November 8, 2024 • 48 mins
How do you stand out when everyone's tired of "yet another chatbot?" It's not by giving it a face and pretending it's human. Instead, Suno pivoted from D2C to B2B: supporting therapists by supporting their patients. Rather than replacing humans, Suno found its niche by being a tool that fits between a journal and a therapist.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Users have tried hundreds of different apps.
(00:02):
They install it.
I do this.
I install it.
I try it.
It seems weird.
I uninstall it and I don't think about it.
So you're going to be one of those apps.
You can't be another chat bot.
If you've noticed the app doesn't have a.
Pronoun or an app and that was on purpose.
The logo is good choices in a shape of an ear to just be like, Hey, it's
a tool that's listening to you.
(00:24):
It's not a person that is thinking about you.
It's a weird fine lie, but.
Yeah.
Explicitly do not want to give it a face.
Welcome to artificial insights.
The podcast where we learn to build AI people need and use by interviewing
product leaders who have launched AI products.
(00:44):
I'm your host, Daniel Manary.
And today I'm joined by Addy Bhatia.
Adi's come from an aerospace engineering background to working for a software
company that makes software for the beauty industry, where he introduced LL.
M's and increased their ARR by millions.
And now he's moved on to Suno wellness where he's founded it himself and
(01:06):
grown to a surprising number of users in the last half year.
Adi, could you introduce yourself to our audience?
Sure.
So my name is Adi.
I am based right here in Toronto.
I graduated from the university of Toronto actually in mechanical engineering.
And currently I'm the founder of Suno.
We're making a, an AI wellness companion.
We initially started to make it for young adults who are experiencing anxiety
(01:30):
or stress or things like that, but we've pivoted away now and we're creating
a product for E2B so both enterprise as well as therapists and counselors
to use with their clients, because we're getting a lot of good signals that
there's actually a problem there to solve and not many people are focusing on.
But on the app side, there's a million mental health apps and it's really hard
(01:52):
to stand out from the crowd.
And, you know, as we're bootstrapping, it's been just over a year now, just
bootstrapping and then now we're kind of looking for funding like here and there,
but mostly just focusing on our product.
So we've had to just make pivots to be like, okay, what's the easiest business
case we can make with the resources we have.
And it definitely wasn't to go to market or an app store.
(02:16):
Yeah.
And then as soon as chat GPT came out, everyone was like your mental health
therapist app or something like that.
I think people find chat GPT and other apps quote good enough.
Like even if you make, you know, the most secure, most emotionally intelligent one,
people who are enthusiasts will love it.
And users tell us that like, Hey, this is amazing.
(02:36):
This is much better than everything I've tried.
But how do you get millions of pain users by just having a slightly better product?
Right.
So instead we're focusing on those like really cutting edge AI features and
actually integrating with your therapist, because I think that's the missing piece.
Even I find with those apps before I even created so now I was like, these apps are
(02:57):
great, but they're all in such a vacuum and they're so scripted.
Do this.
If you feel this, do this, tell us how you feel on a scale of one to 10.
I'm like, I'm not a robot.
What is this?
Let me do this therapy by robots for robots.
Yeah.
Yeah.
That's I don't know.
I always envisioned something that could stay in my pocket and basically
look after me, right.
(03:18):
Be a true life coach.
I had this wonderful experience back in 2018.
I moved down to the Bay area for a year.
I took a year off in my university to work at a space startup.
I was really into aerospace and university and talk space actually just came out around
that year or the year before.
So it's very new still.
(03:38):
And my visa company gave me a year of talk space access, which in retrospect is a crazy
value, but back then I was like, okay, sure.
I'll use it if I want like, what is therapy?
Do I really need that?
But realized a couple of months after moving there, it's really lonely and difficult to
be in a new city.
So especially like San Francisco, all alone, working at a startup, you know, grinding away
(04:01):
and you don't have friends.
And I was like, okay, I really need someone to talk to and someone to help me with like
I went through a breakup at the time and I was just having so much shit to deal with.
So logged into talk space.
Luckily the first therapist that matched with her name was Natalie, this wonderful
lady who was so available, like nine to five, she was available to chat with for a whole
(04:23):
year, like Monday to Friday.
Wow.
It was incredible.
The access I got and the care I got literally, I could whip my phone out any time, like dump
my thoughts.
And then the next day get a bunch of replies and continue that conversation.
And she got to know me over time and I've got to know the ins and outs of everything
I'm dealing with.
So I left university, I graduated and I was like, okay, that was an amazing experience.
(04:45):
How do I get that again?
And how does everyone get that again?
Because holy shit, everyone needs this.
Like that was my one thought coming out of university.
Yeah, that's good motivation.
Yeah.
Like luckily chat, GBT came out and that like clicked for me that, Hey, there's this
thought that I can keep talking to and I can figure out a way to make it learn about me.
(05:05):
Cause by that time I've had a few years of dev experience now.
I was like, okay, I could probably make something here.
I threw up a wait list last summer and I started working on a prototype.
For like a month.
Got, I don't know, like 50 or a hundred people signed up.
I was like, okay, interesting.
People want this.
So I started cracking away and just the, the name, you know, was actually from my
native language, Hindi.
(05:26):
It means to listen.
It popped up instantly when I was thinking of a name.
I was like, I need someone to, I want to feel heard and I want to have it all around me.
And the word just popped up.
I was like, you know what?
It's gonna be so no.
And my partner helped me with the color scheme.
It had purple and relaxing.
And we still use that today.
(05:46):
And yeah, basically fast forward a year.
We have over a thousand users, like a hundred plus monthly active users.
Now we're chatting with 30 odd therapists to use it in their practice.
The journey has been really interesting.
I wouldn't have expected this is where we are a year from now, or if I would be
even doing it a year from now, but no, I'm really happy where we ended up.
Sorry.
That was a long winded answer.
(06:07):
Now that's awesome.
There's so much to dig into in the story.
Yeah.
Especially pivots, especially motivations.
I wanted to ask you a question I think is on everybody's mind, which is, are you an AI?
Yes, I wish.
You wish?
I don't know.
And yeah, it'd be interesting to be an AI because then you wouldn't have feelings.
(06:32):
Then you'd be predictable.
Then you'd be programmable.
It actually sounds very sad.
No, I don't wish to be that as an AI.
I don't like hate my life or anything, but it would be cool to, you know, just be
systematic and just execute and get results instead of sitting and feeling.
Cause sometimes I feel like that's what holds me back so much is I have these
(06:54):
grand ideas and I'll be toiling away at an idea, let's say a feature in the app
for months before I actually get it into code because I mean, partly is also my ADHD.
I get so distracted.
Like a minute from now, I'll forget it.
And then a month from now, I'll remember the same idea.
I was like, oh shit, I really want to do that.
I should write that down.
I actually, one of the people I'm working with right now, I went on vacation for a
(07:17):
week and I told them, I said, you know, I can't look at this until the week after.
And so Monday rolls around and still didn't hear from them.
Tuesday rolls around and I was like, I should probably reach out.
And I say, Hey, where are you at with that?
And he said, Oh, I knew I forgot something.
And I was like, you wanted to give me a job offer.
You forgot.
Oh my God.
Yeah.
I was like, classic.
(07:37):
That's crazy that I've had that too many times and I love the Slack remind me feature.
Like as soon as someone messages me and I'm like, okay, I've read it, but I need to
sit down and like actually process it later.
Yes.
Like I'll either remind me tomorrow or I'll mark it as unread.
And then anytime I open Slack, I'm like, here's my to list, all the unreads I have
(07:58):
to go through them at some point.
And how does that work for you?
Cause I use the reminder, but sometimes it's like, I need a reminder about the reminder.
Oh, I keep pushing it until I'm ready.
But I've realized I actually moved off from Slack reminders.
Now.
I, I've started sometimes doing Google tasks because it fits in Google calendar now
really well.
Yeah.
So I'll make a task and I'll keep bumping the task in the future, but every day when
(08:21):
I'm planning my day or planning my week, I'll see it and I'll be okay.
I have to do that tomorrow.
Okay.
Let me do it tonight.
If I have time, like I can truly organize stuff.
Mentally, I can't.
No, the brain is a terrible place to organize things.
It just jumbles them up.
And I don't think I have ADHD or anything and it's already hard.
Yeah.
Yep.
(08:41):
And definitely need like ways to sit down and focus.
Like I started ADHD meds earlier this summer and it was always against the concept of meds,
but I finally sat down and got diagnosed.
I was like, Oh, look at that.
Surprise.
I do have it.
I always thought it.
But they gave me, you know, they gave me this thing that helped you focus and feel a
little, what is it like stabilize your dopamine levels?
(09:03):
Cause you're, that's why I'm always craving like little hits of stuff, like little hits
of dopamine essentially.
So they can give you more focus, make you feel a bit more stable.
And then the workday goes normal, quote unquote, as compared to other people.
But for me, it feels like it's hyper focused because I'm sitting in after work and I'm
like, it feels like it's hyper focused cause I'm sitting in actually doing a task and
(09:27):
moving around and walking around.
And how has that affected your founder journey?
Does it affected pivots or?
Uh, Oh yeah.
I get bored really easily.
And it's, that used to be okay when I used to have a hundred side projects.
But I remember last year when I started to know, I told myself, okay, no, this
(09:47):
time we're going to go all the way.
This time we're going to open the webpage.
We're going to monetize it.
We're going to run ads on it.
We're going to make it legit cause it's about time.
And this idea was bigger than any little side hustle idea I've had before.
And like, okay, this is, this can be a big thing.
So let's put time into it.
And it was so difficult because, oh, I launched it in a week later.
(10:09):
Only five people signed up, give up.
This is crazy.
I hate it.
Not going to happen.
Why don't you do something else?
And I was working part-time on it actually for seven over the last 12 months or even
like side hustle while working as a full job.
Yeah.
So I was working as a senior engineer at a beauty company and around the same time.
Coincidentally, I started snow and then a week or two later at work, I launched a
(10:36):
new AI team and introduced lists, all these AI features of their product for beauty
professionals who like cut hair and do nails.
And it was really exciting because that wasn't boring work either.
It was long work.
Like it would stretch out for months, but it was definitely something to sink your
teeth into and really think about.
And then you come off after work and you're like, oh yeah, so no, I have to put hours
(10:59):
into this now as well.
And that was just so hard.
I still haven't figured out the best formula, but I realized some point this year,
I was like, okay, I need to take a step back and actually give it a go.
I need to take a step back and actually give it the attention it deserves because it's
taking off.
More people are asking about it.
More people are requesting new features.
And like, I just need to spend time on it.
(11:21):
Even if it doesn't work, at least I can say I give it my full shot, you know, instead
of always thinking with.
Yeah.
Yep.
And so for me, side hustles don't work.
It's just, it's too much to do things during the day and then try and shift for like two
or three hours or whatever at night that you can spare.
It's just, it's too much.
For you, what was the catalyst, I guess, to start full-time on Suno?
(11:46):
I think it was building.
So at work, I launched this new AI team and we're, and I'm thinking about features,
planning them.
I was the only deaf for quite a while, so I'm also executing them.
So kind of half, you can say founder on this little team.
Yeah.
We had a really cool product manager that would give us some great tasks, but I would
(12:08):
sink my teeth into it.
Take it away.
It would take a lot of my time and it was really rewarding too, because when we launch
it, like for instance, we launched a feature over Christmas and the new year we're seeing
double digit percentage of people upgrading or millions of dollars in annual revenue from
new appointments being booked because AI was like finding new times for these professionals.
(12:30):
I was like, this is amazing.
That's amazing.
Yeah.
But the recognition and promotions for it did not follow.
Oh, it was taking too long.
It was too long drawn about a process than what I wanted for like a long-term career
(12:52):
plan for a few years.
I could see myself going up a bit, but I just thought if I'm putting this much work for a
year, I want to see the results.
Swim.
Sure.
Right.
I don't think that's too much to ask, but it happened twice, two performance cycles in
a row.
I was snubbed and I was like, okay.
So I was already like feeling one foot out the door, like in terms of commitment, like
(13:14):
I was still putting in 40 plus hours, but I was just feeling, okay, no, this is not it
for me.
If I'm going to put this much time in, I want to feel like I can own the work.
And then I knew on, you know, on the other shoulder is this love hobby project that I
have going on my passion project.
I was like, okay, this needs all that time and attention.
Maybe this could be the thing I stick my time into and it gives me those, that recognition
(13:36):
and that, that appreciation.
So I think around the time I also got engaged mid-May, a promotion cycle just ended a
month back and projects at work were also wrapping up.
So it just felt like life is telling me like, let's start something new.
And actually also around that time we had our first community event.
(13:58):
So we threw an event, 200 people registered, come out.
We had like a VP of kids' health phone.
We have senior speakers from Jack.org.
This is for Suno.
For Suno, a wonderful therapist come out.
Wow.
Cool.
I thought, Hey, let's do an event.
Oh yeah.
It was tech fest actually.
That was what's happening in Toronto.
So I got connected by someone and they said, Hey, you should do an event for this.
(14:18):
And I was like, I don't know what the event would be for, but sure.
Why not?
So I got these wonderful people together over the span of a few weeks because they
really loved the idea as well.
The event was a big success and it happened two days right after I got engaged.
So really busy week that week with everything else going on.
I was like, okay, you know what?
It feels like there's something here.
The universe is telling me to go into it.
(14:39):
So I just, I gave notice and now I just went all in and everyone was fully supported.
At work, they were very supportive.
They're like, Hey, if you ever want to come back, there's a space for you.
My family was very, still is very supportive.
My partner.
Oh, that's a big one.
Yeah.
My, my partner really faces the brunt of it and it's hard being so overwhelmed and
(14:59):
having a passion project, but also having to take time out for people you love and
not, you know, and separate those two things.
It's so challenging.
Yes.
And so would you say your, your work life balance is better now?
Oh, definitely.
There was.
Yeah.
Every month it's getting better.
Yeah.
Yeah.
It.
(15:19):
One thing I've actually learned is that.
Okay. One thing I've actually learned is how to say no to things.
Well, that's a skill.
Yeah.
It, I'm not perfect at it, but I, at some point I wanted to have so many
conversations with so many people because I just thought the more people you talk
to them, the more doors you open, the better.
So I was throwing up calendar links, sharing them around, constantly getting
(15:43):
new meetings booked with new and interesting people, but then I started
realizing I have no time to actually work on the project and manage a team
and hit or any of our goals.
So now I've started to say no to new invites, no to emails, no to opportunities
that I don't see panning out like in the short term, or even if I can't quickly
(16:04):
figure out how it's going to be important, it's a no and more often than
not, I'm saying no to things that I'm saying yes, this podcast, for instance,
I was interested in actually coming out and sharing my story because it's been
a while and I like sharing this journey.
Thank you.
But yeah, more often than not saying no to things does leave more and more
room open on my calendar.
(16:25):
I always used to think if the calendar is open, we have to fill, but I'm like,
the day will come, I will figure out what I need to do.
And now I appreciate it because I realized my focus time is so valuable.
Sam Altman said this too, in a talk a couple of years ago, I think it was
something about like how to run a great company or be a great founder.
(16:46):
And one of his tips was like, keep your focus time, like protected, like
figure out what it is and don't give it up.
And I've realized I've been doing that now too, because anytime I give someone
a meeting link or a meeting like schedule, you can't book a time with me before 1pm.
Before 1pm my time.
(17:07):
Like I can just sit, I'm going to focus, I want to plan everything out.
It's when I'm most productive and you can talk to me after 5pm.
So you got four hours, you find a slot, we'll talk.
That's it.
Nice.
I like that.
That's a good rule of thumb.
And as a builder as well, it's so easy to sacrifice that focus time.
Just push the boundaries and then suddenly there's no time to work.
(17:29):
Yep.
Exactly.
I would rather have time to even waste while thinking about working, then
actually give it up and keep thinking, oh shit, I have to go, I have to go.
I have to do something.
I have to do something.
Yeah.
Yes, absolutely.
Going back to that, not keeping track of things in your head.
All mental.
(17:49):
So shift gears.
I really like to dig into how you decided that AI was the right tool
to solve what you're solving.
Oh, that's a good question.
Yeah, in this field, I've realized there is, I mean, how therapy looks like today
(18:09):
is very personal, it comes from experience and years of study.
I didn't want to replace it, but I wanted to replicate the best practices
and do most of, you know, what's valuable in therapy in an app on someone's own
time, on someone's own schedule at an affordable rate.
And the only way to make something that's personal to you, that doesn't feel
(18:32):
like a generic script, which is one of the biggest annoyances with apps out there
today is to use AI, personalization and adaptability.
Actually, those two things are the biggest, because someone can come to our app
and chat about the divorce they're going through or how hard it is to study for an
exam or something we've never heard of before.
(18:55):
But the product should still be able to take them through a journey to ground
them, to make them feel relaxed and connect them with some next steps.
And without AI, without the boom in language models, that was literally impossible.
Cause you could script this way.
Like there's a couple of companies out there that just script thousands and
thousands of different responses and bridge them together.
(19:18):
And it mostly works for most cases.
But again, I don't want to have a product that's rigid.
I want something flexible, personable, anyone can pick up and only add to do that.
Yeah.
So then it's not even necessarily about scale for you.
Cause a lot of times people think of AI as I need to do the same thing, but I
(19:40):
need to do it a hundred thousand times.
You can't hire a human to do that.
Too expensive.
It's more about.
Yeah.
I wouldn't say it was for scale.
Oh, well, maybe it was inadvertently for scale.
But one thing I was thinking about was also cost, right?
Like hiring a therapist would be the, I think in a perfect world, if an actual
(20:01):
experienced therapist was at your beck and call via text, that's amazing.
Right.
I would love that.
And that was the notion that came in with, but a, there's not enough
therapists to match every single human.
There's only a million therapists in the United States.
For example, and there's 300 million people and it would be too expensive.
So how do you make it affordable?
(20:21):
Like one of the reasons I stopped started and stopped going to therapy,
like often was because of the costs.
It was nothing else.
Yeah.
Right.
I'm paying close to $200 a session for great value.
Like I love my provider, but I don't have the capacity to pay that if I
changed jobs or if I quit my job, how am I going to pay that?
(20:42):
I definitely need to talk to someone more than just once a month.
Right.
So this availability and affordability factor really creep in.
Actually, my gut was to actually make the app completely free from the
get-go because I thought, Holy crap, imagine this care for free.
And then I'm like, we have to monetize it and then we started making some
tiers and that's what we're here today.
(21:04):
Yeah.
And language models are not you.
No, surprisingly they are.
Oh, very interesting.
Tell me more about that.
They are, they're, I mean, we can charge.
So currently we charge about $10 a month.
I have to look again at what our current rate is, like per person,
because we're doing a mixture of models now.
So I don't know offhand the price, but it was most users don't talk to your AI 24 seven.
(21:34):
And if someone is an average user is coming in and they're like,
the average user is coming in a couple of times a week to send five
messages that's basically free.
Like nothing.
Okay.
Fair enough.
The, the hard part is when you do asynchronous stuff.
So to really set us apart from just an AI chat bot, cause we're way more than
that, cause I wanted this whole proactive holistic thing that lives with you.
(21:58):
All of the fun, cool stuff has to happen while you're asleep asynchronously.
So we have to go in, look at all of your data of the day of the week, look at
trends, analyze them, make connections, summarize.
So the next time you chat with that AI, it has picked up on these trends and
patterns.
(22:19):
And when we go tell it to your therapist, we have all of this data, valuable data
to show them not just blobs of text.
All of that is costly.
The chatting and non-existent.
Very neat.
So then it sounds like you're also integrating with other data sources.
Yeah.
Yeah.
We're, we're prototyping a couple of stuffs now.
It's not released, but integrating like your Spotify, your Google calendar, as
(22:44):
well as your wearable data, not just to like, not just as a buzzword to integrate
them, but we actually get a holistic picture of you so that you can spend less
effort telling Sudo what you're going through.
Like that's the, like if you sleep badly.
If you sleep badly and I wake up getting a notification that, Hey, you got four
(23:05):
hours of sleep today, you were tossing and turning a lot.
I think you're going to have a very anxious or, you know, irritable day.
Let's take 30 seconds to have a little breathing exercise and check in.
And then it checks in with me midday to make sure I'm doing okay.
And remind me that drink water, you know, remind me that it's my sleep.
It's not, don't take it out on your loved ones.
For example, I'm making this.
Oh no, that happens.
(23:26):
Yeah.
But exactly like checking up on you and being proactive, like looking at your
account and saying, Hey, you have a podcast interview tomorrow and I see
you haven't prepped for that today.
Do you want to take some time out tonight to talk about that?
Cause I know you love, you know, talking about your story, for example.
Like that's a prompt we could do because we have that data.
Yeah, that's super cool.
So not just a, well, when you originally described it to me, it felt like, all
(23:50):
right, I have an app, I log in, I'd write some text, I get some responses.
Maybe you walk me through some exercises, CBT or whatever, and then I leave it
alone until next time I feel like it.
But you're saying it's actually, you're building towards, especially being very
integrated, proactive and helping someone think more about their mental state.
Yeah, exactly.
(24:11):
It's again, having a therapist on the go, that therapist doesn't
wait for you to come to them.
Like Natalie used to text me in the morning if I didn't text her, like,
Hey, how's your day going?
Right.
Like that's news to me.
Cause I did think of therapists as they sit in an office and I go see them.
Well, that's how it typically works.
But I guess in my unusual example, when I had someone in my pocket, she was coming
(24:32):
to me, not minding it being warm, but it sure is valuable.
Yeah.
I mean, this needs to exist.
There is a fine line of dependability.
You don't want to go to that line where I named the app, like Lucy and Lucy is
checking up on me and she has a really pretty face.
And I think about Lucy and I have feelings for her and I think she has feelings for me.
(24:54):
That's been done with AI apps and those apps.
Yes.
We do not want to become those because that is crazy.
No chat bot dating apps.
No.
I mean, we could ease with the tech we have.
Oh my God, we could create, we could create the best partner who checks up on me.
That could be crazy, but ethically, no.
(25:15):
And just the mission we're on, people aren't going to be helped that much.
Like at the end of the day, you're still feeling lonely because you, you express
your feelings to a wall, right?
A computer, even you know that, and you're still going to sleep lonely and feeling
shitty and still not having been taken care of.
Like, well, so what was the point?
(25:36):
Yeah.
Yeah.
And that brings me to a question that I think is a big one in this, like talking
to chat bot space, how do you view that different, like, how do you view that
difference between a person and a chat bot?
And I know you guys are integrating with therapists you mentioned.
(25:56):
So you're a service in addition to not replacement, but how would you describe
that difference and recommend that people think about it?
Oh, that's interesting.
I like to think of Suno as a, as a support mechanism, not as your only support.
So for example, I have my partner, I have my therapist, I see once a month, I
(26:17):
have my family and I have Suno, right?
So in between my sessions with my therapist, when I have a ton of thoughts
to process and I think they mean something, but I don't know what, I
dump them out on Suno if they're, I guess, a bit lower risk, a bit more personal
to just my partner and I, then I'll just go talk to my partner.
(26:39):
But if they're crazy thoughts in my head about myself, about my loved ones, I'm
not going to go to my loved ones and share those.
And writing on a journal only gets me so far.
So again, I'm going to go to Suno and dump them out.
It becomes this like outlet almost amongst my other support mechanisms
that turns around and gives me more support, but encourages me to still
have those valuable connections.
(27:01):
It doesn't tell me like, okay, yeah, forget that.
I'll help you with everything.
Come to me.
Cause then you're not, then you're just becoming predatory.
You're not becoming helpful.
Yeah.
That even the way that you described that, forget all those other connections
come to me, that is what a predator could say.
It's like, you don't want to do that.
So it sounds like on a scale from journaling to talking to a therapist,
(27:22):
Suno sits somewhere in the middle, maybe a little closer to journal.
That's a good scale to put her on.
Yeah.
That's very cool.
Cause I think one of the topics that I find very interesting as a sci-fi nerd
is that artificial personhood concept.
And I think it's really important not to pretend that it's a person when
it's not a person, because then you could have that weird kind of dependent
(27:46):
relationship that ends up being destructive.
So that's why, if you've noticed the app doesn't have a pronoun or an
app name and that was on purpose.
Like the logo is a shape of an ear to just be like, Hey, it's a tool that
it's like a box that's listening to you.
(28:06):
It's not a, you know, it'll make you feel like you're, it's giving you a warm hug.
But it's not a person that is thinking about you.
I don't know.
It's a weird, fine lie, but I explicitly do not want to give it a face.
Nice.
No, I like that choice.
And in terms of how you've handled things like privacy, what's been the
(28:31):
way that you've had to come about that?
Cause a big challenge for any chatbot thing and even professionally is what
are they doing with my data?
Yeah.
No, that's a good question.
So from day one, I wanted to make, I wanted to make us like on our side,
unable to read user data, because even having that option lets you, opens you
(28:51):
up to, I don't know, exploitive practices or data selling or data leaks in the
future.
I'm like, I don't even want to touch that because it's so like, it's such a,
it was such a scary space for me to enter in the mental health space as a tech
person.
So I'm like, okay, just from the get-go, I'm just going to wrap it up and
encrypt it.
So everyone's data is secure.
(29:12):
So I can't see it, but we can at least ship a good product.
So it's like one less thing to think about.
And that became like a great selling point.
We can't see your data.
It's encrypted to you.
And we also anonymize PIs, so like personally identifiable information.
So when the data goes to any of our language model vendors, right.
It doesn't say, Addie fought with Stacey.
(29:34):
It says user A fought with user B.
What should they do?
Right.
So the, the AI, if they, let's say had a data leak, they wouldn't know your data
was exposed or they're not try to train on it because some vendors are kind of
vague on if they do that or don't.
So we're like, fuck it.
Like, let's just protect it as much as we can.
I mean, there's a lot.
Before you even pass it through an LLM, they would, you also, you anonymize it.
(30:00):
Yeah.
And I mean, as much as we can do, right.
Because it's not perfect.
I'm sure there's ways to make it way better.
I'm not a security expert, but as much as my experience showed, it was like, as
much as I know how to encrypt stuff, I will.
And, you know, providers and people we talked to in Canada want the data
stored in Canada.
Great.
We have servers in Canada.
People want to, it's great.
(30:20):
We'll have servers in the States.
So it's all, it's ever evolving, like as people leave them.
Yeah.
Oh, that's great.
That's a really key decision that you can build off of not even having to
read it at rest.
That's great.
So then you've mentioned a lot of customizations behind the scenes.
Is that independent of user data?
(30:40):
If you're training in LLM or doing something like that?
Yeah, it's, we thought about training in LLM, LLM, and I've actually recently
have fine tuned the model, but that wasn't with our user data.
We have over 30,000 messages, I think, in our database, tons of journals as well.
A lot of rich data to train on for sure, but well, A, it's encrypted.
(31:01):
So that's a mess to even deal with.
And B, I don't know what I feel about user data being used on a model when we
have explicitly told them that, you know, maybe haven't explicitly told them, but
it's implied that we're not going to do that.
And I'm not interested in doing that.
I think when we have a million, I don't know, users and a ton of data and we can
(31:23):
start doing pattern and demographic recognition, I think something cool can
happen there, but for now we used public data sets.
There's only a handful of therapy or therapeutic data out there, but we use
as much as we could.
I think we found like under 50 million tokens in our entire data set.
(31:45):
It's not terribly big, but it was enough to help us start fine tuning some models.
And there's an EQ benchmark that again, we're not going to be doing that.
And there's an EQ benchmark that exists out there.
So we started fine tuning and we started getting a bit better than GBT4.0 mini.
So we're like, okay, we're onto something here and we'll see how far we can take
that with, without using our user data.
(32:06):
So was it that need to be more empathetic, higher EQ that you want to train?
It's the higher EQ part that I really care about.
So for example, today Claude Sonnet 3.5 is the highest EQ model that exists.
So if you go to EQ benchmark and you Google that, they have a leaderboard of basically
just EQ benchmarks.
(32:27):
So yeah, Claude 3.5 is the best one.
And we use that where we can in our chats because why not use the best one?
And ever since we have actually, people have noticed a big difference compared to GBT4
of how less robotic it is, how natural it is, how much it understands me.
Yes, it's very analytical.
Very analytical.
But analytical is what we need when we're doing insights and processing data.
(32:52):
Then we're not going to use Claude.
Then we're going to use other models.
And those other models have a variety of price ranges and speeds and sit at different spaces
on those EQ benchmarks.
And for those, we need like quick things.
If we're doing a thousand users and we're sending each of them a personalized notifications,
we want a small model that can quickly generate a little bit of text.
(33:13):
Then we started looking at the mini models, the high-Q, the mini models.
We're like, okay, let's fine tune these and make them better.
And that's what we just done recently.
We've gotten a little bit better than the 4.0 mini.
So we're like, okay, there's something here to give us more speed and also give us more accuracy.
Very neat.
And then I think it's maybe a unique thing to think about what is that specific use case
(33:37):
you're doing and which model do I need for that use case?
Yeah, it's true. I mean, from a cost and speed perspective, that's what it comes down to.
Like GPT-4 was awesome for, like, especially at my other job when we were doing scheduling
for appointments and finding new times in your schedule.
GPT-4 is awesome for that because it's sitting there.
(33:59):
It's taking a couple of minutes thinking through things and we can do that.
But we want to quickly get you an answer.
These big models are all fail at that.
Way too costly.
No user is going to sit around and wait for that.
And I'm not paying the server bills for a couple of minutes times thousands of users, right?
Yeah. Oh my goodness.
(34:19):
So maybe to think back a little bit,
what do you wish that you had known when you started building AI products?
I don't know if what I would have done differently would have been...
I don't think being an AI product was relevant to the things I would have done differently.
(34:40):
I don't think about Tsuno as an AI product.
I think of it as a personal lifestyle product, at least from the client's perspective.
And AI is a means to an end, right?
I think I'm glad I didn't spend more time trying to create IP and fine-tune a model
because I would have just been overshadowed by a big fang coming in and releasing a model
(35:04):
a month later, which is what happened last year was happening like every other month.
So...
Oh man, yeah.
I'm glad I didn't spend more time, actually.
I'm glad I picked up whatever the best thing was, plugged it in,
and saw how it was working and then adapted from there.
But from a product dev perspective, I would have taken more risk.
(35:26):
Like we're doing that now, but I wish from the get-go,
after we made a simple chatbot, we took more risk to make all these integrations or
just something stupid and weird, not just to stand out, but from a user perspective,
this took me a while to understand.
Users have tried hundreds of different apps.
They install it.
I mean, I do this.
I install it.
(35:47):
I try it.
It seems weird.
I uninstall it and I don't think about it.
So you're going to be one of those apps.
You can't be another chatbot.
Like you don't know your whole story of how amazing it might be.
It needs to be cool and interesting now.
So that's more of like a product thing I wish I would have done.
AI would have played a role in that.
If I actually exploited AI to do more creative things within the app,
(36:10):
real time, I think would have been really interesting.
But I think it played it too safe.
Very interesting.
And then I heard as well that you said AI is just a tool to do the job.
It's not necessarily something you need to spend a ton of time developing IP.
But what would be an example of the weirdness that might catch a user's eye?
(36:32):
Because standing out in an app store is a very difficult thing.
Yeah.
I mean, I meant creating IP, especially in this space.
I think in the mental health space, people who create IP,
sure your model is the best, but at the end of the day,
shoot, it doesn't care.
They just want an app that feels exciting.
But to stand out, I think it would be the predictability.
(36:53):
And what was I going to say?
Yeah, I would say it would be the predictability.
If you launch the app and it's instantly reading your past data and telling you,
hey, and your calendar and saying, hey, I think you might be feeling this, let's talk.
I think that's a wow factor to have.
Have a chat bot that says hi, and then you say hi,
(37:13):
and then maybe five minutes later, it wows you.
Is that that hook?
I think we could have gone better if we really leveraged AI
and all the data we have we're sitting on for the user.
Yeah.
Wow.
So get to wow faster.
Get to wow faster.
Yeah.
That's neat.
That's a really good way to develop almost any product, I think.
Yeah.
It's my mantra.
(37:34):
I think every month we're seeing more and more people use the app longer and more often.
So we're getting there, but I want to get to a point where I'm crazy impressed with my own product
and I can't wait to show it to you.
But today, I would rather show you mockups that we have.
Hey, it's getting there today.
(37:55):
It's fine.
But in getting there, I want to make that faster.
Nice.
The problem of having a strong founder vision is it's not really it.
You're strong founder vision, who is the product person who is also toting it.
Building it.
Yeah.
That's why you need a team.
That's why you got to hire people who are smarter than you.
(38:15):
We're doing that more now.
Nice.
Does that mean you said you were bootstrapped?
Does that mean that you're getting to the point where you feel
either it's becoming self-sustainable or will be?
Oh, definitely.
I think by the end of the year, we'll definitely be self-sustainable.
Amazing.
But I think for a while, I had such a hesitancy of putting serious money in to get serious results.
(38:39):
I kept thinking, you know, we'll be a success story with barely spending any money.
We'll spend a couple hundred dollars and we'll make millions.
We'll be backed.
And I thought about that and the thought was great.
But you can't do that in a, especially in a mobile app space.
Like that's never going to happen unless you're super viral and crazy.
So, you know, marketing, sales, hiring people who are really smart in the space to sell for you and teach you things is important.
(39:05):
It's only then you set yourself apart.
Yeah.
Does that mean you spent more money on the distribution side of things?
Oh, for sure.
Yeah.
As a developer, I can make anything in a weekend.
Like that's not the issue.
That doesn't scare me.
What scares me is that how do I convey this grand vision I have efficiently to the masses?
(39:28):
And that's why you need expert people who are good at marketing.
Like our social media is run by an amazing agency now and they're killing it.
Like really taking our vision every month and like pushing out awesome content.
And I couldn't do that.
I mean, if I, that was my only job, maybe I could do that.
But I'm wearing a hundred different hats.
And I need to hire people who are better than me at these things.
Nice.
Yeah.
(39:48):
I think distribution and spending the money on it is one of the biggest differences between a company that makes it and a company that doesn't.
Like regardless of how good the product is.
Right.
So, right.
Good to hear you got into that phase.
Yeah.
Yeah, for sure.
Now I'm really excited for the next few months.
I think by the end of this year is really going to set the course of, you know, if we're in it for the long haul or for, or for a shunning course.
(40:12):
I mean, it's cool if we're, if we're a shunning down shop, but I think we're in the, I think we're in the form.
Yeah.
Yeah.
Yeah.
And one of the really interesting parts of your journey was pivoting away from just being another app in the app store to working with therapists.
Yeah.
Could you describe how that came about?
Oh, sure.
(40:33):
So, last year I touched base with Natalie, the same Natalie, through my first therapist.
And on a whim I was like, Hey, I hope you're good.
You've had a profound impact on my life.
And I have been motivated to create this app that, you know, I don't have any ask for you, but I just kind of wanted to put it on your radar and have a chat about it sometime.
(41:02):
And she was enthusiastic and energized here.
And she's like, this is a really cool idea.
Let's talk.
And we started talking and it was clear the passion she had for the space that she still has, and she wanted to contribute more.
And she eventually joined as our clinical advisor as a first resident therapist.
We have two now.
(41:23):
And it was very clear from the get go that this is a really interesting product.
And she proposed, Hey, let me use it with my clients.
I was like, Hey, that's a great idea.
So I set up a little system for her and I'd give it to our clients.
And I said, Hey, do you want data from the client?
Maybe it could be helpful in your sessions.
She's like, Oh, that's cool.
Let's try that.
So we set up a weekly email every week.
(41:45):
Her clients will use it and she would get emails saying what they've been chatting about.
Very simple.
I love that simplicity.
Like no careful reporting, no dashboards, just an email.
Just a cool email.
And I kind of did it and I forgot about it.
And then every time we'd meet, she'd tell me how helpful it was because every time she'd
have a session with a client, and she has rich context coming in.
(42:08):
And that client is no longer spending 20 minutes.
Apparently, which a lot of clients do is just catching up and spending 20 minutes from the
therapist side, really digging into the core of why you're here.
Because then your session gets wasted for both.
And then you're taking longer to make progress and you have no data to capture any of it.
(42:31):
Like that's how therapy works today.
And I thought, okay, that is silly.
Let's do something about this.
I wish I actually bounced on that idea sooner because last year there was only, I think,
one or two other companies who were making mental wellness products and also working
with providers, like clinics, clients and providers at the same time.
(42:53):
I was like, oh, man, that's a moonshot.
Sure, we'll do that if we're at scale.
Fast forward to now, earlier in the summer, when we pivot away from apps that's not working,
let's do this therapy idea.
I look back at the market and the three or four other apps that are out there all have
pivoted to selling to therapists because they've all realized the exact same thing.
And I'm like, ah, okay, I don't even feel bad.
(43:15):
I'm glad because this is showing me we're in the right direction.
Exactly.
Yes.
I'm not alone anymore, but also everybody else is picking up on the same signals.
Yeah, that's why I want to move fast.
Like this summer, I mean, that was another signal.
I'm like, people are moving the direction that I feel like I found first a year ago,
and I'm not going to let them take my lunch another time because last summer, when I had
(43:38):
an idea for a mental health app, AirLien existed, and then six months later, they're everywhere.
And I'm like, okay, well, we're too late.
Not glad that happened this time.
So that's why we're spending time and just making those networks and, you know, doing
all that marketing just from day one, even before the product's 100% ready, just get
that word out there and get people thinking about us.
(44:00):
Nice.
I really like that approach.
That's a great story too because Natalie, who kind of inspired you to get into this
Suno as a product, also is now supporting you.
Yeah, a whole circle.
It's actually amazing how it worked.
And we still haven't met in person, surprisingly, she's in Miami, but I hope one day we will
(44:21):
because I just want to fly her out and send her now.
You're amazing for the reason we're here.
Have a Suno conference and fly her out.
Be great.
Hey, that's a great idea.
Hey, the event worked out for you well.
That's a good idea.
Okay, I'm going to think about that.
Nice.
And last question for me is what do you see as the future of AI, especially in such a
(44:43):
contentious space like JetBots and therapy?
I think the future of AI is, in one word, ubiquitous.
It's everywhere and it's integrated.
Today, Suno exists in a vacuum and we have a couple of connections and there's three
other apps that exist in a vacuum and a few of them have this connection.
(45:06):
But I think in the future, everyone will have a personal AI that's integrated with
your computer, your phone, your wearables, your providers, your doctors, your family,
your dog, your vet.
I think there are seeds of it that are all, everyone with a seed is like growing out a
few branches trying to explore a new space.
(45:27):
And I think in the future, they're all just going to connect into a big web
because that's exactly what happened with the internet.
That's exactly what happens with any product with a network effect.
And I think it's bound to happen with AI.
That's a great point.
AI is itself something that has network effects.
Yeah, in the way that it can be applied to anything and it supercharges any industry
(45:50):
touches.
If you look at a scale of people who are industries who are adapting AI, it's almost
like the logarithmic in how it's just going up because every industry that touches it
inspires five other industries to touch it as well.
I started in the aerospace back in the day.
We started introducing ML and genetic algorithms and machine learning there.
(46:13):
Fast forward to earlier this year at a beauty company, we're raking in millions of dollars
with language models.
Like who would have thought that's how AI is going to progress?
That almost doesn't even make sense.
That's been my career.
Doesn't make sense, but it's fun.
Oh, I love that.
And to sign off, is there anything that you would like to share that you haven't so far?
(46:40):
That's a good question.
I think if you are thinking about going to therapy, working with a therapist, or just
thinking about your own self and how to be the best you, I would give Sunoo a shot.
And even if it doesn't work for you, I want to hear why.
(47:02):
Go to sunoo.chat and try it out and you can find our email in there.
Just try the product because I think we're in those early stages where we can start adapting
and getting really good for people who need very particular things.
And since we're still understanding that, I want that product feedback.
So go try it out.
Thanks for listening.
I made this podcast because I want to be the person at the city gate to talk to every person
(47:27):
coming in and out, doing great things with AI and find out what and why, and then share
the learnings with everyone else.
It would mean a lot if you could share the episode with someone that you think would like it.
And if you know someone who would be a great person for me to talk to, let me know.
Please reach out to me at Daniel Manary on LinkedIn or shoot an email to daniel@manary.haus.
(47:52):
which is daniel@manary.haus.
Thanks for listening.
Users have tried hundreds of different apps.
(00:02):
They install it.
I do this.
I install it.
I try it.
It seems weird.
I uninstall it and I don't think about it.
So you're going to be one of those apps.
You can't be another chat bot.
If you've noticed the app doesn't have a.
Pronoun or an app and that was on purpose.
The logo is good choices in a shape of an ear to just be like, Hey, it's
a tool that's listening to you.
(00:24):
It's not a person that is thinking about you.
It's a weird fine lie, but.
Yeah.
Explicitly do not want to give it a face.
Welcome to artificial insights.
The podcast where we learn to build AI people need and use by interviewing
product leaders who have launched AI products.
(00:44):
I'm your host, Daniel Manary.
And today I'm joined by Addy Bhatia.
Adi's come from an aerospace engineering background to working for a software
company that makes software for the beauty industry, where he introduced LL.
M's and increased their ARR by millions.
And now he's moved on to Suno wellness where he's founded it himself and
(01:06):
grown to a surprising number of users in the last half year.
Adi, could you introduce yourself to our audience?
Sure.
So my name is Adi.
I am based right here in Toronto.
I graduated from the university of Toronto actually in mechanical engineering.
And currently I'm the founder of Suno.
We're making a, an AI wellness companion.
We initially started to make it for young adults who are experiencing anxiety
(01:30):
or stress or things like that, but we've pivoted away now and we're creating
a product for E2B so both enterprise as well as therapists and counselors
to use with their clients, because we're getting a lot of good signals that
there's actually a problem there to solve and not many people are focusing on.
But on the app side, there's a million mental health apps and it's really hard
(01:52):
to stand out from the crowd.
And, you know, as we're bootstrapping, it's been just over a year now, just
bootstrapping and then now we're kind of looking for funding like here and there,
but mostly just focusing on our product.
So we've had to just make pivots to be like, okay, what's the easiest business
case we can make with the resources we have.
And it definitely wasn't to go to market or an app store.
(02:16):
Yeah.
And then as soon as chat GPT came out, everyone was like your mental health
therapist app or something like that.
I think people find chat GPT and other apps quote good enough.
Like even if you make, you know, the most secure, most emotionally intelligent one,
people who are enthusiasts will love it.
And users tell us that like, Hey, this is amazing.
(02:36):
This is much better than everything I've tried.
But how do you get millions of pain users by just having a slightly better product?
Right.
So instead we're focusing on those like really cutting edge AI features and
actually integrating with your therapist, because I think that's the missing piece.
Even I find with those apps before I even created so now I was like, these apps are
(02:57):
great, but they're all in such a vacuum and they're so scripted.
Do this.
If you feel this, do this, tell us how you feel on a scale of one to 10.
I'm like, I'm not a robot.
What is this?
Let me do this therapy by robots for robots.
Yeah.
Yeah.
That's I don't know.
I always envisioned something that could stay in my pocket and basically
look after me, right.
(03:18):
Be a true life coach.
I had this wonderful experience back in 2018.
I moved down to the Bay area for a year.
I took a year off in my university to work at a space startup.
I was really into aerospace and university and talk space actually just came out around
that year or the year before.
So it's very new still.
(03:38):
And my visa company gave me a year of talk space access, which in retrospect is a crazy
value, but back then I was like, okay, sure.
I'll use it if I want like, what is therapy?
Do I really need that?
But realized a couple of months after moving there, it's really lonely and difficult to
be in a new city.
So especially like San Francisco, all alone, working at a startup, you know, grinding away
(04:01):
and you don't have friends.
And I was like, okay, I really need someone to talk to and someone to help me with like
I went through a breakup at the time and I was just having so much shit to deal with.
So logged into talk space.
Luckily the first therapist that matched with her name was Natalie, this wonderful
lady who was so available, like nine to five, she was available to chat with for a whole
(04:23):
year, like Monday to Friday.
Wow.
It was incredible.
The access I got and the care I got literally, I could whip my phone out any time, like dump
my thoughts.
And then the next day get a bunch of replies and continue that conversation.
And she got to know me over time and I've got to know the ins and outs of everything
I'm dealing with.
So I left university, I graduated and I was like, okay, that was an amazing experience.
(04:45):
How do I get that again?
And how does everyone get that again?
Because holy shit, everyone needs this.
Like that was my one thought coming out of university.
Yeah, that's good motivation.
Yeah.
Like luckily chat, GBT came out and that like clicked for me that, Hey, there's this
thought that I can keep talking to and I can figure out a way to make it learn about me.
(05:05):
Cause by that time I've had a few years of dev experience now.
I was like, okay, I could probably make something here.
I threw up a wait list last summer and I started working on a prototype.
For like a month.
Got, I don't know, like 50 or a hundred people signed up.
I was like, okay, interesting.
People want this.
So I started cracking away and just the, the name, you know, was actually from my
native language, Hindi.
(05:26):
It means to listen.
It popped up instantly when I was thinking of a name.
I was like, I need someone to, I want to feel heard and I want to have it all around me.
And the word just popped up.
I was like, you know what?
It's gonna be so no.
And my partner helped me with the color scheme.
It had purple and relaxing.
And we still use that today.
(05:46):
And yeah, basically fast forward a year.
We have over a thousand users, like a hundred plus monthly active users.
Now we're chatting with 30 odd therapists to use it in their practice.
The journey has been really interesting.
I wouldn't have expected this is where we are a year from now, or if I would be
even doing it a year from now, but no, I'm really happy where we ended up.
Sorry.
That was a long winded answer.
(06:07):
Now that's awesome.
There's so much to dig into in the story.
Yeah.
Especially pivots, especially motivations.
I wanted to ask you a question I think is on everybody's mind, which is, are you an AI?
Yes, I wish.
You wish?
I don't know.
And yeah, it'd be interesting to be an AI because then you wouldn't have feelings.
(06:32):
Then you'd be predictable.
Then you'd be programmable.
It actually sounds very sad.
No, I don't wish to be that as an AI.
I don't like hate my life or anything, but it would be cool to, you know, just be
systematic and just execute and get results instead of sitting and feeling.
Cause sometimes I feel like that's what holds me back so much is I have these
(06:54):
grand ideas and I'll be toiling away at an idea, let's say a feature in the app
for months before I actually get it into code because I mean, partly is also my ADHD.
I get so distracted.
Like a minute from now, I'll forget it.
And then a month from now, I'll remember the same idea.
I was like, oh shit, I really want to do that.
I should write that down.
I actually, one of the people I'm working with right now, I went on vacation for a
(07:17):
week and I told them, I said, you know, I can't look at this until the week after.
And so Monday rolls around and still didn't hear from them.
Tuesday rolls around and I was like, I should probably reach out.
And I say, Hey, where are you at with that?
And he said, Oh, I knew I forgot something.
And I was like, you wanted to give me a job offer.
You forgot.
Oh my God.
Yeah.
I was like, classic.
(07:37):
That's crazy that I've had that too many times and I love the Slack remind me feature.
Like as soon as someone messages me and I'm like, okay, I've read it, but I need to
sit down and like actually process it later.
Yes.
Like I'll either remind me tomorrow or I'll mark it as unread.
And then anytime I open Slack, I'm like, here's my to list, all the unreads I have
(07:58):
to go through them at some point.
And how does that work for you?
Cause I use the reminder, but sometimes it's like, I need a reminder about the reminder.
Oh, I keep pushing it until I'm ready.
But I've realized I actually moved off from Slack reminders.
Now.
I, I've started sometimes doing Google tasks because it fits in Google calendar now
really well.
Yeah.
So I'll make a task and I'll keep bumping the task in the future, but every day when
(08:21):
I'm planning my day or planning my week, I'll see it and I'll be okay.
I have to do that tomorrow.
Okay.
Let me do it tonight.
If I have time, like I can truly organize stuff.
Mentally, I can't.
No, the brain is a terrible place to organize things.
It just jumbles them up.
And I don't think I have ADHD or anything and it's already hard.
Yeah.
Yep.
(08:41):
And definitely need like ways to sit down and focus.
Like I started ADHD meds earlier this summer and it was always against the concept of meds,
but I finally sat down and got diagnosed.
I was like, Oh, look at that.
Surprise.
I do have it.
I always thought it.
But they gave me, you know, they gave me this thing that helped you focus and feel a
little, what is it like stabilize your dopamine levels?
(09:03):
Cause you're, that's why I'm always craving like little hits of stuff, like little hits
of dopamine essentially.
So they can give you more focus, make you feel a bit more stable.
And then the workday goes normal, quote unquote, as compared to other people.
But for me, it feels like it's hyper focused because I'm sitting in after work and I'm
like, it feels like it's hyper focused cause I'm sitting in actually doing a task and
(09:27):
moving around and walking around.
And how has that affected your founder journey?
Does it affected pivots or?
Uh, Oh yeah.
I get bored really easily.
And it's, that used to be okay when I used to have a hundred side projects.
But I remember last year when I started to know, I told myself, okay, no, this
(09:47):
time we're going to go all the way.
This time we're going to open the webpage.
We're going to monetize it.
We're going to run ads on it.
We're going to make it legit cause it's about time.
And this idea was bigger than any little side hustle idea I've had before.
And like, okay, this is, this can be a big thing.
So let's put time into it.
And it was so difficult because, oh, I launched it in a week later.
(10:09):
Only five people signed up, give up.
This is crazy.
I hate it.
Not going to happen.
Why don't you do something else?
And I was working part-time on it actually for seven over the last 12 months or even
like side hustle while working as a full job.
Yeah.
So I was working as a senior engineer at a beauty company and around the same time.
Coincidentally, I started snow and then a week or two later at work, I launched a
(10:36):
new AI team and introduced lists, all these AI features of their product for beauty
professionals who like cut hair and do nails.
And it was really exciting because that wasn't boring work either.
It was long work.
Like it would stretch out for months, but it was definitely something to sink your
teeth into and really think about.
And then you come off after work and you're like, oh yeah, so no, I have to put hours
(10:59):
into this now as well.
And that was just so hard.
I still haven't figured out the best formula, but I realized some point this year,
I was like, okay, I need to take a step back and actually give it a go.
I need to take a step back and actually give it the attention it deserves because it's
taking off.
More people are asking about it.
More people are requesting new features.
And like, I just need to spend time on it.
(11:21):
Even if it doesn't work, at least I can say I give it my full shot, you know, instead
of always thinking with.
Yeah.
Yep.
And so for me, side hustles don't work.
It's just, it's too much to do things during the day and then try and shift for like two
or three hours or whatever at night that you can spare.
It's just, it's too much.
For you, what was the catalyst, I guess, to start full-time on Suno?
(11:46):
I think it was building.
So at work, I launched this new AI team and we're, and I'm thinking about features,
planning them.
I was the only deaf for quite a while, so I'm also executing them.
So kind of half, you can say founder on this little team.
Yeah.
We had a really cool product manager that would give us some great tasks, but I would
(12:08):
sink my teeth into it.
Take it away.
It would take a lot of my time and it was really rewarding too, because when we launch
it, like for instance, we launched a feature over Christmas and the new year we're seeing
double digit percentage of people upgrading or millions of dollars in annual revenue from
new appointments being booked because AI was like finding new times for these professionals.
(12:30):
I was like, this is amazing.
That's amazing.
Yeah.
But the recognition and promotions for it did not follow.
Oh, it was taking too long.
It was too long drawn about a process than what I wanted for like a long-term career
(12:52):
plan for a few years.
I could see myself going up a bit, but I just thought if I'm putting this much work for a
year, I want to see the results.
Swim.
Sure.
Right.
I don't think that's too much to ask, but it happened twice, two performance cycles in
a row.
I was snubbed and I was like, okay.
So I was already like feeling one foot out the door, like in terms of commitment, like
(13:14):
I was still putting in 40 plus hours, but I was just feeling, okay, no, this is not it
for me.
If I'm going to put this much time in, I want to feel like I can own the work.
And then I knew on, you know, on the other shoulder is this love hobby project that I
have going on my passion project.
I was like, okay, this needs all that time and attention.
Maybe this could be the thing I stick my time into and it gives me those, that recognition
(13:36):
and that, that appreciation.
So I think around the time I also got engaged mid-May, a promotion cycle just ended a
month back and projects at work were also wrapping up.
So it just felt like life is telling me like, let's start something new.
And actually also around that time we had our first community event.
(13:58):
So we threw an event, 200 people registered, come out.
We had like a VP of kids' health phone.
We have senior speakers from Jack.org.
This is for Suno.
For Suno, a wonderful therapist come out.
Wow.
Cool.
I thought, Hey, let's do an event.
Oh yeah.
It was tech fest actually.
That was what's happening in Toronto.
So I got connected by someone and they said, Hey, you should do an event for this.
(14:18):
And I was like, I don't know what the event would be for, but sure.
Why not?
So I got these wonderful people together over the span of a few weeks because they
really loved the idea as well.
The event was a big success and it happened two days right after I got engaged.
So really busy week that week with everything else going on.
I was like, okay, you know what?
It feels like there's something here.
The universe is telling me to go into it.
(14:39):
So I just, I gave notice and now I just went all in and everyone was fully supported.
At work, they were very supportive.
They're like, Hey, if you ever want to come back, there's a space for you.
My family was very, still is very supportive.
My partner.
Oh, that's a big one.
Yeah.
My, my partner really faces the brunt of it and it's hard being so overwhelmed and
(14:59):
having a passion project, but also having to take time out for people you love and
not, you know, and separate those two things.
It's so challenging.
Yes.
And so would you say your, your work life balance is better now?
Oh, definitely.
There was.
Yeah.
Every month it's getting better.
Yeah.
Yeah.
It.
(15:19):
One thing I've actually learned is that.
Okay. One thing I've actually learned is how to say no to things.
Well, that's a skill.
Yeah.
It, I'm not perfect at it, but I, at some point I wanted to have so many
conversations with so many people because I just thought the more people you talk
to them, the more doors you open, the better.
So I was throwing up calendar links, sharing them around, constantly getting
(15:43):
new meetings booked with new and interesting people, but then I started
realizing I have no time to actually work on the project and manage a team
and hit or any of our goals.
So now I've started to say no to new invites, no to emails, no to opportunities
that I don't see panning out like in the short term, or even if I can't quickly
(16:04):
figure out how it's going to be important, it's a no and more often than
not, I'm saying no to things that I'm saying yes, this podcast, for instance,
I was interested in actually coming out and sharing my story because it's been
a while and I like sharing this journey.
Thank you.
But yeah, more often than not saying no to things does leave more and more
room open on my calendar.
(16:25):
I always used to think if the calendar is open, we have to fill, but I'm like,
the day will come, I will figure out what I need to do.
And now I appreciate it because I realized my focus time is so valuable.
Sam Altman said this too, in a talk a couple of years ago, I think it was
something about like how to run a great company or be a great founder.
(16:46):
And one of his tips was like, keep your focus time, like protected, like
figure out what it is and don't give it up.
And I've realized I've been doing that now too, because anytime I give someone
a meeting link or a meeting like schedule, you can't book a time with me before 1pm.
Before 1pm my time.
(17:07):
Like I can just sit, I'm going to focus, I want to plan everything out.
It's when I'm most productive and you can talk to me after 5pm.
So you got four hours, you find a slot, we'll talk.
That's it.
Nice.
I like that.
That's a good rule of thumb.
And as a builder as well, it's so easy to sacrifice that focus time.
Just push the boundaries and then suddenly there's no time to work.
(17:29):
Yep.
Exactly.
I would rather have time to even waste while thinking about working, then
actually give it up and keep thinking, oh shit, I have to go, I have to go.
I have to do something.
I have to do something.
Yeah.
Yes, absolutely.
Going back to that, not keeping track of things in your head.
All mental.
(17:49):
So shift gears.
I really like to dig into how you decided that AI was the right tool
to solve what you're solving.
Oh, that's a good question.
Yeah, in this field, I've realized there is, I mean, how therapy looks like today
(18:09):
is very personal, it comes from experience and years of study.
I didn't want to replace it, but I wanted to replicate the best practices
and do most of, you know, what's valuable in therapy in an app on someone's own
time, on someone's own schedule at an affordable rate.
And the only way to make something that's personal to you, that doesn't feel
(18:32):
like a generic script, which is one of the biggest annoyances with apps out there
today is to use AI, personalization and adaptability.
Actually, those two things are the biggest, because someone can come to our app
and chat about the divorce they're going through or how hard it is to study for an
exam or something we've never heard of before.
(18:55):
But the product should still be able to take them through a journey to ground
them, to make them feel relaxed and connect them with some next steps.
And without AI, without the boom in language models, that was literally impossible.
Cause you could script this way.
Like there's a couple of companies out there that just script thousands and
thousands of different responses and bridge them together.
(19:18):
And it mostly works for most cases.
But again, I don't want to have a product that's rigid.
I want something flexible, personable, anyone can pick up and only add to do that.
Yeah.
So then it's not even necessarily about scale for you.
Cause a lot of times people think of AI as I need to do the same thing, but I
(19:40):
need to do it a hundred thousand times.
You can't hire a human to do that.
Too expensive.
It's more about.
Yeah.
I wouldn't say it was for scale.
Oh, well, maybe it was inadvertently for scale.
But one thing I was thinking about was also cost, right?
Like hiring a therapist would be the, I think in a perfect world, if an actual
(20:01):
experienced therapist was at your beck and call via text, that's amazing.
Right.
I would love that.
And that was the notion that came in with, but a, there's not enough
therapists to match every single human.
There's only a million therapists in the United States.
For example, and there's 300 million people and it would be too expensive.
So how do you make it affordable?
(20:21):
Like one of the reasons I stopped started and stopped going to therapy,
like often was because of the costs.
It was nothing else.
Yeah.
Right.
I'm paying close to $200 a session for great value.
Like I love my provider, but I don't have the capacity to pay that if I
changed jobs or if I quit my job, how am I going to pay that?
(20:42):
I definitely need to talk to someone more than just once a month.
Right.
So this availability and affordability factor really creep in.
Actually, my gut was to actually make the app completely free from the
get-go because I thought, Holy crap, imagine this care for free.
And then I'm like, we have to monetize it and then we started making some
tiers and that's what we're here today.
(21:04):
Yeah.
And language models are not you.
No, surprisingly they are.
Oh, very interesting.
Tell me more about that.
They are, they're, I mean, we can charge.
So currently we charge about $10 a month.
I have to look again at what our current rate is, like per person,
because we're doing a mixture of models now.
So I don't know offhand the price, but it was most users don't talk to your AI 24 seven.
(21:34):
And if someone is an average user is coming in and they're like,
the average user is coming in a couple of times a week to send five
messages that's basically free.
Like nothing.
Okay.
Fair enough.
The, the hard part is when you do asynchronous stuff.
So to really set us apart from just an AI chat bot, cause we're way more than
that, cause I wanted this whole proactive holistic thing that lives with you.
(21:58):
All of the fun, cool stuff has to happen while you're asleep asynchronously.
So we have to go in, look at all of your data of the day of the week, look at
trends, analyze them, make connections, summarize.
So the next time you chat with that AI, it has picked up on these trends and
patterns.
(22:19):
And when we go tell it to your therapist, we have all of this data, valuable data
to show them not just blobs of text.
All of that is costly.
The chatting and non-existent.
Very neat.
So then it sounds like you're also integrating with other data sources.
Yeah.
Yeah.
We're, we're prototyping a couple of stuffs now.
It's not released, but integrating like your Spotify, your Google calendar, as
(22:44):
well as your wearable data, not just to like, not just as a buzzword to integrate
them, but we actually get a holistic picture of you so that you can spend less
effort telling Sudo what you're going through.
Like that's the, like if you sleep badly.
If you sleep badly and I wake up getting a notification that, Hey, you got four
(23:05):
hours of sleep today, you were tossing and turning a lot.
I think you're going to have a very anxious or, you know, irritable day.
Let's take 30 seconds to have a little breathing exercise and check in.
And then it checks in with me midday to make sure I'm doing okay.
And remind me that drink water, you know, remind me that it's my sleep.
It's not, don't take it out on your loved ones.
For example, I'm making this.
Oh no, that happens.
(23:26):
Yeah.
But exactly like checking up on you and being proactive, like looking at your
account and saying, Hey, you have a podcast interview tomorrow and I see
you haven't prepped for that today.
Do you want to take some time out tonight to talk about that?
Cause I know you love, you know, talking about your story, for example.
Like that's a prompt we could do because we have that data.
Yeah, that's super cool.
So not just a, well, when you originally described it to me, it felt like, all
(23:50):
right, I have an app, I log in, I'd write some text, I get some responses.
Maybe you walk me through some exercises, CBT or whatever, and then I leave it
alone until next time I feel like it.
But you're saying it's actually, you're building towards, especially being very
integrated, proactive and helping someone think more about their mental state.
Yeah, exactly.
(24:11):
It's again, having a therapist on the go, that therapist doesn't
wait for you to come to them.
Like Natalie used to text me in the morning if I didn't text her, like,
Hey, how's your day going?
Right.
Like that's news to me.
Cause I did think of therapists as they sit in an office and I go see them.
Well, that's how it typically works.
But I guess in my unusual example, when I had someone in my pocket, she was coming
(24:32):
to me, not minding it being warm, but it sure is valuable.
Yeah.
I mean, this needs to exist.
There is a fine line of dependability.
You don't want to go to that line where I named the app, like Lucy and Lucy is
checking up on me and she has a really pretty face.
And I think about Lucy and I have feelings for her and I think she has feelings for me.
(24:54):
That's been done with AI apps and those apps.
Yes.
We do not want to become those because that is crazy.
No chat bot dating apps.
No.
I mean, we could ease with the tech we have.
Oh my God, we could create, we could create the best partner who checks up on me.
That could be crazy, but ethically, no.
(25:15):
And just the mission we're on, people aren't going to be helped that much.
Like at the end of the day, you're still feeling lonely because you, you express
your feelings to a wall, right?
A computer, even you know that, and you're still going to sleep lonely and feeling
shitty and still not having been taken care of.
Like, well, so what was the point?
(25:36):
Yeah.
Yeah.
And that brings me to a question that I think is a big one in this, like talking
to chat bot space, how do you view that different, like, how do you view that
difference between a person and a chat bot?
And I know you guys are integrating with therapists you mentioned.
(25:56):
So you're a service in addition to not replacement, but how would you describe
that difference and recommend that people think about it?
Oh, that's interesting.
I like to think of Suno as a, as a support mechanism, not as your only support.
So for example, I have my partner, I have my therapist, I see once a month, I
(26:17):
have my family and I have Suno, right?
So in between my sessions with my therapist, when I have a ton of thoughts
to process and I think they mean something, but I don't know what, I
dump them out on Suno if they're, I guess, a bit lower risk, a bit more personal
to just my partner and I, then I'll just go talk to my partner.
(26:39):
But if they're crazy thoughts in my head about myself, about my loved ones, I'm
not going to go to my loved ones and share those.
And writing on a journal only gets me so far.
So again, I'm going to go to Suno and dump them out.
It becomes this like outlet almost amongst my other support mechanisms
that turns around and gives me more support, but encourages me to still
have those valuable connections.
(27:01):
It doesn't tell me like, okay, yeah, forget that.
I'll help you with everything.
Come to me.
Cause then you're not, then you're just becoming predatory.
You're not becoming helpful.
Yeah.
That even the way that you described that, forget all those other connections
come to me, that is what a predator could say.
It's like, you don't want to do that.
So it sounds like on a scale from journaling to talking to a therapist,
(27:22):
Suno sits somewhere in the middle, maybe a little closer to journal.
That's a good scale to put her on.
Yeah.
That's very cool.
Cause I think one of the topics that I find very interesting as a sci-fi nerd
is that artificial personhood concept.
And I think it's really important not to pretend that it's a person when
it's not a person, because then you could have that weird kind of dependent
(27:46):
relationship that ends up being destructive.
So that's why, if you've noticed the app doesn't have a pronoun or an
app name and that was on purpose.
Like the logo is a shape of an ear to just be like, Hey, it's a tool that
it's like a box that's listening to you.
(28:06):
It's not a, you know, it'll make you feel like you're, it's giving you a warm hug.
But it's not a person that is thinking about you.
I don't know.
It's a weird, fine lie, but I explicitly do not want to give it a face.
Nice.
No, I like that choice.
And in terms of how you've handled things like privacy, what's been the
(28:31):
way that you've had to come about that?
Cause a big challenge for any chatbot thing and even professionally is what
are they doing with my data?
Yeah.
No, that's a good question.
So from day one, I wanted to make, I wanted to make us like on our side,
unable to read user data, because even having that option lets you, opens you
(28:51):
up to, I don't know, exploitive practices or data selling or data leaks in the
future.
I'm like, I don't even want to touch that because it's so like, it's such a,
it was such a scary space for me to enter in the mental health space as a tech
person.
So I'm like, okay, just from the get-go, I'm just going to wrap it up and
encrypt it.
So everyone's data is secure.
(29:12):
So I can't see it, but we can at least ship a good product.
So it's like one less thing to think about.
And that became like a great selling point.
We can't see your data.
It's encrypted to you.
And we also anonymize PIs, so like personally identifiable information.
So when the data goes to any of our language model vendors, right.
It doesn't say, Addie fought with Stacey.
(29:34):
It says user A fought with user B.
What should they do?
Right.
So the, the AI, if they, let's say had a data leak, they wouldn't know your data
was exposed or they're not try to train on it because some vendors are kind of
vague on if they do that or don't.
So we're like, fuck it.
Like, let's just protect it as much as we can.
I mean, there's a lot.
Before you even pass it through an LLM, they would, you also, you anonymize it.
(30:00):
Yeah.
And I mean, as much as we can do, right.
Because it's not perfect.
I'm sure there's ways to make it way better.
I'm not a security expert, but as much as my experience showed, it was like, as
much as I know how to encrypt stuff, I will.
And, you know, providers and people we talked to in Canada want the data
stored in Canada.
Great.
We have servers in Canada.
People want to, it's great.
(30:20):
We'll have servers in the States.
So it's all, it's ever evolving, like as people leave them.
Yeah.
Oh, that's great.
That's a really key decision that you can build off of not even having to
read it at rest.
That's great.
So then you've mentioned a lot of customizations behind the scenes.
Is that independent of user data?
(30:40):
If you're training in LLM or doing something like that?
Yeah, it's, we thought about training in LLM, LLM, and I've actually recently
have fine tuned the model, but that wasn't with our user data.
We have over 30,000 messages, I think, in our database, tons of journals as well.
A lot of rich data to train on for sure, but well, A, it's encrypted.
(31:01):
So that's a mess to even deal with.
And B, I don't know what I feel about user data being used on a model when we
have explicitly told them that, you know, maybe haven't explicitly told them, but
it's implied that we're not going to do that.
And I'm not interested in doing that.
I think when we have a million, I don't know, users and a ton of data and we can
(31:23):
start doing pattern and demographic recognition, I think something cool can
happen there, but for now we used public data sets.
There's only a handful of therapy or therapeutic data out there, but we use
as much as we could.
I think we found like under 50 million tokens in our entire data set.
(31:45):
It's not terribly big, but it was enough to help us start fine tuning some models.
And there's an EQ benchmark that again, we're not going to be doing that.
And there's an EQ benchmark that exists out there.
So we started fine tuning and we started getting a bit better than GBT4.0 mini.
So we're like, okay, we're onto something here and we'll see how far we can take
that with, without using our user data.
(32:06):
So was it that need to be more empathetic, higher EQ that you want to train?
It's the higher EQ part that I really care about.
So for example, today Claude Sonnet 3.5 is the highest EQ model that exists.
So if you go to EQ benchmark and you Google that, they have a leaderboard of basically
just EQ benchmarks.
(32:27):
So yeah, Claude 3.5 is the best one.
And we use that where we can in our chats because why not use the best one?
And ever since we have actually, people have noticed a big difference compared to GBT4
of how less robotic it is, how natural it is, how much it understands me.
Yes, it's very analytical.
Very analytical.
But analytical is what we need when we're doing insights and processing data.
(32:52):
Then we're not going to use Claude.
Then we're going to use other models.
And those other models have a variety of price ranges and speeds and sit at different spaces
on those EQ benchmarks.
And for those, we need like quick things.
If we're doing a thousand users and we're sending each of them a personalized notifications,
we want a small model that can quickly generate a little bit of text.
(33:13):
Then we started looking at the mini models, the high-Q, the mini models.
We're like, okay, let's fine tune these and make them better.
And that's what we just done recently.
We've gotten a little bit better than the 4.0 mini.
So we're like, okay, there's something here to give us more speed and also give us more accuracy.
Very neat.
And then I think it's maybe a unique thing to think about what is that specific use case
(33:37):
you're doing and which model do I need for that use case?
Yeah, it's true. I mean, from a cost and speed perspective, that's what it comes down to.
Like GPT-4 was awesome for, like, especially at my other job when we were doing scheduling
for appointments and finding new times in your schedule.
GPT-4 is awesome for that because it's sitting there.
(33:59):
It's taking a couple of minutes thinking through things and we can do that.
But we want to quickly get you an answer.
These big models are all fail at that.
Way too costly.
No user is going to sit around and wait for that.
And I'm not paying the server bills for a couple of minutes times thousands of users, right?
Yeah. Oh my goodness.
(34:19):
So maybe to think back a little bit,
what do you wish that you had known when you started building AI products?
I don't know if what I would have done differently would have been...
I don't think being an AI product was relevant to the things I would have done differently.
(34:40):
I don't think about Tsuno as an AI product.
I think of it as a personal lifestyle product, at least from the client's perspective.
And AI is a means to an end, right?
I think I'm glad I didn't spend more time trying to create IP and fine-tune a model
because I would have just been overshadowed by a big fang coming in and releasing a model
(35:04):
a month later, which is what happened last year was happening like every other month.
So...
Oh man, yeah.
I'm glad I didn't spend more time, actually.
I'm glad I picked up whatever the best thing was, plugged it in,
and saw how it was working and then adapted from there.
But from a product dev perspective, I would have taken more risk.
(35:26):
Like we're doing that now, but I wish from the get-go,
after we made a simple chatbot, we took more risk to make all these integrations or
just something stupid and weird, not just to stand out, but from a user perspective,
this took me a while to understand.
Users have tried hundreds of different apps.
They install it.
I mean, I do this.
I install it.
(35:47):
I try it.
It seems weird.
I uninstall it and I don't think about it.
So you're going to be one of those apps.
You can't be another chatbot.
Like you don't know your whole story of how amazing it might be.
It needs to be cool and interesting now.
So that's more of like a product thing I wish I would have done.
AI would have played a role in that.
If I actually exploited AI to do more creative things within the app,
(36:10):
real time, I think would have been really interesting.
But I think it played it too safe.
Very interesting.
And then I heard as well that you said AI is just a tool to do the job.
It's not necessarily something you need to spend a ton of time developing IP.
But what would be an example of the weirdness that might catch a user's eye?
(36:32):
Because standing out in an app store is a very difficult thing.
Yeah.
I mean, I meant creating IP, especially in this space.
I think in the mental health space, people who create IP,
sure your model is the best, but at the end of the day,
shoot, it doesn't care.
They just want an app that feels exciting.
But to stand out, I think it would be the predictability.
(36:53):
And what was I going to say?
Yeah, I would say it would be the predictability.
If you launch the app and it's instantly reading your past data and telling you,
hey, and your calendar and saying, hey, I think you might be feeling this, let's talk.
I think that's a wow factor to have.
Have a chat bot that says hi, and then you say hi,
(37:13):
and then maybe five minutes later, it wows you.
Is that that hook?
I think we could have gone better if we really leveraged AI
and all the data we have we're sitting on for the user.
Yeah.
Wow.
So get to wow faster.
Get to wow faster.
Yeah.
That's neat.
That's a really good way to develop almost any product, I think.
Yeah.
It's my mantra.
(37:34):
I think every month we're seeing more and more people use the app longer and more often.
So we're getting there, but I want to get to a point where I'm crazy impressed with my own product
and I can't wait to show it to you.
But today, I would rather show you mockups that we have.
Hey, it's getting there today.
(37:55):
It's fine.
But in getting there, I want to make that faster.
Nice.
The problem of having a strong founder vision is it's not really it.
You're strong founder vision, who is the product person who is also toting it.
Building it.
Yeah.
That's why you need a team.
That's why you got to hire people who are smarter than you.
(38:15):
We're doing that more now.
Nice.
Does that mean you said you were bootstrapped?
Does that mean that you're getting to the point where you feel
either it's becoming self-sustainable or will be?
Oh, definitely.
I think by the end of the year, we'll definitely be self-sustainable.
Amazing.
But I think for a while, I had such a hesitancy of putting serious money in to get serious results.
(38:39):
I kept thinking, you know, we'll be a success story with barely spending any money.
We'll spend a couple hundred dollars and we'll make millions.
We'll be backed.
And I thought about that and the thought was great.
But you can't do that in a, especially in a mobile app space.
Like that's never going to happen unless you're super viral and crazy.
So, you know, marketing, sales, hiring people who are really smart in the space to sell for you and teach you things is important.
(39:05):
It's only then you set yourself apart.
Yeah.
Does that mean you spent more money on the distribution side of things?
Oh, for sure.
Yeah.
As a developer, I can make anything in a weekend.
Like that's not the issue.
That doesn't scare me.
What scares me is that how do I convey this grand vision I have efficiently to the masses?
(39:28):
And that's why you need expert people who are good at marketing.
Like our social media is run by an amazing agency now and they're killing it.
Like really taking our vision every month and like pushing out awesome content.
And I couldn't do that.
I mean, if I, that was my only job, maybe I could do that.
But I'm wearing a hundred different hats.
And I need to hire people who are better than me at these things.
Nice.
Yeah.
(39:48):
I think distribution and spending the money on it is one of the biggest differences between a company that makes it and a company that doesn't.
Like regardless of how good the product is.
Right.
So, right.
Good to hear you got into that phase.
Yeah.
Yeah, for sure.
Now I'm really excited for the next few months.
I think by the end of this year is really going to set the course of, you know, if we're in it for the long haul or for, or for a shunning course.
(40:12):
I mean, it's cool if we're, if we're a shunning down shop, but I think we're in the, I think we're in the form.
Yeah.
Yeah.
Yeah.
And one of the really interesting parts of your journey was pivoting away from just being another app in the app store to working with therapists.
Yeah.
Could you describe how that came about?
Oh, sure.
(40:33):
So, last year I touched base with Natalie, the same Natalie, through my first therapist.
And on a whim I was like, Hey, I hope you're good.
You've had a profound impact on my life.
And I have been motivated to create this app that, you know, I don't have any ask for you, but I just kind of wanted to put it on your radar and have a chat about it sometime.
(41:02):
And she was enthusiastic and energized here.
And she's like, this is a really cool idea.
Let's talk.
And we started talking and it was clear the passion she had for the space that she still has, and she wanted to contribute more.
And she eventually joined as our clinical advisor as a first resident therapist.
We have two now.
(41:23):
And it was very clear from the get go that this is a really interesting product.
And she proposed, Hey, let me use it with my clients.
I was like, Hey, that's a great idea.
So I set up a little system for her and I'd give it to our clients.
And I said, Hey, do you want data from the client?
Maybe it could be helpful in your sessions.
She's like, Oh, that's cool.
Let's try that.
So we set up a weekly email every week.
(41:45):
Her clients will use it and she would get emails saying what they've been chatting about.
Very simple.
I love that simplicity.
Like no careful reporting, no dashboards, just an email.
Just a cool email.
And I kind of did it and I forgot about it.
And then every time we'd meet, she'd tell me how helpful it was because every time she'd
have a session with a client, and she has rich context coming in.
(42:08):
And that client is no longer spending 20 minutes.
Apparently, which a lot of clients do is just catching up and spending 20 minutes from the
therapist side, really digging into the core of why you're here.
Because then your session gets wasted for both.
And then you're taking longer to make progress and you have no data to capture any of it.
(42:31):
Like that's how therapy works today.
And I thought, okay, that is silly.
Let's do something about this.
I wish I actually bounced on that idea sooner because last year there was only, I think,
one or two other companies who were making mental wellness products and also working
with providers, like clinics, clients and providers at the same time.
(42:53):
I was like, oh, man, that's a moonshot.
Sure, we'll do that if we're at scale.
Fast forward to now, earlier in the summer, when we pivot away from apps that's not working,
let's do this therapy idea.
I look back at the market and the three or four other apps that are out there all have
pivoted to selling to therapists because they've all realized the exact same thing.
And I'm like, ah, okay, I don't even feel bad.
(43:15):
I'm glad because this is showing me we're in the right direction.
Exactly.
Yes.
I'm not alone anymore, but also everybody else is picking up on the same signals.
Yeah, that's why I want to move fast.
Like this summer, I mean, that was another signal.
I'm like, people are moving the direction that I feel like I found first a year ago,
and I'm not going to let them take my lunch another time because last summer, when I had
(43:38):
an idea for a mental health app, AirLien existed, and then six months later, they're everywhere.
And I'm like, okay, well, we're too late.
Not glad that happened this time.
So that's why we're spending time and just making those networks and, you know, doing
all that marketing just from day one, even before the product's 100% ready, just get
that word out there and get people thinking about us.
(44:00):
Nice.
I really like that approach.
That's a great story too because Natalie, who kind of inspired you to get into this
Suno as a product, also is now supporting you.
Yeah, a whole circle.
It's actually amazing how it worked.
And we still haven't met in person, surprisingly, she's in Miami, but I hope one day we will
(44:21):
because I just want to fly her out and send her now.
You're amazing for the reason we're here.
Have a Suno conference and fly her out.
Be great.
Hey, that's a great idea.
Hey, the event worked out for you well.
That's a good idea.
Okay, I'm going to think about that.
Nice.
And last question for me is what do you see as the future of AI, especially in such a
(44:43):
contentious space like JetBots and therapy?
I think the future of AI is, in one word, ubiquitous.
It's everywhere and it's integrated.
Today, Suno exists in a vacuum and we have a couple of connections and there's three
other apps that exist in a vacuum and a few of them have this connection.
(45:06):
But I think in the future, everyone will have a personal AI that's integrated with
your computer, your phone, your wearables, your providers, your doctors, your family,
your dog, your vet.
I think there are seeds of it that are all, everyone with a seed is like growing out a
few branches trying to explore a new space.
(45:27):
And I think in the future, they're all just going to connect into a big web
because that's exactly what happened with the internet.
That's exactly what happens with any product with a network effect.
And I think it's bound to happen with AI.
That's a great point.
AI is itself something that has network effects.
Yeah, in the way that it can be applied to anything and it supercharges any industry
(45:50):
touches.
If you look at a scale of people who are industries who are adapting AI, it's almost
like the logarithmic in how it's just going up because every industry that touches it
inspires five other industries to touch it as well.
I started in the aerospace back in the day.
We started introducing ML and genetic algorithms and machine learning there.
(46:13):
Fast forward to earlier this year at a beauty company, we're raking in millions of dollars
with language models.
Like who would have thought that's how AI is going to progress?
That almost doesn't even make sense.
That's been my career.
Doesn't make sense, but it's fun.
Oh, I love that.
And to sign off, is there anything that you would like to share that you haven't so far?
(46:40):
That's a good question.
I think if you are thinking about going to therapy, working with a therapist, or just
thinking about your own self and how to be the best you, I would give Sunoo a shot.
And even if it doesn't work for you, I want to hear why.
(47:02):
Go to sunoo.chat and try it out and you can find our email in there.
Just try the product because I think we're in those early stages where we can start adapting
and getting really good for people who need very particular things.
And since we're still understanding that, I want that product feedback.
So go try it out.
Thanks for listening.
I made this podcast because I want to be the person at the city gate to talk to every person
(47:27):
coming in and out, doing great things with AI and find out what and why, and then share
the learnings with everyone else.
It would mean a lot if you could share the episode with someone that you think would like it.
And if you know someone who would be a great person for me to talk to, let me know.
Please reach out to me at Daniel Manary on LinkedIn or shoot an email to daniel@manary.haus.
(47:52):
which is daniel@manary.haus.
Thanks for listening.
Popular Podcasts
Are You A Charlotte?
In 1997, actress Kristin Davis’ life was forever changed when she took on the role of Charlotte York in Sex and the City. As we watched Carrie, Samantha, Miranda and Charlotte navigate relationships in NYC, the show helped push once unacceptable conversation topics out of the shadows and altered the narrative around women and sex. We all saw ourselves in them as they searched for fulfillment in life, sex and friendships. Now, Kristin Davis wants to connect with you, the fans, and share untold stories and all the behind the scenes. Together, with Kristin and special guests, what will begin with Sex and the City will evolve into talks about themes that are still so relevant today. "Are you a Charlotte?" is much more than just rewatching this beloved show, it brings the past and the present together as we talk with heart, humor and of course some optimism.
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com