How Wispr Flow Uses AI to Save Professionals Hours Every Day

The average person today takes 5 hours every single day. And if you can half that, you give them back 20 to 30 full days of their life back, which is

unreal. And so what Whisper does is it

unreal. And so what Whisper does is it hears you when you speak naturally, your rambles and all, and then it writes just how you would have written. And the most

magical thing about it that we measure is how long does it take for people after they see what Flo wrote to press enter? Half a second. Nobody even reads

enter? Half a second. Nobody even reads what Flo writes anymore after they've used it for a few days cuz they just trust FL to just get them. Welcome to

Subversive, a podcast dedicated to sharing stories from the best consumer subscription apps in the world. We'll

bring you lessons for how to grow your consumer subscription business, including insights and inflection points that led to exponential growth from leaders at category defining companies

and innovative startups. Let's get into the show.

My guest today is Tene Cathari, co-founder and CEO of Whisper, the company behind an AI powered voice dictation app called Flow. In this

episode, I talked to TA about how Flow's zeroedit design philosophy differentiates it from other voice dictation tools, about the technical complexities that Whisper has needed to overcome to make Flow work across

multiple platforms and more than 25,000 different applications, and about how the company's clever approach to influencer marketing has led to lots of free organic growth. All right, welcome

today to the Subversive podcast. I'm

really excited to have you on today to talk about Whisper and your first product, Flow. Uh this is an AI powered

product, Flow. Uh this is an AI powered dictation app that I've personally been using a lot uh at my company and I'm hearing more and more people talk about it recently. So would love to hear a

it recently. So would love to hear a little bit more about you and the origin story behind this company and then we'll get into some of the early moves you've made to drive growth for the business.

Of course. And by the way, thanks for having me here. This is I was I've been looking forward to this. So, we started working on Whisper Flow about a year ago. And for me, this has actually been

ago. And for me, this has actually been a personal project I've been working on for the last 16 years, which is what is the next generation of personal computing feel like? Cuz back in 2008, when the first Iron Man movie came out,

I wanted to build Jarvis. And so, me and a friend of mine built what was then one of the world's first voice assistants.

This is before Siri, before Alexa. And

to people, it just felt like magic. It

grew to 2 and a half million users.

Google then shut us down because they didn't like what we were doing. And over

the years, I built about 60 70 different products in AI personalization, voice systems. And I think today the

technology is finally ready to actually build something that feels like Jarvis, something that makes interacting with devices just as natural and effortless

as talking to a close friend. And that

is what we want to build with Whisper Flow.

I love it. Well, and one of the things that is most inspiring about that story, one is that you were so young when you had this inspiration, but two is that you actually stuck with it. I I feel

like there are a lot of 10, 12y olds running around being like, you know, I want to be an astronaut or I want to be an archaeologist. But you you saw the

an archaeologist. But you you saw the Iron Man movie and you very specifically said, I want to go build Jarvis. And

then you've built not one, not two, but 60 or 70 different iterations of a product like it. So that brings us to Whisper today. Uh you said you've been

Whisper today. Uh you said you've been working on it for around a year. Um tell

me a little bit about the team, your co-founder and sort of the origin story around getting this product off the ground in the first place.

Yeah, of course. So I started this company with Sahed, who is my co-founder and CTO. He and I actually go way back.

and CTO. He and I actually go way back.

We met first day of undergrad back at Stanford and then we ended up grooming together for the next three years.

So I probably know this man better than most people really should.

And uh with him, the thing he and I really aligned on was life is short, right? You live, you you're born, you

right? You live, you you're born, you live for a little bit, then you die. And

so you want to spend your time doing the most impactful and meaningful work.

And so for us it wasn't when we were talking the thing that we were excited about was going for the big swing and actually

looking at not building a specific app but how people interface with technology cuz the thing that we cared about is making it more seamless. Like we are

extremely annoyed by the fact that hey you walk around and you just see people stuck on their phones looking at screens all day long like this. And I

wanted to step away from that. And so

that was the starting point for us. And

flow was actually the eighth product inside whisper that we built. The first

seven had to do with taking actions. You

could they were the the standard AI agent companies that you would see. And

the challenge with those was the never build habit.

And we then spent time just sitting beside people looking at them do work on their computer. And the non-obvious

their computer. And the non-obvious obvious insight hit us that what people spend a majority of their time doing on their computers and phones is typing.

You type messages to friends, you type prompts to AI, you type emails to your customers and whoever else. And that is something that if you tap into, you

start to have a massive impact on people's lives. The average person today

people's lives. The average person today types 5 hours every single day. And if

you can half that, you give them back 20 to 30 full days of their life back, which is unreal actually.

If you think about that and so that was the the inspiration to build flow and then pull the initial version. It

started getting a lot of traction within the company. we all started using it for

the company. we all started using it for everything that we were doing and and that to us was the point where we decided to just focus the entire company onto that and uh make it something that

just an average person on the street could start using.

So I love that you called it a nonobvious obvious insight. It's one of these things that we all do so often and we spend so much of our times doing it.

just typing emails, typing out memos, reports, presentations that you sort of take it for granted at some point.

Um, and so sometimes these big insights can just be hiding in plain sight. Um,

so you've built this AI powered voice dictation tool, but it's certainly not the first voice dictation tool that's existed. There are many other versions

existed. There are many other versions of a product like this that have existed for years. So what is it in particular

for years. So what is it in particular that Whisper Flow does today that is already differentiated from the other dictation tools out there and then where do you see it evolving in the coming

years?

It's it's really simple how we differentiate ourselves in this specific domain because every single other dictation platform, right? There's the Google, Apple,

right? There's the Google, Apple, OpenAI, everybody else in the world, they try to put down every single word that you're saying word for word. They

pride themselves on that. All the

metrics, everything else is on it, which is fantastic actually for video transcriptions. It's fantastic

transcriptions. It's fantastic if uh you want to record something, but if you're talking about putting our thoughts down as people, as messages or

whatever, you speak differently than you write. And that's the key insight we

write. And that's the key insight we went after. And so what whisper does is

went after. And so what whisper does is it hears you when you speak naturally, your rambles and all, and then it writes just how you would have written.

H what that means is if you say, "Hey, let's meet at 5. actually know let's do six. It's going to write, "Hey, let's

six. It's going to write, "Hey, let's meet at 6 because that's what you wanted it to write." So, imagine how a smart person sitting on your computer would have written what you're telling them to

write. That's how I have the mental

write. That's how I have the mental model of what flow should be doing for you. And there are so many small nuances

you. And there are so many small nuances that that come. Uh when you're saying a long rambly list of things, it puts it into bullets. When you're writing an

into bullets. When you're writing an email, it formats it like an email. when

you're sending a text message, it's lowercase, it's got emojis, it's got all of that. And so, the biggest thing that

of that. And so, the biggest thing that we care about is how do you output something that people just look at?

They're like, "This is perfect." And

they just press send. And the most magical thing about it that we measure is how long does it take for people after they see what Flo wrote to press enter?

Yeah.

A second. Nobody even reads what Flow writes anymore after they've used it for a few days because they just trust Flow to just get them.

Wow. I love that. I I mean, it's so compelling when a company has built something so novel and interesting that they identify a metric that is very

bespoke to their business that represents sort of that magic of the experience. I think one of the

the experience. I think one of the canonical examples is Facebook's X number of friends in Y days as the activation moment when if you cross that

threshold then you you Facebook has a lot of confidence that you're going to be a highly retained user. And so for you it makes a lot of sense that it used to be voice dictation tools were just so

clunky that they were borderline non-usable. I think for a long time now

non-usable. I think for a long time now that technology has been good enough to get an output that is identical to what you said, but what you said isn't necessarily what you want in a written

communication. And so these autoedits

communication. And so these autoedits that you've built into flow just lead to a more compelling output. It's not what the user said, it's what they want. And

that's actually two different things.

Um, so let's talk a little bit more about some of the early design decisions that you all made that were were different from some of your predecessors.

Um, what are some of the other specific features that you've built into Flow's user experience that that make it look and feel different?

The first one I'd say is most other apps need you to go inside their application to voice dictate something. Then you

have to copy paste it back into where you want. That's extra steps. Our whole

you want. That's extra steps. Our whole

company is built upon removing those extra steps.

So we did the harder upfront work of making it something that runs in the background and works in literally any single application. You just hold on a

single application. You just hold on a button, speak, let go. You don't have to think about any integrations. You don't

have to switch tabs. You don't have to do anything at all. And that from an engineering side is insanely hard to build. But that to us was a

build. But that to us was a non-negotiable from when we were starting because the other thing you get when you build something that's seamless is it's just not a product that is stuck within

Silicon Valley. There are people who are

Silicon Valley. There are people who are the least tech friendly tech forward people who can just start using flow cuz it's trivial to use.

The other big major design decision we made was most of the dictation systems show you the text as it's streaming in word for word.

We actually decided to not do that for two reasons. The first one is it is

two reasons. The first one is it is actually distracting to read what you said half a second ago

and there is very little value to that.

And so what we do is instead we show you everything that you said at the end of it. And so you speak for a while and

it. And so you speak for a while and then Flo just said cool here you go here's what you said. The other thing you get from it is I don't know if

you've ever sat in a in a Tesla and you know how it would track the path of where the car is going and it would just like vibrate like nobody's business and you'd be like I don't know if this car knows what it's doing.

Yeah.

When these voice dictation systems like are streaming text in real time, they often go and edit things afterwards. Uh

you've probably probably seen this with all of them. that actually loses user trust because they're like it doesn't seem like you you've got like your handle on this really

and if instead you're like I'm going to give you the perfect thing at the end and I'm not going to really tell you like all the hard work that goes into coming with this changes the user mindset generates a lot more trust and

those were two design decisions that were very different than what any other system had done so far.

Yeah. Well, and it seems very complimentary because you said earlier that these auto edits and the sort of zeroedit send within 0.5 seconds after completing the message is is the number

one thing that differentiates flow from all of its predecessors. And it's very complimentary with the design decision to not show how the sausage gets made.

Right? We're not going to show you everything you're saying as you're saying it. We're just going to save that

saying it. We're just going to save that to the end. And that also gives us the additional degrees of freedom to be able to make modifications before we present the final output. So the whole is almost greater than the sum of its parts

because these features play nicely together. to the extent that you can

together. to the extent that you can share. I know you don't want to go deep

share. I know you don't want to go deep into the technology, but you you talked about how number one seamlessly integrating this experience across all of the different contexts,

the different applications, the different use cases that a knowledge worker or other early adopter of your product would want to use is technically complex. to the extent you can share

complex. to the extent you can share like what what what are some of the challenges that you have faced and how have you gone about overcoming some of those challenges?

Here's one uh if you try to create a bulleted list in notion versus Slack versus Gmail versus Outlook versus a

Notepad text editor that doesn't support bulleted lists. It all looks and is

bulleted lists. It all looks and is different under the hood. M

oh so we needed to have some way where it would just work where in every single application and for reference in this

year people have used whisperflow in over 25,000 apps and websites.

Wow.

With zero that it did not work in.

And so that's like one kind of complexity. The other one is, this is

complexity. The other one is, this is actually really annoying when you paste things into Outlook or or Gmail.

I was hoping I was hoping you were going to bring up this example because it is so frustrating.

Oh my god, it's so bad. And then that was another thing that we had to fix and and then get to work in those systems. And it's all these small details, right, that that add up, but once you start

knocking out each of them, and this is such a system that works. whenever any

of our users had tried out another product, they just immediately come back to flow a day later. And it was like some small frustration that the other product didn't solve cuz it's it's hard.

You need lots of attention to detail.

And I think that that's one of the uh the unique things that I do pride our team on is they're insanely just maniacal about getting

everything perfect and right and every small detail just gets surfaced up.

Well, it's one of the nice things about having a product that while it's extraordinarily technically complex under the hood, yeah, it's actually relatively simple in terms of the amount of product surface area

you're showing to the end consumer. And

that allows you the focus to be maniacal about getting every last detail right.

Yes. One other implication of what you said I think but keep me honest here is because you've had to do all of this engineering around understanding the

context that a user is in down to in many cases sounds like the specific app that also means you can infer what sort of message they're

trying to send an email type out a presentation write a memo a Slack message and you talked earlier about how

the tone and the specific words that somebody is trying to say often look different across those different contexts. And so am I

different contexts. And so am I understanding correctly that when you use flow to dictate a message into email versus Slack versus notion, your project is actually making different decisions

on the margins about how to take the exact words it heard and then format it into a final output that is well suited to that application.

Exactly. It it does just that. And this

is a segue into the next thing that we're going to be adding soon, which is the the insight that we had is yes, even though your emails may sound different

than your Slack messages, even within that there's nuances. You might talk differently in a public thread versus in a single DM with your best friend who works at the same company. And so that's

the next thing that we're going to be shipping where it understands the context of every single thread that you're in and then understands your tone there and tries to match that

which is the next magic that we're going to be shipping soon.

I love that. Well, and I I spent a few years early in my career in venture capital as an associate. And one of the things I would look for is, you know, a

founder whose vision goes like multiple levels deep on the level of detail and specificity that you're paying attention to to solve problems at the level of nuance that, you know, a lay person just wouldn't even think of.

And to to me, this is such a great example of that. Like as a consumer writing an email or writing a message in Slack or using voice dictation tool to do that, you would never think about all of the nuance under the hood of why

these messages are different and how the applications are coded differently and what that means in terms of formatting bullet points. But you're all the way

bullet points. But you're all the way down to that level and that's to me that's what leads to um truly exceptional uh product design. I guess

another question is how did you decide to launch across platforms sequentially and what sort of lessons did you learn as you went from one platform to the next to the next to the point where you are now where it now works seamlessly

across platforms. We started off essentially balancing out the technical effort that it would take with where are the markets that are

easiest for us to tap into. So for the first one, we knew that we have an incredible audience in Silicon Valley, New York, LA, some of the leading metro

cities in the world. And Macs are most popular there. So we decided to start

popular there. So we decided to start off with Mac on desktop. It was also the platform that had the most support for us to be able to grab context, build a

really delightful experience. And so we decided to launch for for Mac first. And

when we launched for that, it became very clear that that is a lot of work that goes into building this system work

well on any platform because whisperflow works at an operating system level. And

so it's gathering a lot of context there. And so once we had built it for

there. And so once we had built it for that, we understood what we needed to do with that. And then we pushed it over to

with that. And then we pushed it over to Windows. Now you'd imagine like desktop

Windows. Now you'd imagine like desktop desktop same thing except it it wasn't.

And so that again took a few months of work. And for us, we never want to ship

work. And for us, we never want to ship a half-ass product. Our bar for shipping something is we actually have about 10 times the number of features inside the company that you as an external person would see.

Mhm.

And every single one of them goes through the ringer until it heads product market fit internally until half the company's like, "Holy I cannot live without this thing. This thing is magical."

magical." Mhm.

Then and only then do we ship it externally. And so to pass that high bar

externally. And so to pass that high bar takes time. And so we're like, okay,

takes time. And so we're like, okay, then we're going to prioritize Windows, which we did. And after that, my my initial assumption was, okay, we hit product market fit with dictation on

desktop. Uh, dictation on mobile, same

desktop. Uh, dictation on mobile, same thing. Biggest, no, no, not biggest

thing. Biggest, no, no, not biggest mistake, but rookie mistake, right?

Because it wasn't completely different surface, completely different platform.

The only thing common was you speak and it writes things for you. And so very soon within a month of using the mobile app, it was clear that we need to go

from 0 to one on product market fit on this thing again. And that honestly took 6 months. It took a lot of hard work.

6 months. It took a lot of hard work.

The mobile app was launched about 3 weeks ago. And right now it's on build

weeks ago. And right now it's on build 152.

Wow.

It has gone through 150 revisions to get to this point where we were proud to ship it to an external user base. And um

I think right now they did a phenomenal job of it. But that was the essentially like core of it. And now that we're building for Android, again building for

Android and iOS, you cannot assume the same thing cuz Android is fundamentally different on the operating system level.

And so we're essentially going to have to do a lot of these things from scratch again. And Android, I think, is going to

again. And Android, I think, is going to be one of the best experiences we would have shipped uh given some of the things that we're seeing uh internally and and I cannot wait to have that go out.

Yeah, that's exciting. One thing that doesn't get talked a lot about is I guess what I'll call product culture fit, right? On one end of the spectrum,

fit, right? On one end of the spectrum, you have Facebook move at least in the early days of Facebook where they were a social network and it

was all just about rapid experimentation and keeping up with uh the rapid early adoption of users on college campuses to on the other end of the spectrum you've got Apple as a good example where you

know Steve Jobs's bar was famously high and in a context where you're building hardware tools and specifically trying to build some of the best hardware products in the world you know that makes sense.

So where on that spectrum do you think Whisper falls as a company and Flow falls as a product? And to what extent do you think Whisper's early culture is a function of you and your co-founders

personalities versus a deliberate decision you made because of the products you're building and the customers you're serving.

I think we're pretty close to the Steve Jobs bar on things. And that is because of how fundamental the product that we're building is to

people's lives. Imagine if

people's lives. Imagine if every 100 times your keyboard stopped working. You would throw that keyboard

working. You would throw that keyboard out the window.

Yep.

If every hundred friend requests you send on Facebook, one of them didn't go through, you wouldn't even notice.

Right. And it's so so fundamental about the kind of product you're building and how it fits into people's lives and the bar of reliability that you need. And so

that is one of the reasons why we have been so focused and reliability is one part of it. The other part of it is just like taste and product design how

intuitive things feel because the product is all about building building habits. And so I think those two things

habits. And so I think those two things make it uniquely uh very focused on uh a lot of detail and iteration and and polish that needs

to happen. funnily enough that is not my

to happen. funnily enough that is not my personal way in which I build things as um somebody who is my my mental model is

a lot towards growth right like historically for all the products that I've built it's like cool here's an MVP after 24 hours of like pulling an allnighter let's ship it let's see how

people are reacting let's go and update it afterwards and build that but this was one of the things that I personally had to adapt uh because it needed me to adapt and

there are some things that we'll do where we're going to be moving at that pace.

Uh but the hard thing was how do you manage that insanely fast pace of building with also that product quality and so the the answer to that question is you just do a lot of those reps internally.

Yeah, I love that. And I feel like a lot of the best founders do adapt their culture and their approach to the product they're building and the customer they're building it for. But uh

it's it's also very easy to just get set in your ways and fall back on your own tendencies. And so um that's great to

tendencies. And so um that's great to hear that this was a deliberate choice and it sounds like it's um very compatible with the company you're trying to build. So now let's

also my co-founder Sahes like he is he is very much more on the spectrum of like focus like just do things right and so he's been an incredible balancing force to me as well. And so

I I add in the hustle and move fast. He

adds the polish culture and together it's polish with speed that we end up delivering as a company.

Sounds like a good a good yin-yang team.

Our relationship genuinely is.

Well, we we've talked about product culture fit. Let's talk about product

culture fit. Let's talk about product market fit. And let's start with who

market fit. And let's start with who have your early adopters been? Like is

there a particular segment of customers that you find you're resonating most with and and why?

Our user group has been surprisingly very very varied. And so a few kinds of people who love this product is a lot of founders and business leaders. For them,

Jared's job is primarily communication first and that's how they get leverage.

That's how they make the biggest decisions and manage large groups of people across their customers, teams, whatnot.

Apart from that, the other groups we've really seen really strong product market fit are people who are very thoughtful.

They have a lot of thoughts. They want

to put those down. They want to think through it. That's how they kind of

through it. That's how they kind of reason through things. And finally, it's a lot of developers and engineers who today are using AI and prompting for a

lot of work that they're doing. And for

them actually the the insightful thing is when they are speaking in English and they're writing code with their fingers, there's actually no context switching happening versus if they're prompting

and writing code and constantly switching between the two of them before.

And so that actually lets them stay in the flow. And so we saw these people

the flow. And so we saw these people come in as early adopters, just had incredible um organic word of mouth growth through them. And the other group

of people which is actually about 30% of our user base is what I would call nontechnical users.

Mhm. These are people who have maybe used chat GPT, right? And the way we got these people is uh people like you or me would find this tool and just be like,

"Wait, this is great cuz I have like my my dad or my grandfather has like this issue, arthritis, can't type, very slow at typing, like hunt and peck, you're doing this, and if this is there for

them, like this would be incredible."

So, it's all of these people who are onboarding their parents. And so we actually had to like spend a lot of time with with old people, we call them silvers, uh to to sit with them and figure out

what parts of it are not intuitive, how can we make this even more intuitive for you? Cuz if that is the lowest common

you? Cuz if that is the lowest common denominator we're building for, I think the product becomes intuitive for the average human being.

Yeah. Yeah. That's so interesting. So it

sounds like you started with what I would have expected to hear, which is you have a lot of these business leaders, tech co-founders, Silicon Valley insiders using the product. You

also got a lot of engineers and software developers using the product.

But then what's been surprising is it is such a horizontal use case, right? We

all spend lots of time typing on our desktops and our mobile devices. And so

you've been pleasantly surprised to find that a lot of counterintuitive use cases have emerged. One of them being um

have emerged. One of them being um senior citizens or people who are getting a little bit older and so their typing is slowing down or they have a physical disability that makes typing difficult and this this has been a real

godsend for them. You talked earlier about this metric you have which is half a second between finalizing a message and sending it which which is so fascinating. Are there any other

fascinating. Are there any other particular metrics or key performance indicators you've been using to track your progress towards product market fit and then strengthening that product market fit over time?

The biggest one we've noticed is this metric we built inhouse which is called zero edit rate which is what percentage of messages are perfect and ready to

send. And if you compare Apple, Google a

send. And if you compare Apple, Google a lot of the other basic dictation platforms that at about 10 to 15%. Mhm.

So only 10 to 15% of the time is what you have said and what they produce is you just press send on. The rest of the time you have to edit it.

For whisper flow that today is at 81%.

Four-fifths of everything that Whisperflow produces people just press send on. Very rarely do they need to

send on. Very rarely do they need to edit it.

Yeah.

And when they're usually editing, we found that it's because they changed their mind afterwards or they're like, "Ah, this is not exactly what I wanted to say." And so the the thing that is

to say." And so the the thing that is what we found is the closer or the higher we get that metric even a 1 percentage point improvement on that massive impact and how happy users are

with it. And this is something that has

with it. And this is something that has been growing as we've been improving the the technology learning more from how people are using the system. And so I think it's gone from about 40% from when

we started the early prototypes to 80% now. And it just keeps increasing as we

now. And it just keeps increasing as we keep improving the product. And this is interesting because this goes to one of our core goals of making a voice interface people trust

and you will trust us if you don't have to go and correct everything it does and you're just mostly like okay FL's got this I I'll let it do its thing.

Yep. So it's both the edit rate which if I heard you correctly 80% of voice memos that get dictated into flow do not get edited. So four out of five are edit

edited. So four out of five are edit free. And then the amount of time that

free. And then the amount of time that it takes between when a user completes the message and when they send it. And

that is averaging half a second now, which basically means they're not editing it at all.

Editing, they're not even reading it is the bigger point.

They don't even need to look at the content. They've used it enough and they

content. They've used it enough and they trust it enough that they'll just hit send without without even giving it a second review. Uh, which

second review. Uh, which yeah, I mean, if if you need an indicator of of strong product market fit, that's that's about as good as you're going to get. We talked about who your early customers have been. I'm also

curious, you talked about there are these 25,000 different applications that Flow has been used on so far, but I imagine there's a very steep power curve. And so what are the top, call it

curve. And so what are the top, call it 5 to 10 applications that people are using Flow with? And I'm particularly interested in the degree to which people

are using flow to dictate into LLMs like Chat GBT and Claude because I find that there's such a complimentary relationship between what you've built

in those tools because the tools are smart enough that they can take an even an imperfect output and they can they can create something really powerful out of it and then you can iterate back and forth. And so that's one of the number

forth. And so that's one of the number one things I've been using flow for. I'm

curious if you're seeing that with your customers as well. Yeah, the power law is actually not as strong as I would have expected. So the top 20 apps

have expected. So the top 20 apps actually just account for a third of the usage.

Okay.

And some of the top ones that we've seen used, you'll get your usual suspects which is like claude, chat, GPT, perplexity that people use flow in. The

very interesting thing I found about that was those all products have their own mic buttons. They have their own voice input.

Mhm. But what we found is people still love using Flow for it.

Yeah.

Faster, more accurate. And more than anything, people have the habit now of using Flow wherever they are. And so you just wouldn't go press that button there cuz you're just like, I know Flow. I

know it's going to work. I have my habit of doing this. I'm just going to keep doing that. We see people use that in

doing that. We see people use that in these LLMs a lot. The other place where we see people use it a lot is actually emails,

which to us is a is a big compliment to the product because people today send emails for their professional communications, especially ones that are

going out to customers, people external to the company who you want to have a really good impression in front of. And

what flow essentially does is it just formats your messages so coherently, so easy for the other person to read that people trust it for those external

messages. And then finally, the the

messages. And then finally, the the category that we see um people use it a lot for as well is in coding. So cursor

is I think one of the biggest platforms that we have and there are a couple of users who've genuinely dictated one and a half million words into cursor in the last 6 months.

I don't even know what they're building but it it's wild.

That is wild.

Yeah.

Um all right, last related question around product market fit is just was there a particular product launch or moment you can point to when you really knew you had product market fit? I feel

like product market is one of those things where it's sort of like you know it when you see it. So what was that moment?

I I'd always heard about that. You'll

know it when you see it. But then we felt that and it was like okay yeah this we know exactly what people meant by this. That started happening in January

this. That started happening in January of this year. It wasn't a launch.

There's nothing else. Apart from the fact that every day there were 5 to 10 people who unprompted would post about how much they loved flow on LinkedIn on

Twitter. For most of these people, if

Twitter. For most of these people, if you go and look at their posts, they have never posted about a product in their life.

And suddenly, this was this one company that came up and they were like, "This has completely changed how I live my life." And then other people in the

life." And then other people in the thread was like, "Okay, I just tried it out and this is incredible. Thank you so much for finding it." And you just saw that happen purely organically.

And I think at that point when people describe the product as this has changed my life, you don't even need to ask them the question like would you be disappointed if this goes away cuz sure more than disappointed if this goes away.

Yeah.

Um and the second thing that happened around then was was January was the time when we started really seeing uh exponential growth happen within the company and that's when a lot of systems

started breaking.

Mhm. Every time flow would be slower for an hour, our support inbox would be filled to the brim. Initially, I was like, "Oh god, so much to respond to."

But then my reaction was like, "Wait a second. If this goes out for a little

second. If this goes out for a little bit, these are the number of people who care enough about the product and where it matters to them that they have taken out the time to send us a message like

asking about how to fix it. Some of them are very frustrated and then we fix it and they're like, "I love you guys."

Like that that happens all the time. But

it's like, "Wow, this is making people feel things and people that's way better than just them being ambivalent to what you're doing."

you're doing." Yeah. You know, it's funny that anecdote

Yeah. You know, it's funny that anecdote reminds me of an experience I had when I was in Mature Capital. I mentioned that earlier. And this was between 2013 and

earlier. And this was between 2013 and 2015. And so it was right around the

2015. And so it was right around the time that a lot of VC firms were starting to try to apply more and more quantitative analysis to sort of pattern matching what early indicators leading

indicators of rising star you know seed series A series B companies would look like. Obviously once you get to a

like. Obviously once you get to a certain stage you have tons of data but in in the early days when a company's just getting escape velocity it can be hard to recognize from the outside. And

one of the insights I I stumbled upon was that specifically within the social networking category, if you looked at um the number of ratings and reviews that a

social network got and then you looked at the average rating out of five stars, you wanted a very large volume of ratings and reviews. That's sort of obvious, right? If you don't have a

obvious, right? If you don't have a certain amount of volume, it means not enough people are interested in the product to be rating it. But

counterintuitively, some of the most successful, fastest growing social networking startups actually had middling ratings. So what

so what you were looking for was the combination of a very large and rapidly growing number of ratings and reviews, but actually somewhere more in the like 3.5 to four star range, not necessarily

4.5 to five star. Because what that meant was you were pulling in enough people fast enough who had strong enough opinions that they cared about this product. and and where it was going. And

product. and and where it was going. And

if everybody was happy with it, that probably meant it wasn't being bold and ambitious enough with its vision or it wasn't different enough from what already existed.

And so what I hear you saying is something similar and that's that as the founder and CEO of this company, you're sort of this is your baby and it hurts to hear people write in and say that they're frustrated, but at the same time it's sort of a backhanded

compliment because for somebody to take time out of their busy day to write you that email, they must have cared enough to want you to fix it. Yeah. Um, and so I think that's that's a good reminder to all of the founders out there who are

putting in late hours trying to get their company uh to the point where um people care enough to be frustrated.

Um, all right, let's move into our last section here, which is around growth strategy and how you've gotten early adoption. Subversive at the end of the

adoption. Subversive at the end of the day is ultimately a podcast about growth. So much of your early growth has

growth. So much of your early growth has been organic. How have you managed to

been organic. How have you managed to generate so much viral word of mouth among adopters outside of the obvious which is just building a world-class product? Is there anything else you've

product? Is there anything else you've done to really get people to be talking about this online or offline?

There are a couple of very intentional things that we learned and that has driven both of what we do and what we don't do. The thing we realized very

don't do. The thing we realized very interesting about the space we're in is if I show up to somebody and I was like, "Hey, I'm building a voice dictation tool. You should try it." No one's going

tool. You should try it." No one's going to try it because they're like, "Yeah, so did the last hundred people who said that and I know what they delivered."

Yeah, there's a lot of distrust here. But if

you went and you told somebody, "Hey, I'm trying this voice dictation tool and it's insane." No one has ever said that

it's insane." No one has ever said that before about any voice dictation tool.

So that is that is what generates trust.

And so that is the thing that we realized very early on that this is what is going to sell this product. It's not

me. It's not Whisper saying things about itself. It's other people talking about

itself. It's other people talking about the value prop.

And so what we did was early on um we reached out to a lot of influencers, lots of people who make content. And for

me, it wasn't even like sponsorship. It

was like, "Hey, here's a tool. I think

you might like it. Use it for free."

Y some people did. And the great thing about those influencers make using it is it just showed up in every single one of their videos.

As a few of these creators, you go on their website, every single video they're just using flow. They're talking

about whatever else. There's other

companies who are sponsoring and paying for it. But Whisper Flow is the star of

for it. But Whisper Flow is the star of the show for the next two and a half hours. And every third comment is, "What

hours. And every third comment is, "What are you using for voice text? What are

you using for voice to text?" Most of these people, not affiliates, not sponsors, no nothing. And like they just reply like whisper flow, whisper flow, whisper flow, whisper flow.

Yeah.

And that has been huge because what we what we found is like yeah, these people just want to learn from the best and copy whatever they're doing.

Yeah.

Uh so that worked really well and also focused our team and that's why our growth team is so small. Like we don't produce a lot of content by ourselves.

It's not going to do well. we have other people produce all that content for us uh for essentially giving them a free euro flow.

Yep. I mean it's one of the advantages of being such a horizontal use case.

There are certainly plenty of challenges and I had Rachel Hepworth on from from Notion to talk about notion's growth story.

You talked about the challenge of being a product that has so many different use cases that people don't where know where to begin. But one of the advantages is

to begin. But one of the advantages is if you can get past if you if you can get enough activation energy to get past a tipping point, then it can be used in all of these different contexts. And I

would have never thought of of that idea, but it makes so much sense that rather than pay an influencer to go do a sponsored UGC video for your product, just give them the product for free, let

them play with it, and then if it's compelling enough, they'll start using it all their other videos, and you sort of become a free rider on all of the UGC content they're doing with all these other companies. So that's that's

other companies. So that's that's amazing. What what a great strategy.

amazing. What what a great strategy.

Sometimes there's one or two people who then message me and ask me to do a podcast with them. There you go. There

you go. Free marketing for Flow right here.

Yeah. Um,

I didn't even give it to you for free.

That's true. That's true. Although, I do think I have a free membership now, but I did have to pay for that first month.

So, maybe I should ask for a refund.

That's a good point. You You'll be getting a customer support ticket from me later this afternoon.

Um, how are you thinking about your medium to long-term growth strategy?

Like, you you at this point, you've got product market fit. You're starting to see this exponential growth. the

momentum is building, but but where do you think your growth will come from over the next two to three years? And

then I'm always curious to ask, how are you structuring your go to market team to support that strategy?

That is a phenomenal question. That's

actually one of the key things that I've been thinking about a lot uh recently and there are there's two motions that I've broken it into. The first one is

acquiring a lot of individual users and grabbing a lot of mind share.

Mhm. And then there are a lot of kind of growth tactics to go and do that and very intentionally within the company every growth initiative is broken down into is this for awareness or is this for conversion.

Yeah.

Right. Because then that doesn't confuse those two things and just keeps the goals very clear. Um and the second part of it is okay we have an individual user

on board and the unique thing about flow is it has really easy distribution within companies. So, how do we take

within companies. So, how do we take that individual person and then upsell them into getting their entire company onto FL? And we did that recently with

onto FL? And we did that recently with companies like Superhum and Substack and Mercury and a lot of law firms, a lot of other financial institutions and hundreds of other companies which has been been

doing well. And so when I think about

doing well. And so when I think about our growth motion, it's it's interesting. It's a combination of this

interesting. It's a combination of this B2C and B2B that we're running together.

Culture- wise, I think the biggest thing that we wanted to set up from a B2C perspective is the is a norm of

experiments and a norm of doing things fast. So, if if you're on my growth

fast. So, if if you're on my growth team, Phil and you come up with an idea like I think we should do this thing.

Mhm.

And it doesn't sound ridiculous, I want that to be done and shipped anywhere from between the next 1 to 5 days.

Yep. I don't want to have any extended like back and forth of does this make sense? Can we do this any better? It's

sense? Can we do this any better? It's

like no, ship it. Let's try it. We learn

if we want to make it better. And the

more we could do to support that velocity within the organization and make people very independent to be able to do that, I think the faster we move overall. And that's that's hard to have

overall. And that's that's hard to have that mindset cuz sometimes people are afraid of making mistakes. Sometimes

people are perfectionists and you want to get things done well like myself included.

Yeah.

And that has been one of the things that we're we're kind of like pushing forward as the mindset for every single person in the company to have is just ship it.

And and this is interesting, right?

Because you'll notice in the first part of the conversation I was talking about how we're very intentional and thoughtful about building products and everything and in growth it's different. And so

what we essentially have to balance is the growth mindset of like the the shipping fast and encapsulated to a certain set of projects and the deeply

thoughtful lots of iterations, perfection inside the product and have these two cultures exist side by side within the same company.

Yeah, I'm so glad you brought that up because I was going to go back to that as well. It sounds like on core product

as well. It sounds like on core product you have a very high Steve Jobsian bar of what makes it out into the world, but on growth product it's the complete

opposite. It's move fast and break

opposite. It's move fast and break things. It's let's throw stuff against

things. It's let's throw stuff against the wall and see what sticks. And it

reminds me so much of my time at Quizlet where um I came in as uh the first growth product leader. uh we had one product

product leader. uh we had one product manager at the time working on growth and we expanded um the growth product

team um into a much larger group of PMs, designers, engineers, but we had to very much instill this culture of the way you approach a problem as a growth PM or a

growth engineer looks very different than how you would approach it on core product because a significant percentage of the experiments we run are going to be failures and then we're going to learn from our mistakes and move on. And

the last thing we want to do is goldplate the code for an AB test that may not go anywhere. So,

it is a different mindset.

Um, last question before we get to our lightning round. So, you've already seen

lightning round. So, you've already seen global adoption. More than 50% of your

global adoption. More than 50% of your users are outside of the US. What's been

your secret to expanding internationally so quickly or is it all just happened because people love the product and they're just talking about it across borders?

Huge part of it is organically. The

second part was unlike most other companies the first version of whisper flow that we launched worked across a 100 languages and so from day one we built it for

international audience. Uh part of it

international audience. Uh part of it also came because in our team of 16 people we speak a total of 30 languages which is 30 different languages which is wild and so

the the team is very international and so that that culture just seeped into the product. It's it certainly helps

the product. It's it certainly helps when you're solving for your own need.

And so if you have if you have the luxury of having a a early team that speaks that many different languages, then I could see how that would make things easier.

Well, let's get to our lightning round.

I want to ask you a series of rapidfire questions and you can just give me short responses. The first one I always like

responses. The first one I always like to ask is just what do you think is the best part about working at Whisper and what's one thing that you would change?

I wish our team was in person. Our team

is 30% remote and whenever they come into the office it is the most fun week and I just I just wish uh and and most people can't can't move because family

reasons but I I wish they could.

Yeah, that's something I hear from founders in particular and to me it's so emblematic of you have a lot of individual contributors who will say the opposite. It's such a relief to work

opposite. It's such a relief to work remotely and it's so much easier to balance work and family, especially if you have kids. But as a founder or as a founding team, you know, this is this is

one of your families, right? You're the

company you're building is it very much feels like your work family. And I can totally relate to

family. And I can totally relate to that. Like being together with your

that. Like being together with your colleagues in the office, there's there's no feeling quite like it. And

so, while there are many wonderful aspects of remote work, I can I can certainly understand where you're coming from. Um, what is the coolest AI app

from. Um, what is the coolest AI app you're using right now that most people haven't heard of?

I would say Sunno.

I use it to make music and it is maybe the most underrated AI product of all time cuz I think video generation is just about getting to a point where it

starts to feel real. Image generation

questionable how good you think it is.

But Sunno is incredible.

So good.

I'm going to have to try it. I've heard

SUNO come up a couple times actually on this podcast because there are a lot of creative teams that are using Sono to apply music tracks to uh video ads, but I haven't used it myself. So, I'm going to have to go to try.

It's great for that, but the songs I make, I listen to more than Spotify now.

Oh, wow.

I did not expect that to happen.

Very cool. I'll have to give it a try.

Um, if you could add one premium feature to Flow tomorrow, what would it be?

I have so many. I have so many. Uh the

biggest one would be for it to be able to do things for you and take actions on your behalf.

So we've come full circle back to the agentic vision you had when you first founded Whisper because you said the first seven products that Whisper built were sort of agentic actionoriented products. flow is this voice dictation

products. flow is this voice dictation tool, but now you want to build on the foundation you've already built with voice dictation to be exactly because now you have the habit and now you can stack on a lot of things. Starting off with that would

things. Starting off with that would have been a mistake.

Yeah, makes sense. Um, what do you think is the toughest user experience challenge for voice first apps over the next two to three years? I think we're going to see more and more of these products that are using voice as sort of

the native interface of the product. But

what do you think the UX challenges look like there? The biggest one that I'm

like there? The biggest one that I'm starting to see happen, this is especially on Apple, is all of the privacy limitations and popups and

restrictions that they are starting to put in. And that I think harms the user

put in. And that I think harms the user experience the most. That's one thing that I wish they would change.

Yeah. Well, you certainly aren't the first to complain about Apple and their approach to privacy restriction.

Last one. If you could go back, you mentioned that you started to code, I believe, when you were around 12 years old. So, if you could go back, what

old. So, if you could go back, what would you say to your 12-year-old self who was just learning to code for the first time about building technology products?

Honestly, I think you did a fine job. If

I if I changed anything about that, I wouldn't be where I am and I wouldn't be the person I am today. through all the struggles and all that I had to go through myself. All the apps that barely

through myself. All the apps that barely had any users that led to all the apps that had millions and I think that learning is a part of the process. So

I would just I would just tell him to keep doing what he thought was the right thing to do and he'll figure it out.

I love that. I mean it goes back to the growth mindset you talked about at the beginning of the episode and it's one thing to say you have a growth mindset but it's another thing to actually live it. And it sounds like uh you've been

it. And it sounds like uh you've been very authentic to that even even from the time you were you were just getting started as a young kid. Well, thank you so much tonight for coming on. This has

been a really enjoyable discussion and I I've certainly learned more about about the product than I even knew even though I've been a user for a while now. Any

parting words you want to give the audience before we wrap up here?

I the biggest thing I'd say is we just raised a $30 million series A and now on track to grow faster and faster exponentially. And if you're somebody

exponentially. And if you're somebody who's excited about growth, excited about building incredible products, especially ones that put smiles on people's faces day in day out, then drop

me a note and I'd love to chat more.

Well, you certainly have my vote of confidence. It's a great product and as

confidence. It's a great product and as a growth professional myself, I think it's a great company to be building your career. So, thanks again, TA. This has

career. So, thanks again, TA. This has

been a pleasure and um I I really appreciate your time today.

Phil, this is fantastic. Thanks for

having me.