yeah, it isn't the problem with friends too, that if you really
6:07
need them, you really have a specific question, maybe they'll
6:11
be there for you. But then you got to sit through
6:14
with their question and their thing.
6:16
It can never just be like, hey man, I'm having a hard time, can
6:19
you help me with this? They're like, yeah, absolutely.
6:21
And honestly, I've had a worse time.
6:23
So now you got to listen to my I'll say, Oh no, Oh no, it's a
6:27
double sided knife. OK, I get it.
6:30
You know, so like with this, what's kind of interesting is
6:32
that it's all about me. You know what I mean?
6:34
I don't have to listen to the AIS interpretation of or I don't
6:40
have to hear about their problems.
6:41
It's all about my problems. And perhaps that was the promise
6:45
of like what a like a therapist would be there for, right is the
6:50
therapist could be there to help guide you through your things
6:52
without any commitment. There's a very like clear kind
6:56
of exchange. I give you cash for time.
6:59
You listen to my sorrows. If you can help walk me through
7:02
this, great. If not, then we'll keep it
7:04
moving and we'll push to the next one.
7:06
But it seems like this, although it's not, it's not kind of like
7:10
specifically looking to be a therapist.
7:12
It looks like it is really trying to be kind of like a a
7:15
good friend, you know, maybe like that wise friend you wish
7:19
you had. Yeah, I appreciate all my
7:23
friends. The first of that, you know,
7:24
like all my friends, I appreciate them 100%.
7:27
And this product is not creative because of I'm really happy with
7:34
all my friends and but this product is just there to to give
7:40
you what you need at whatever time that it is.
7:43
And that's not a friendship relation because you have to
7:48
adjust the older person in a friendship and it's important.
7:52
But with committify, it's will always give you what you want
7:57
and anytime you want it. And that is a very different
8:02
than a friend, but it still feels like a friend.
8:05
That's what we're trying to create here.
8:09
So cool. And I, I wonder too, you know,
8:11
the very beginning of GPT, what was that 22?
8:16
I started playing with it and talking to it.
8:18
I was like, this is cool. But like, I don't really like
8:20
the attitude. I think that there's been a lot
8:22
of like criticism about GPT because it's kind of too
8:26
sycophantic. You know, it's too
8:28
complimentary. It's like, I don't want this.
8:30
So what I would do is I would start saying like, OK, if I am
8:35
talking to Derek, these are the attributes I want from Derek.
8:38
Boom, boom, boom, boom, boom. And then if I'm talking to
8:41
Steven, these are the attributes I want.
8:43
So what I was doing is kind of trying to figure out like how to
8:46
create individual personas to get what I want very similarly
8:50
to what you've built. Now.
8:51
The problem was it wasn't great about remembering who it was or
8:56
like how to respond in a certain way.
8:57
So if I'm asking to talk to Derek, it'd just be like normal.
9:01
I'd be like, ah, it stinks. And every once in a while I'll
9:03
go back and check and be like, what can you tell me about
9:05
Derek? And it'll like regurgitate some
9:07
of the attributes that, you know, three years ago I put on
9:09
it. So I'm like, OK, it's operating
9:11
in there somewhere. So with committify, if I'm
9:15
talking to sleigh bestie because that's the kind of energy I need
9:18
this morning, you know, is sleigh bestie going to remember
9:22
the conversation so that when I come back to it, it's we can
9:24
kind of pick up like an old friend.
9:26
I mean, oftentimes the way that I consider a a true friend, it's
9:30
like I might not see him for five years, but the second I see
9:32
him, it's like, oh, what's up, man?
9:33
It's just it's right back to where you were.
9:35
Is it right back to where you were with with the this team
9:40
now? How's the memory?
9:44
Yeah. So, so memory is built into the
9:45
product, so basically exactly in the way that you're describing
9:48
it right now. It's definitely something we
9:50
need to tweak a little bit still, right?
9:51
Because, you know, as opposed to humans who can kind of see
9:56
memory in the context of today's day, like an AI agent may not do
10:00
this. So they maybe have like a memory
10:02
about something that isn't really relevant anymore a week
10:05
later. But still, because, OK, it
10:08
wasn't their main memory somewhere they may bring it up
10:10
basically. So to really make the memory
10:14
feel more natural to interact with.
10:15
It's still something we're working on, I would say.
10:17
But of course, the, yeah, the concept of adding all the past
10:21
conversations to the context window and the memory of the AI,
10:24
it's quite of a straightforward thing.
10:25
And it's something that we, we do to, yeah, create this, this
10:30
ID of memory, yeah. Yeah, well, it's only to get
10:33
better, right? I mean, this is going to be
10:36
absolutely common to the point that like, you know, the, the
10:41
teddy bear that you're going to put in your brand newborn baby's
10:44
crib is obviously going to have AI in it and it will learn the
10:47
habits and it'll learn the sounds.
10:50
It'll learn everything about your baby and it will grow with
10:53
the child, you know, So at some point, you know, it the, it's
10:57
not a, it's not a little baby bear anymore.
10:59
Maybe it's a blankie, but it's still it's, it's learning, you
11:02
know, So by the time the kid is walking and can coach it, by the
11:05
time it's in school, it knows how to like how to, how to tutor
11:10
your child just totally based on who they are.
11:13
So like this idea of an AI friend is just in its infancy.
11:16
But truly we are going to create like a relationship with AI that
11:21
will span our entire life, that will know us better than
11:24
ourself, that will have a photographic memory of every
11:27
instance of our entire life because it'll be with us forever
11:31
whether we like it or not. That's what's coming for.
11:34
Sure. So I, I think like a lot of
11:36
companies now are looking for the exact UX or form factor to,
11:40
to kind of bring this what you already described, right.
11:43
So it can either be a teddy bear or it can be Jetty BD or it can
11:46
be a wearable. OK.
11:47
What we are playing around with is a cellular phone call that
11:49
you can ideally you know, you get it every day at 9:00 AM or
11:53
every day at 5:00 PM, for example.
11:56
And indeed, like the real value or one of the ways that it could
12:00
provide real value is indeed that it has a very long term
12:03
memory about everything that you ever discussed with it.
12:05
And it can reason based on that memory.
12:09
But of course that there's other other values to it as well.
12:12
Like, OK, it can, you know, interact with you, like talk to
12:16
your problems, etcetera. Memory is just one part, of
12:19
course, of the of the product that we're building right now.
12:23
And I'm really interested on why you decided on the phone.
12:25
And actually I really like it because I feel like the human
12:29
computer interface is going to change radically in the next
12:31
four or five years. I think probably all of us are a
12:34
little bored with rectangles, glowing rectangles.
12:38
You know, I, I remember taking a lot of phone calls and it was
12:40
great 'cause I would walk around the park or I'd pace around.
12:43
But now, you know, at post COVID, everyone's on Zoom.
12:46
So now you got to sit down. You got to look at my, my stupid
12:48
face. And I can't like be distracted,
12:51
you know it. And like walk.
12:52
There's something like, I would say romantic about a phone call
12:56
because you can kind of imagine the other person.
12:58
It's not necessarily you have to sit see them sitting there, you
13:01
know? So why did you decide to make
13:03
this a a phone chat bot as opposed to like something more
13:07
interact or visual on the screen?
13:11
Yeah. So I think one of the reasons
13:12
was because no one really, that's maybe one of the most
13:15
interesting things about it as well.
13:16
Like no one really does work with like cellular interactions.
13:22
You know, with AI, of course there's like companies that that
13:25
do like, I don't know, customer service and stuff like if the,
13:28
the person reaches out through the phone to a certain company,
13:32
like an AI can pick up, but we've we've not seen it yet.
13:35
The other way around. Like a lot of people that do AI
13:39
products that you interact with through voice kind of work with
13:42
like this like glowing blurbs indeed, or like, you know, like
13:46
apps. We think the phone makes a lot
13:49
of sense because it's basically it's like the, the how do you
13:55
say it? Like the the UX that everyone is
13:57
used to Like it's the the UX it doesn't have been around since
14:00
30 years plus ago already, right.
14:02
So everyone from your grandma to like the young generation would
14:06
know how to use a phone. Interestingly, the the phones
14:11
themselves, like the iPhones, iOS on these phones, they are
14:14
also optimised for phone call, right?
14:16
So because we the AI reaches out through phone, you can add
14:20
voicemail like ringtones to it. You can you know it will will
14:23
respond to silent mode, it can leave voicemails, etcetera.
14:27
These are all things that are not built or optimized natively
14:31
for Yeah. For example, standard apps on
14:34
your phone. So that's, yeah, one of the
14:37
interesting things about using the phone route at the same time
14:41
for it. It's like an experience for us,
14:43
right? Like, is indeed the phone the
14:45
most interesting way to interact with AI or does it end up being
14:48
more interesting to just have like a, yeah, an app on the
14:52
phone instead of the the cellular phone call?
14:55
It's something that we're finding out along the way.
14:58
I think also the the most beautiful thing is with a phone
15:04
call you can. So I think that right now
15:07
there's so much UI importance, but what if we don't need UI?
15:12
It's like it's just something there and it just knows you and
15:17
adjusts to you. That's like the vision that
15:19
there is no UI needed because if there's no UI needed, then it
15:24
will just go just set it at one time.
15:27
That's the division. You set it up one time and then
15:30
you just never have to open a browser or open an app again to
15:34
just call you. And just based on what you say,
15:36
based on what you do, maybe add some add more integrations with
15:39
order systems. Then it'll even know that's what
15:42
kind of things you want. Like I say you Google Calendar
15:45
and with that information, you can completely drop the app at
15:50
all. That's new.
15:53
No one does that. So that's why we're thinking
15:56
this, this space. I love it.
15:59
You know, they say that like audio is a warm medium and video
16:03
is a cool medium, you know, so there's something like that
16:06
brings you closer about audio and there's something that gets
16:09
you off the computer. I really like it.
16:12
Now let's talk about these personas you developed.
16:14
What, what kind of like decision making did it?
16:16
Did it? Did it come down to why you
16:18
decided on these specific characters?
16:23
Yeah. So I think like maybe to give
16:25
the full context, it's also like an interesting story.
16:28
Like we started building this product during a hackathon
16:32
actually a few months ago. And the goal of this hackathon
16:36
was or, or, or like the requirements where either make
16:40
money within the 1st 100 hours of of, you know, the start of
16:44
the hackathon or become the most viral product basically.
16:47
So if you wanted to win, you needed to either go viral or
16:49
make money. So this was a 100 hour Sprint
16:53
and basically in the Monday morning.
16:56
So the very first morning of the Hagathon, Mehdi pitched ID of
16:59
OK, let's do AIS that call your phone to do check insurance.
17:04
And then once we settle that ID, we needed to think about how is
17:06
this going to, you know, produce virality in in in one way shape
17:12
or form. And of course, an way to do that
17:15
is to make an AI that's funny or relatable, for example, like,
17:19
you know, your favorite one, the slave bestie or maybe the real
17:23
Sergeant if you need that that extra military push.
17:28
So yeah, quite frankly, we just started brainstorming on these
17:31
on these types like very quickly.
17:32
So we wanted two types that were more appeal to Gen.
17:37
Z. So that will be like Sne bestie
17:39
and high beast and then a few more generic ones like the drill
17:45
Sergeant and the CEOs. And then there we also wanted to
17:48
have one that's a bit more serious, which will be the life
17:49
coach for case. People wanted to actually wanted
17:52
to use it in a more, you know, natural interaction as opposed
17:55
to, you know, kind of communicating with a mean, so to
18:00
speak. But there.
18:03
Yeah. So I didn't we we just came up
18:04
with it like in in a very short time span, basically.
18:07
Yeah. And, and the idea was, OK, it's
18:08
easy to build content around these types.
18:11
And again, and I think like in general, like if you're doing
18:14
consumer AI and you want to, you know, distribute quickly, then
18:17
the pirality is a good way to go.
18:20
So if you can build it in into your agent types or you know the
18:25
way your AI talks, it's yes, definitely an interesting
18:28
approach. Yeah.
18:30
You know, Carl Jung talked about it kind of famously of like all
18:33
these different elements in our in our psyche, you know, that
18:37
like we're kind of an amalgamation of all these
18:39
different personas or like they have what that he's got like 12
18:43
kind of archetypes, personality archetypes and all of them
18:47
coexist and kind of emerge at different times for different
18:50
reasons. You know, so it's interesting
18:52
when I'm thinking about this product, it's like, no, actually
18:54
the characters that they're pulling out are actually going
18:57
to be characters that I want to be personified in me.
19:00
Like there is a time when you need that drill Sergeant, you
19:03
know it, it's for me, it's getting my ass out of bed at
19:06
6:30 in the morning. It's not going to happen now for
19:10
me. And maybe other people relate to
19:11
that. You know, usually it's a lot of
19:13
really negative self talk. That drill Sergeant is really
19:16
mean it it and like says horrific things to me in my own
19:20
head that I'm like, all right, all right, all right, I'm going
19:21
to get up. I'm going to go do some push
19:23
ups, whatever. So the, the the opportunity is
19:27
to like create a true voice of the drill Sergeant.
19:30
That's not that mean. That could call you and be like,
19:32
hey, buddy, these are your goals.
19:33
You already planned it out. These are the things you asked
19:35
me to say. You know, in the future, I can
19:37
see if I'm listening to a podcast, I can be like, OK,
19:41
drill Sergeant, I want you to remember this.
19:42
This is good stuff. And it's just like, you know,
19:45
Goggins going bananas for 30 seconds.
19:47
I'm like, give me this Goggin speech tomorrow morning.
19:49
I need to hear it. And he's like got you so he can
19:51
call you. So there's an element of like
19:54
what you're building has this like very kind of this pedigree
19:58
of psychology where you're emerging multiple people from
20:03
our own subconscious. Now it seems like the challenge
20:05
is is if you really start to develop these personas, the the
20:08
one that kind of sticks out for in union psychology would be
20:11
like the lover. Now ultimately we know that half
20:16
the bandwidth of. The Internet is porn at any
20:20
given time. You know what I mean?
20:21
Like it it, people don't talk about it because it's not
20:24
necessarily polite conversation. But like, the demand for porn
20:28
and those kind of relationships is so great that people are
20:31
going to be like having conversations immediately.
20:33
They're going to start testing the boundaries of like, how can
20:36
I make this a little more risque?
20:38
Like how naughty can this conversation get?
20:41
You know, so everything would probably devolve ultimately down
20:44
to the lover. So when you're thinking about
20:46
that, what kind of parameters, whether that's like sexual
20:49
fantasy talk or kind of violent idiation or self harm, you know,
20:53
when you're creating an AI persona, there's there's already
20:57
documented cases of people going completely off the rails.
21:00
So how do you kind of put protections in there to not
21:02
allow conversations to get into that terrain?
21:06
OK, I've got a confession. As a podcaster and founder, I'm
21:10
constantly meeting incredible people, guests, listeners or
21:13
other entrepreneurs. But here's my little secret.
21:16
I'm terrible at following up. My inbox is basically a
21:19
graveyard of missed opportunities.
21:21
Does that sound like you? You meet someone amazing,
21:23
exchange emails, have great intentions, and then life
21:26
happens. Six months later.
21:28
You're kicking yourself thinking I should have stayed in touch
21:30
with that person. This is exactly why I started
21:32
using Warmstar dot AI and honestly, it's informed how I
21:36
think about my network. Instead of letting relationships
21:39
go cold, Warm Start analyzes my existing contacts and helps me
21:43
identify who I should be reaching out to and why.
21:46
It even writes the outreach messages for me.
21:49
But here's the key. I review everything before it
21:52
goes out. I'm not just talking about this
21:54
because they're a sponsor. I'm talking about it because
21:56
watching a healthcare executive use this to build this entire
21:59
Advisory Board in one weekend before CEO signups made me
22:03
realize I was leaving. Money on the table every.
22:05
Single day your network is your net worth, but only if you
22:09
actually nurture it. Check out warmstart dot AI.
22:13
Use code aiff to get 50 bucks off your first month.
22:17
Stop letting your relationships turn into miss revenue.
22:21
Now you and I know startups are built in boardrooms.
22:24
They're built in the wild during 3:00 AM, bug fixes, botched
22:27
launches, and that one meeting where everything almost fall
22:30
apart. That's what Code Story is all
22:32
about. It's a podcast where tech
22:34
founders breakdown the make or break moments behind their
22:37
products, What went wrong, what went right, and what nearly
22:40
wrecked everything. Hosted by Noah Labhart, Code
22:44
Story goes past the highlight reel and into the honest stuff,
22:47
late pivots, brutal bugs, and somehow coming out alive with a
22:50
product people actually love. If you're in the build stage and
22:54
the scale stage or the please let this feature work stage, I
22:57
can relate. This show will speak to your
22:59
soul. Listen to codestory@codestory.co
23:03
or wherever you get your podcast.
23:04
Just search code story. It'll be.
23:05
Right at the top, because great. Startups aren't born, they're
23:09
forged, one hard moment at a time.
23:11
Code story.co. Yeah.
23:13
So my take is, and maybe you can, you can add to it if you
23:16
have a different take. But I, I believe since we're
23:18
using like the the, the standard foundational models, right, So
23:22
either Gemini, TCPD, etcetera, like we can choose one of the
23:25
foundational models for our AI text generation.
23:29
We're basically adhering to their boundaries as well.
23:32
And I believe if you want to, you know, go of, you know, when
23:36
I go too far or talk to Jetty PD, it won't allow you to do it
23:39
because, I don't know, it will say it's obscene or whatever.
23:42
So thereby we're kind of naturally, yeah, it's already
23:46
built into the system because of it's built into the foundational
23:49
models that we're using. Cool, let's talk about
23:52
foundational models. What are you using to make it?
23:55
Well, first I want to ask you a question before we go into that.
23:58
Like what kind of things would you like to use an an AI that
24:03
calls you on your phone? Like what kind of dimensions
24:06
would you think you would feel helps or in such a way?
24:11
Like where would you see yourself wanting to grow
24:13
together with an AI with? Boy, in all ways, you know, so
24:18
let's say we flash forward five years and my, my air pods are
24:22
just now just like, you know, surgically implanted into my
24:25
ear. And so I can have, you know,
24:28
anything at any time that I want.
24:30
And I want my AI persona to be tuned in.
24:34
Now. I like this.
24:36
I like what you're doing to create multiple people.
24:38
Like I said, I was kind of already trying to do that.
24:41
I think it would. It's going to be, I think it's
24:46
going to be awesome. I mean, I'm trying to think of
24:48
the negative sides. And of course, there's always
24:50
going to be cases of like, oh, somebody fell in love with their
24:53
life coach and they jumped off a Cliff.
24:55
It's like, yeah, well, people eat freaking chair cushions.
24:59
People, people eat balloons. Like what are you going to do?
25:01
You think it's a is fault that people eat balloons?
25:04
I don't think so. I think for most of us it's
25:06
going to be like this unbelievable learning experience
25:09
and growth opportunity to have these personas.
25:12
So if I'm thinking about it like how how would I need help?
25:15
I would like it to be recording everything, everybody at all
25:18
times. So let's say in in six years,
25:22
you know, Mehdi, you and I somehow crossed paths in New
25:25
York City and someone would be like, hey, remember that podcast
25:28
you did? You don't.
25:29
Here's a snippet of it. Well, there's Mehdi over there.
25:32
He made $2 billion with Comidify.
25:33
It became actually me. You should go say what's up and
25:37
then I could go DAP you up. And while I'm walking over, it
25:39
recognizes me. And it's like, that's Ryan
25:41
Estes. You guys had a conversation
25:42
about XYZ boom. You know, I think the
25:45
opportunity for to enhance our memory, you know, give it 100 X,
25:50
you know, and that's not to say that our actual biology is going
25:52
to be affected, but rather now we have data centers that are
25:55
storing every single moment of our lives that will be able to
25:59
enrich every experience and and show connections if you want to,
26:03
you know, so you have that going, you know, perhaps it has
26:06
the ability to then coerce you or talk you into or out of
26:11
decisions that maybe you're not so good.
26:13
You know, we're always kind of forced what they think ethical
26:17
quandaries if whether it's like with our personal relationships
26:19
or in business, that if you have someone who's non biased, that
26:22
has the wisdom of a 10X human that's smarter than you might be
26:26
able to see the future. Maybe isn't, as you know,
26:29
persuaded by money or persuaded in the case of like, let's say a
26:35
therapist. Therapist is always going to
26:36
want you to keep coming back, right?
26:39
You know, or if you're talking to a friend, maybe they have a
26:41
personal stake, an emotional stake in that the relationship
26:44
you're talking about. So they're going to try and
26:46
persuade you in a certain way. It seems like there's an
26:48
opportunity here for kind of like less bias in a decision
26:52
making process when you can talk people through kind of potential
26:55
outcomes. So I mean that that's just like
26:57
a kind of a, a, a way I would think about using it, you know,
27:01
which I suppose is, is a little bit intense because it's not
27:05
real. I'm not talking about
27:06
entertainment, you know, like I'm talking about like, how is
27:09
this going to help me in my life?
27:12
Yeah, interesting. Like there's like a very
27:17
interesting topic here, which right now, because there's is
27:21
there a limit in what's possible?
27:23
I think there's there's the possibilities are limitless and
27:27
they are so big that anything is really possible in this field.
27:31
And we are here at the really early stage in this field
27:34
because this topic of AI voices that sounds so great, almost
27:41
human and natural. It's so new.
27:43
And we are really seeing the space develop so such a way that
27:48
we, we are here to get out of view by talking about this, to
27:54
shape the future of this, how it will grow towards.
27:57
And that's fascinating by itself.
27:59
And I think of course, it's, we also have to really value the
28:04
human connection. So I think also, well, there's
28:07
like a space for this one of you have human connection where you
28:09
talk of all the humans, but at one time you also have like AI
28:13
connection where you talk about AIS.
28:14
And there's like this balance that we have to see what is,
28:21
what's the right balance? And no one knows.
28:24
And that's the cool thing as well, yeah.
28:27
So, you know, if we're thinking about a tech stack, what you
28:30
guys used to build, it's like GPT 11 labs, AWS.
28:34
Is that the basic configuration? This like it doesn't really
28:42
matter, like we can't choose anything.
28:44
So open source is also possible and every nap is possible.
28:49
It's just based on what is best at that more important time we
28:52
pick. And that allows us to be super
28:55
adaptive and also not to reach it.
28:58
But that's also important. Be flexible and be able to swap
29:02
out components when you need it. Yeah, totally.
29:06
And characters, you know, 'cause I, this also reminds me of like
29:09
the classic cartoons, like there's like a devil on your
29:13
shoulder and an Angel on your shoulder, and they're both
29:15
talking and telling you information.
29:16
You ever seen that in cartoons? Because then you could just
29:21
think about, you could do like, OK, like I want to have Jesus,
29:25
Moses, Abraham, David and Solomon.
29:28
Those are the characters I want. And it's like, hey, my my
29:31
predicament is like, I really need to go to my kids football
29:35
game, but also my boss has called me in for a meeting and
29:40
you're like, all right, you guys work it out and then give me
29:43
your response. And then Abraham, David, all
29:47
these people can like then chime in like, Hey, here's what I
29:49
would do. And you're like, no, in this
29:51
case, I kind of agree with Solomon.
29:53
So thank you, Solomon. I'm gonna go with your way.
29:55
You know, these characters become like the real IP.
29:59
You know that the the characters become the important part.
30:02
You know, being first to market and it with a product like this
30:04
is really important because you're going to get to know the
30:07
person, you're going to understand the texture of their
30:10
voice and you're going to want the texture of their voice.
30:12
For example, I don't know what it was 12 months ago or
30:15
something. Siri, the Siri voice changed
30:18
from a lady to a guy now, like I'm like, who is this new Siri?
30:22
I just don't like it. I'm like, I want the old Siri
30:24
back because I had bonded with her terrible directions, giving
30:27
me, you know, maps directions. So, you know, getting this
30:31
right, I think is really important.
30:33
So because people are going to connect with it early.
30:35
So let's talk about the the business a little bit.
30:38
Now you guys launched in for a hackathon, which is awesome from
30:44
from that point till now, like how much has have you sought
30:47
investment? Have you guys bootstrapped it?
30:49
How much have you spent on the project?
30:54
Yeah, so, so far we've basically completely bootstrapped it.
30:57
Beautiful. So, so the like, like Mehdi and
31:01
I, you know, this was both started as a side project or the
31:04
Agathon itself also was just like as a fun little thing to
31:07
do. And then, OK, we actually got
31:08
some organic traction. Like there was, for example, 1
31:12
completely random guy somewhere in the US that found the
31:16
comidified product through a rock search because he searched
31:18
for, I want to have like an AI call agent that, you know, calls
31:21
me and keeps your attraction for like that.
31:23
And then broke maybe based on Twitter posts that we posted
31:26
during the hackathon or LinkedIn proposed to look at humidifiers.
31:30
So this guy made an account. So two weeks after the
31:34
hackathon, basically we suddenly or medicine, you know, check the
31:38
stride balance or, or whatever and suddenly sound like, yo,
31:41
someone is paying for the product.
31:43
And we actually already, we kind of, you know, moved on already
31:46
from it, but then we saw someone paid for it.
31:48
And then we're like, OK, actually we should kind of see
31:50
how we can develop this a bit more somewhat still on a side
31:53
project set up I would say. But in the evenings basically we
31:57
start working on OK, optimizing the product and we kind of work
32:01
towards doing our product on large.
32:04
And that's a launch that we did I guess two months after the
32:07
hackathon and that was very well received as well.
32:11
So we ended the the day on the third spot.
32:15
And then of course gives you know, a fair amount of organic
32:18
users that will, you know, check out your, your product on board,
32:21
on the platform, etcetera. So yeah, that's basically what
32:26
has been going on over the over the past few, few weeks and
32:29
months. We have been talking to some
32:32
PCs, but more like exploratory because we still are kind of
32:35
deciding like which way we want to take it in general, I would
32:39
say. And on top of that, we're still
32:42
also trying to figure out more who actually are the people
32:46
that, you know, are the the big users of our product.
32:50
Yeah, who is our ideal customer profile, our ICP.
32:53
So we're a bit more in the the customer discovery phase, I
32:56
would say calling our customers, asking about the experience with
33:00
the product, why they stopped using it after one call, why
33:04
they actually kept on using it, etcetera.
33:06
We we still need to find out a lot about our customers before
33:09
we can really, yeah, decide on the direction that we want to
33:12
take it. Yeah.
33:14
Cool. And you have a very interesting
33:16
business model too. I'd love to hear how you kind of
33:18
came to that. You have kind of four different
33:20
like pricing options where basically you're pricing it on
33:24
calls per week. So it's like 375 EUR for is 3
33:29
calls a week, you know, and then you have for €15 you get 21
33:33
calls per week. So how did you make the decision
33:35
to go to actually like a calls per week model?
33:42
Yeah. So I I would say like that we
33:45
were kind of wanted to make it comparable to for example,
33:52
coffee because that's kind of, you know, what we were looking
33:55
at. OK.
33:55
Like it should have cost more than a coffee per week, for
33:59
example, to add your AI call you.
34:02
OK, so then let's say that the cheapest package is, you know,
34:06
not more than the amount that you spend per week on coffee.
34:10
And OK, then we end up end up, you know, having it doing it per
34:14
week because if it is per month, then the the number sounded a
34:16
bit too, too large. Again, these are kind of things
34:19
that we kind of, you know, decided on on the fly during the
34:21
heck, it's all like thesis on these things sometimes go and
34:24
started right. So we kind of rolled with it.
34:27
Definitely something we may want to look into right, because the,
34:30
the big drawback of a weekly subscription is that the like
34:36
the the payment is on every week as well.
34:38
So it's kind of in your face, like every week you see, OK,
34:40
you're paying for something. Whereas if you do a yearly
34:42
subscription, that's like obviously a yearly subscription,
34:44
right? Because people paid once and
34:46
they forget about it and the year after we are probably going
34:48
to forget about it again. Of course, like do you have a
34:51
good product? It should it, should it matter
34:52
either way? But these are all details that
34:54
we still kind of need to think about.
34:56
I don't know if you wanted to add something something to that
34:58
Lady. Yeah, I think it feels most
35:03
honest to me as well. If you based on every week you
35:09
are people keep doing that then if they must find it valuable.
35:13
And it also seems like we we let me say right way.
35:24
So with law companies a freemium model.
35:26
So they just have €50 million in this throwaway tried to take
35:29
over the market and but we felt like let's just do something
35:33
right from the start where people have to pay for it.
35:35
So you get it one off, you get 3 calls free.
35:37
And if you want to have another time, just get a rich
35:39
description and you can cancel at any time.
35:42
So in that way, it still feels like a small investment.
35:47
But at the same time it's you can just bootstrap it because
35:53
it's also expensive to run these models, right?
35:55
So this way you can bootstrap it by having a weekly subscription
35:59
and it feels most honest. That's why we also took the
36:02
weekly one. Cool.
36:05
And I, I, I love where you guys are at.
36:07
You know, it's like you launched a product, you're like, oh, that
36:10
was cool. You find somebody's paying for
36:12
it. You're like, oh, let's give it
36:13
some attention, bring it to market and like, see what's
36:15
going on. And sounds like you're in this
36:16
really kind of interesting like information gathering to see
36:21
like where do we want to go with this product, you know, So if
36:24
you think about the next 12 months, what's on the road map?
36:31
Yeah. So I would say that the main
36:35
goal at this point, let's say over the coming three months is
36:38
really defining this ICP. Like we want to ideally find
36:43
like a handful of users that can like maybe 10 people's already
36:46
enough, like a handful of users that love the product enough to
36:50
use it as it is, you know, without like any other tweaks or
36:54
stuff like they just really love it the way it is.
36:56
And these are the people that I think, you know, have an opinion
37:00
that weighs the heaviest on where we, you know, what, which
37:04
direction we need to go, which direction we need to take it or
37:07
which changes we need to make to the product.
37:09
Because, you know, as you know, with like a software product
37:11
like this, we already have a list of like probably 100 things
37:14
that we can think about that we can improve a little bit, but it
37:17
will take way too much time to start working through it, right?
37:21
So we really need to as we need some help with prioritizing
37:26
which probably we're going to tackle first.
37:28
And to do that we just need to have, yeah, some ideal customers
37:32
that that that have a opinion that we can actually value on,
37:37
on this aspect. And the other thing.
37:41
So that's more may maybe on, you know, the side of building a
37:43
product. The other thing for us about
37:45
marketing and distribution. And again, we need these ICP
37:48
customers for it because, OK, if for example, all the people that
37:52
love our products are all people that struggle with ADHD, for
37:55
example, then or marketing is going to be ADHD focused.
37:58
If it's all people that want to do their affirmations every
38:01
morning at 9:00 AM, OK, then we're going to do it completely
38:02
different. We're going to work with
38:03
influencers that, you know, are big in the affirmation space or
38:07
maybe more spiritual phase space, maybe even the, you know,
38:13
some people into union psychology, for example, you
38:15
know, like, who knows? That's like the, the where the
38:18
real, where the real interest is.
38:21
So I guess for a lot of consumer and I mean any adult companies
38:26
comes down to finding your ICP and, and build the product and
38:30
the marketing strategy around these type of people.
38:33
And we, we, we're still really looking for these people because
38:36
that's actually a bit harder than than you would think to
38:39
find them. Yeah.
38:41
Yeah. And there's a couple of
38:42
questions early to ask. You know, it's like, are we
38:44
going to go serious? Like if we're going like we're
38:46
going deep, then like charging on a weekly basis seems totally
38:49
normal because you'd be like, wow, I really look forward to
38:51
these three phone calls. It's really helping me in a deep
38:54
way. Or you go novelty, you charge it
38:56
yearly and it's just some something ridiculous and
38:59
something fun to pop into your day.
39:01
You know, more entertainment was, you know, and then and then
39:04
you're just like, oh, that was cool.
39:06
You know, you just it creates that little burst of joy in your
39:09
day. So and there's so much to so
39:11
much ground or ground to like cover, you know, if you wanna go
39:15
serious it, there's so much opportunity because humans are
39:19
in such a silly predicament. If you think about just
39:21
relationships, you know, what we are is kind of like a a
39:25
hairless, you know, non committal, like ape that for
39:29
some reason our culture and the technology has made us, you
39:32
know, have monogamous relationships, but we're not
39:35
built to have monogamous relationships.
39:37
So there's just innate built in tension to our relationships,
39:41
you know, so like literally like if there was a do one of these,
39:46
you know, one of the characters for for a woman would just be
39:50
like empathetic and want to listen to her day every day.
39:53
How was your day? Was Nancy?
39:55
Was Nancy up to that again? You know, that must have been
39:58
really hard and just was like compassionate and that could
40:01
just let her wind down from her day be amazing.
40:04
And then on the counter side for a, for a man, you know, it'd be
40:07
like just somebody was like, how was your day?
40:09
That sucks. You're doing a great job and
40:11
it's just encouraging and just like reassuring the things that
40:14
like we all kind of want but oftentimes don't get from our
40:17
partners because we're in a ridiculous situation that goes
40:20
against our own biology. The fun thing is, is we know it,
40:23
we know these, these biological things are in place and so we
40:27
can now create characters that would help us.
40:29
So, you know, I, I love what you guys are doing.
40:32
You know, it's it, these are big questions.
40:34
This is a big problem to tackle and it, you've done it so far in
40:38
like a really cool way. I actually really like the
40:40
characters you you chose. I think they're like really
40:43
smart. Yeah.
40:46
Thanks, Hun. Appreciate it.
40:47
Appreciate it. Yeah.
40:48
Thank you. Can we wrap it up?
40:51
I need to go in a few minutes so.
40:54
Get out of here man. I ain't holding you.
40:55
I can talk to guys all day. Now we'll wrap it up.
40:58
Where can people find you online?
41:00
Yeah, we can find me on at least some websites, committify dot me
41:04
for sure. And also you can find me on
41:06
LinkedIn, Medicaid first, or on Twitter also Medicaid first.
41:10
Great. And we will link all that in the
41:12
show notes. Guys, how about how about you?
41:15
Yeah, same thing. So committify dot me it.
41:17
It lists my full name there as well.
41:19
So I don't if people hear my name, they probably can't spell
41:22
it. So you can see it on the website
41:24
and then search at the LinkedIn. And you can, you can shoot me a
41:27
a message or an invite. Yeah.
41:29
Tremendous. Thanks so much for coming on the
41:30
podcast, guys. I really appreciate it.
41:33
Thank you, Ryan. Have a Good Friday.
41:35
Have a good day. This episode is brought to you
41:40
by MO Burst, the AI first growth engine trusted by iconic
41:43
start-ups like Calm, Robin Hood, Reddit, Uber and Google.
41:49
Moburst is a digital full service, mobile first marketing
41:52
agency that helps companies to scale and become category
41:55
leaders. Whether you're finding product
41:57
market fit or ready to accelerate growth, Moburst
42:01
delivers creative, data-driven solutions that help with scale
42:04
and efficiency and impact. At Moburst, we're not just
42:08
guessing what works, we're predicting it.
42:10
As an AI first agency, Moburst uses cutting edge tools to
42:14
analyze user behavior, adapt creative in real time, and cover
42:19
hidden growth opportunities. The result?
42:21
Marketing that's faster, smarter, and built to convert
42:25
from early stage startups to global brands.
42:27
We've helped hundreds of companies overcome complex
42:30
challenges, unlock new audiences, and optimize every
42:33
metric that matters. At Moeburst, we believe that
42:37
creative thinking, advanced technology, and data are the
42:40
core drivers of success at scale.
42:42
If you're serious about scaling your business with precision and
42:45
purpose, it's time to talk to Moeburst.
42:48
Visit moeburst.com and let's build something category
42:51
defining together. All right, all right.
42:57
Coffee is great, but coffee won't prep your.
42:59
Meetings for you. It'll just keep you awake while
43:02
you panic, hastily scrolling through LinkedIn at 6:59 AM.
43:06
Ambient, however, will prep your meetings.
43:09
Each morning, you get a founder tuned briefing, sourced
43:12
overnight from your calendar, your meeting history, and the
43:15
web. It's like walking to every.
43:17
Call with a cheat. Sheet that you didn't have to
43:19
write. No more scrambling, no more
43:21
guessing. Just context, confidence, and
43:24
focus. And if you're raising money,
43:26
ambient investor dossiers are an absolute must have.
43:30
You'll know the investor's portfolio, their interests, and
43:33
their latest moves before you even step into the room.
43:36
You'll even get a heads up on red flags.
43:39
Think of it as the caffeine upgrade that your calendar
43:41
actually needs. Prep while you sleep, perform
43:44
while you're awake. Sign up for free at ambient dot
43:48
US. Go check it out at ambient dot
43:50
US. That's ambient dot US.
43:55
Hey founders, let me ask you this.
43:57
Are you building with AI or getting left behind by it?
44:01
If you're serious about scaling smarter, faster and bolder, you
44:05
need to check out AI for founders.co.
44:09
It's your go to resource for cutting through the noise and
44:12
actually using AI to grow your startup.
44:15
Each week, AI for Founders drops a sharp, no fluff newsletter,
44:20
dives deep with expert guests on the podcast, and hosts monthly
44:24
workshops where you can get your hands dirty with the real tools
44:28
and strategies. Whether you're precede or Series
44:31
B and beyond, launching something new or just staying
44:34
sharp, AI for Founders gives you the edge you need to win in 2025
44:39
and beyond. Hit AI for founders.co to
44:43
subscribe to the newsletter, tune into the podcast, and grab
44:46
your spot in the next workshop, and it's all free.
44:50
Build smarter, move faster, think bolder with AI for
44:54
Founders. Now back to the show.