July 8, 2025 • 21 mins
Interested in being a guest? Email us at admin@evankirstel.com
What if your meetings could remember everything so you don't have to? That's the promise behind Granola AI, the innovative note-taking tool designed to augment human capabilities rather than replace them.
In this revealing conversation with founder Christopher Pedregal, we explore how Granola differentiates itself in the crowded AI productivity space by focusing on what truly matters - fitting seamlessly into users' workflows and addressing the small "paper cuts" that plague our daily meeting routines. Unlike other solutions that deploy bots to meetings (creating what Chris calls an "uncanny valley" effect), Granola uses your device's microphone to work consistently across all meeting formats without storing potentially sensitive audio or video recordings.
The philosophy behind Granola stands apart in today's AI ecosystem. Chris articulates two competing visions for AI products: those that replace human capabilities versus those that augment them. Granola firmly commits to the augmentation path, giving users what he calls a "steering wheel for AI" - meaningful control over how the technology serves them. This approach recognizes that outsourcing busy work is valuable, but outsourcing thinking itself leads down a problematic path.
Most fascinating is Granola's evolution beyond mere note-taking. With their 2.0 release introducing team features, Granola is positioning itself as a comprehensive AI-powered workspace that understands your context and helps you work more effectively. As Chris explains, this puts them on a trajectory to compete not with other note-takers but with foundational AI providers like Anthropic and OpenAI.
For anyone interested in how AI is transforming productivity tools or entrepreneurs looking to build in this space, this conversation offers invaluable insights from someone at the forefront of reimagining how we work. Ready to let AI remember your meetings so you can focus on what truly matters? Granola might just be your answer.
More at https://linktr.ee/EvanKirstel
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Hey everyone, fascinating chat today.
What if your meetings couldremember everything so you don't
have to diving into the worldof invisible AI?
Today, with granolaai, chris,how are you?
I'm doing great.
Thanks for having me, evan.
Well, thanks for being here,really intrigued by your mission
(00:21):
and vision.
Let's start with the basics.
What was the big idea behindgranolaai?
What problem were you trying tosolve?
Speaker 2 (00:30):
So, for folks who don't know, granola is an AI
notepad for people inback-to-back meetings.
So the idea is kind of the wayyou would take notes in Apple
Notes or an app like that.
You can do that in granola.
The difference is that granolalistens in the background,
transcribes and then, when themeeting's over, it takes
whatever notes you've writtenand then makes them great like
(00:50):
fills in the stuff you didn't.
You didn't write in terms ofwhat was the original idea, the
original idea.
I came across gpt3 at the timethis was maybe seven, eight
(01:13):
months before chat, gpt launchedand um I I was just fascinated
by this technology.
It felt like alien technology.
It was like really smart.
In some ways it could write acollege essay, but it couldn't
do basic math and it was reallyconfusing and interesting and I
started playing with it andexperimenting.
(01:34):
We use for work like knowledge,work, productivity, tooling.
All of that is going to getreinvented on top of AI, on top
of these large language models.
So my co-founder, sam, and I wealways envisioned okay, what is
(01:56):
the AI native tool of thefuture to help us do work, and
that's still what we're tryingto build with Granola.
We just decided to start withone very specific use case,
which is helping you take notesduring meetings.
Speaker 1 (02:08):
Fantastic.
And before Granola, you foundedSocratic, a great name, and
sold it to Google.
Tell us about your Googlejourney and how did that help
you build Granola?
Speaker 2 (02:20):
So it's just so different to do a startup than
to be in a big company.
You know, I first I joinedGoogle out of college.
I was it's like my first joband then I quit it after four
years and went to go to Socratic.
And I did that for five yearsand then when I came back to
(02:40):
Google, it had just become amuch larger company on one hand
and I had gotten used to workingon a very small team on the
other hand.
So it's a bit of a cultureshock.
I think when you're buildingsomething from scratch, the
types of problems you have arejust very, very different than
the types of problems like aGoogle or Apple or Microsoft
(03:00):
have.
When you're building somethingfrom scratch, the hard thing is
to get people to care about whatyou're doing right, and I think
the the problems with the thegoogle's have to deal with is
like okay, you know we alreadyare getting a lot of stuff right
, a lot of people trust us, alot of people use us.
How can we innovate or makechanges without breaking what we
already have or breakingpeople's expectations?
(03:23):
It's just a very differentworld to navigate.
Speaker 1 (03:26):
I bet.
So a lot of AI tools come andgo.
I know I use a dozen or sodifferent tools and I have my
favorite tech stack any giventime.
Granola seems to be stickingaround.
People are using it.
Congratulations on that.
So what's been a secret or thesecret for keeping users happy
and giving them what theyactually want and value?
Speaker 2 (03:50):
yeah, I, I think in ai it's very easy to build an
incredible demo and I think wesee these all the time.
It's like, oh my god, all of asudden I can do this crazy thing
for me and, and that's greatand that's incredible.
It's really hard to buildsomething that's actually very
(04:15):
useful day in and day out, and Ithink doing that has very
little to do with AI.
It's a lot about taking all thelittle paper cuts, all the
little annoying things inpeople's workflows, and trying
to fix those or to streamlinethose, and I think we've tried
to do that with granola.
(04:35):
We spent a lot of time on that.
So it's a little bit when Idescribe granola to someone,
it's like, oh, it doesn't soundthat different from anything
else out there and it's kind ofhard to articulate why they
should use it.
But the moment someone usesgranola they tend to say, oh my
God, it was just so effortless,it just fit into my workflow, I
didn't have to think about itand kind of did exactly what I
wanted.
(04:55):
And of course you know we tryto stay up with the latest and
greatest AI, which is changingday in and day out.
But a lot of the things thatmake.
I think granola have stickingpower.
Are things about quitethoughtful, by the the way
notifications show up right andhow when you use granola it's
(05:16):
not invasive, and all the kindof small details that we had to
do along the way.
Speaker 1 (05:20):
Brilliant, and you've talked about giving people a
steering wheel for AI.
What do you mean by that?
Why is control important?
Speaker 2 (05:30):
Yeah, great question.
I think there's two main ideasthere.
One is, I think there's avision of AI where the AI kind
of replaces you, in terms of youdelegate a task and just the AI
go and does it.
And then there's a vision forAI where the AI augments you, so
instead of replacing you, itjust gives you superpowers, it
(05:51):
lets you do the task way betterthan you would have been able to
do it without AI.
And that's the vision of AIproducts that we believe in,
that we're excited about, thatwe're building towards products
that we believe in, that we'reexcited about, that we're
building towards um and and tobe able to do that effectively,
you need to give people theright controls and how to steer
(06:12):
the ai in the way you want.
And I think we are in the we'rein like the caveman days of
that in as an industry.
Or maybe a better metaphorwould be um, like like under
Kirk Pothy, I think, used thisthe other day where he said like
chatting with AI today feelslike using the terminal, felt in
like the seventies or earlyeighties, right Before like the
(06:35):
graphical user interface camealong.
And I think as an industry,we've been giving this
incredible technology and westill haven't figured out the
most human uh way to let peopleinteract and make the most of it
, and it's one of the things wewe spent a lot of time working
on at granola.
Speaker 1 (06:54):
Brilliant.
So many of us in business orenterprises are using team
collaboration tools and haveSlack or Zoom or Workspace etc.
Etc.
And you've come out with teamfeatures.
I guess you call it Granola 2.0.
What's changed?
How does that work and how arebusinesses using your service?
Speaker 2 (07:16):
Yeah, when we launched Granola, it was very
much a tool just for you as anindividual and you're in control
and the notes are bespoke foryou.
For example, if you use Granolaand I use Granola on the same
meeting, granola would generatecompletely different notes for
you and for me, because I wouldthink about what's most
(07:37):
important for me to take awayfrom the meeting versus what's
most important for you to takeaway from the meeting.
But the thing about meetings isthat oftentimes information
shared in meetings is inherentlysocial and inherently relevant
to, like, a group of yourcoworkers, of the people you
work with, and we kept hearingfrom users that they'd take the
(07:57):
granola notes and they'd shareit with their team over slack.
That works fine, um, but themoment you so that like
something that's really powerfulin granola is if you take a set
of meetings and you startchatting with our ai across the
set of meetings, like granola,like the underlying models are
really good at pulling outthemes or ideas across that set
(08:18):
of meetings and, uh, well, youcould, you couldn't, you
couldn't do that in Granola 1.0.
So we basically built theability to share a folder, Like
you can create a folder whereyou can put a bunch of meetings
or files, whatever context isrelevant to a specific project.
Let's say and share that withinyour company and then everybody
can chat with that set ofmeetings.
(08:38):
I guess an example for us wouldbe every sales meeting someone
like Granola does goes into oursales folder that's public to
the whole company and you cankind of go in there and be like,
okay, what are customers askingfor this week that they weren't
asking for last week?
Or whatever question you mayhave.
Speaker 1 (08:56):
Very cool, so I use a number of transcription, um,
and they use bots to joinmeetings on my behalf.
I recently showed up to ameeting and there were more bots
sitting.
People, yeah, and people it wasa little strange.
Speaker 2 (09:17):
You don't use bots.
Why?
Two reasons.
One is we want, like, granolais a tool, and great tools are
very consistent in the way theywork and they always work.
So I want Granola to be as easyand consistent to use as a
pencil and a pad of paper, right, and when you have a pencil and
(09:40):
a pad of paper, you don't thinkto yourself, oh, is this
meeting on Zoom and thereforecan my bot join Zoom?
But it doesn't work.
If it's, you know, on a Slackhuddle or you know, an iMessage
call, or maybe it's just anin-person meeting, and there you
know, there's no.
Basically, I just wantedGranola to always work, and so
(10:00):
what that means is we don't.
There's no bot that joins themeeting.
Granola uses your microphoneand it listens to the audio on
your computer, which means youcan have an in-person meeting
right, there's no Zoom link oranything like that and use
Granola and it'll still work thesame.
So that's why we originallydesigned it that way.
Uh, we found that, um, bots,bots, just they're in that
(10:25):
uncanny valley, I think oftechnology, like you're, you're,
you know if you and I arehaving a one-on-one conversation
over zoom, but now there's likethis black square on the screen
and it's neither providing anyvalue in the moment, but it's
kind of making you feel likeyou're you're watched and
recorded.
Um, it just changes the, thedynamic of the conversation.
(10:46):
Also, the, the ai note takersthat had bots that, as far as I
know, most of them record theaudio and the video and they
store the audio and the videofrom the, from the calls, which
granola does not do.
Granola does not store theaudio, it only generates notes.
Yeah, and that's by design,because I think what people want
most of the time is they justwant to have great notes, like
the equivalent of if you justsat there and were furiously
(11:09):
typing notes through the wholemeeting, like that's kind of
what you want.
The audio that could be usefulfor sure, but again, that's an
invasive act to record audio.
That's like a different.
It changes the dynamic of theconversation.
Speaker 1 (11:24):
Yeah, well done.
I guess the other concern mightbe that you know note takers
that are AI driven will justmake us less engaged, or you
know tendency to zone out.
How do you think about that,balancing, you know, ai and
automation with keeping peopleengaged and present?
Speaker 2 (11:43):
Yeah, I think it's.
I think it's a really importantquestion and I think I think
our competencies as humans willwill change here.
I'll give you a great example.
So, um, uh, there are a fewcities I lived in before Google
Maps came out on a cell phone.
(12:04):
I'm old enough where I canremember this.
You know there's a life prebeing able to look up where the
heck you are at any moment onyour phone and post that and the
cities I lived in before thatyou could.
You could drop me in thosecities today and I'd probably be
able to navigate them fine,because I had to really learn my
way around those cities and thecities I lived in after Google
Maps on your phone came out.
(12:24):
I don't have nearly as strongof a mental model of the city
because whenever I was lost Icould just look it up on my
phone.
So I do think there's like anthere's an atrophy of skill
whenever a new technology kindof comes out, and I do think
that will happen with AI.
I think the important questionis are you outsourcing the busy
(12:48):
work or are you outsourcing thethinking right?
And I do think the way wedesign AI tools have a big
impact on this.
Speaker 1 (12:57):
And.
Speaker 2 (12:57):
I think there's a real danger that we design AI
tools where people by default,start outsourcing their thinking
to AI, and I think that's it'sprobably not the future society
or vision that we want to livein.
Right and sure, if you'reoutsourcing busy work, like if
it's, for example, maybe todaypeople are really good at typing
very quickly during a meetingright, because that's just been
(13:19):
a necessity to be able to takenotes, and now in the future,
people aren't able to type sowell during a meeting because
they don't need to, becausethere are things like granola
that take notes.
That's probably a fine tradeoff, right, because now people can
be more present.
If people stop paying attentionwhen they're talking to each
other, right, or thinkingcritically about the decision
(13:40):
because they think AI will catchit for them, I think that's
really problematic.
Speaker 1 (13:45):
Yeah, what a balancing act.
Speaker 2 (13:48):
And just one point on that like we did design and
again, this is just one step,but like a big difference
between Granola and other notetakers is when we launched is
that Granola is still a notepad,it's still a place where you
write your notes, so and andwhat the ai does is that it
takes your notes and makes thembetter.
So we're designing the product.
(14:08):
So it's still a place, it'sstill a tool for thought for you
, so you're still supposed toput your thoughts in there and
the ai adjusts and um andimproves what you do, as opposed
to some of these othernote-taking services where it's
just the AI dissolving right andyou just kind of trust that the
AI gets it right, and then yououtsource it and you don't think
about it.
(14:29):
And I do think that small detail.
Actually, it really changes theway Granola feels and it really
changes the way the notes thatGranola generates feel.
Speaker 1 (14:38):
Brilliant.
The way the notes that thatkernel generates feel brilliant.
Speaking of other ai notetakers, you have lots of
competition which has pluses andand minuses.
Other note takers you know bigtech companies.
What do you see as yourcompetition now?
In the future, what does thelandscape look like?
Speaker 2 (14:56):
yeah, when.
So when we, when we entered themeeting note-taking landscape,
it was already incrediblysaturated.
Companies like otter andfireflies I think they're eight
or nine years old at this point.
We've been at it for a long time, and there are lots and lots of
other companies out there doingthat, as well as the googles,
the zooms and the microsoft's.
(15:17):
Um, that's not when I thinkabout competition.
That's not what I, what I thinkabout.
I think about so that, and toexplain that, I probably just
need to talk about the future ofgranola a little bit.
So earlier I mentioned thatwith granola, we're starting off
with meeting notes, but ourvision had always been to be
(15:39):
something much larger, which isthe place where you do work with
the help of ai.
So we want Granola to be theplace that has a lot of context
about you knows about all thedifferent things that you're
doing, and we have that fromyour meetings.
And then now, how can we use AIto help you actually do your
work and your thinking?
And I think that's the futureof Granola.
(16:01):
And I think when I think aboutcompetition, I think about who's
, who's going to be buildingthat, or who's gonna building
tooling towards in thatdirection.
Um, and from that lens, I thinkit's much more the anthropics
and the open ais of the world,uh, that are going to be like a
longer term competition for us,or or a brand new startup that
(16:24):
hasn't launched yet.
Right, the space is moving soquickly, um and and you see,
like these really innovativestartups pop up, so like those
are the companies that I kind ofthink about because I think
about the future vision.
I'm less worried aboutnote-taking, specifically, I
think I think yeah, I know it'sgoing to be commoditized, Like I
(16:48):
think we have the best notesright now and hopefully that
continues, but I think the thingthat's really going to be
powerful for people is thisfuture tool that lets you do
work five times better and fivetimes faster than you'd be able
to do without it.
Speaker 1 (16:57):
Wow, very thoughtful.
You've integrated with many ofthe top AI models.
How do you decide which AImodel to use and do you see
yourself building your own modelin the future or leveraging
open source or some of theseamazing emerging?
Speaker 2 (17:12):
Yeah, we, we use whatever is the best model for
our use case at that time wealways try to use so, so one is
the best model at any, giventhat changes frequently right
within a matter of months.
Sometimes you have one of thebig labs launch one model and we
switch to that and then anotherlab launches another model, so
(17:34):
we switch to that and we alwaystry to use the highest end uh
model.
So if you're using granola uh,you should be getting the the
best possible premium experience, like on the market.
That's possible today.
Um, in terms of uh doing ourown model, it's possible.
We'll do it in the future.
The pace of innovation of thefoundation model layer has just
(17:57):
been so incredible over the lastfew years.
I expect that to continue.
It's possible at some somepoint that fine-tuning our own
model on.
We have a lot of, obviously wehave a lot of user data on like
what are good notes, what arenot good notes, like what's
actually the most importantthing for meetings, so like we
could do that in the future.
We found that that hasn't been,that hasn't been like the best
(18:20):
use of our time.
So far We've been kind ofletting the models get better
and therefore, granola gettingbetter and we're spending our
time on the user experience, thedesign, the actual product
layer, because there's a lot todo there.
Speaker 1 (18:32):
A lot to do, indeed.
And for other founders outthere, builders just getting
started or thinking of gettingstarted in the AI space, what's
your advice for them?
You've seen it all.
At this point, what would yougive?
Them for guidance or persuasion.
Speaker 2 (18:52):
Yeah, I think this space is changing faster than
any other space that I canremember since I've been alive,
(19:14):
and I think that has a fewimplications.
I think one um, a lot of thehistorical advice or best
practices might not hold true inthis.
In this, like the laws it'slike the laws of physics have
kind of changed.
We're still all trying tofigure out the new laws.
So if you're a new builder,that's a great, that's an
advantage in a way, becausewe're all kind of trying to
figure it out together.
So I think that's anopportunity.
And because the space is movingso quickly, once you launch a
(19:45):
product, once you're in the ratrace, you just kind of have to
be sprinting just to keep upright.
It's like, oh, now this companyhas launched this feature, now
you have to launch it.
So everyone's sprinting.
But when you're sprinting youdon't have a whole lot of time
to really explore and thinkdeeply about a problem.
So if you're like starting offright now, you might feel like,
(20:06):
oh no, I'm being left behind.
I need to immediately justchoose something, work on it.
I would say, actually take abreath and maybe choose a
problem you care about andcherish that time and just
really go deep and really try tounderstand it well and build
your intuitions about what'simportant in that problem area
(20:26):
or not, because once you launcha product, you probably won't
have time to do that and there'sso much value in being able to
have that deep thought on atopic.
So that'd be my second piece ofadvice.
Speaker 1 (20:39):
Wonderful advice.
So what are you looking forwardto the rest of the summer?
What's on your radar?
Any travel events or otherplans heading into the fall?
Speaker 2 (20:51):
So we're based in London and summer in London is
like a wonderful time of year.
It's kind of like that yeah,it's the time of year, the
winter I wouldn't recommend somuch much, but the summer is
really fantastic.
So, uh, no, we just have a, wehave a great set of.
It's just a fun time.
Like we have an office, we havea rooftop, like the team goes
out there, we eat lunch there,we do happy hours there, um, so,
(21:15):
yeah, just lots to build, um,and yeah, we'll be here
wonderful.
Speaker 1 (21:20):
well, thanks so much for joining.
Can't wait to try granola.
Personally, maybe ditch mycurrent transcription service
and onwards and upwardsCongratulations on all the
success.
Thank you so much.
Have a great day and thankseveryone for listening, watching
.
Be sure to check out our new TVshow, tech Impact TV, now on
Bloomberg and Fox Business.
(21:41):
Thanks everyone.
Thanks, chris.
Hey everyone, fascinating chat today.
What if your meetings couldremember everything so you don't
have to diving into the worldof invisible AI?
Today, with granolaai, chris,how are you?
I'm doing great.
Thanks for having me, evan.
Well, thanks for being here,really intrigued by your mission
(00:21):
and vision.
Let's start with the basics.
What was the big idea behindgranolaai?
What problem were you trying tosolve?
Speaker 2 (00:30):
So, for folks who don't know, granola is an AI
notepad for people inback-to-back meetings.
So the idea is kind of the wayyou would take notes in Apple
Notes or an app like that.
You can do that in granola.
The difference is that granolalistens in the background,
transcribes and then, when themeeting's over, it takes
whatever notes you've writtenand then makes them great like
(00:50):
fills in the stuff you didn't.
You didn't write in terms ofwhat was the original idea, the
original idea.
I came across gpt3 at the timethis was maybe seven, eight
(01:13):
months before chat, gpt launchedand um I I was just fascinated
by this technology.
It felt like alien technology.
It was like really smart.
In some ways it could write acollege essay, but it couldn't
do basic math and it was reallyconfusing and interesting and I
started playing with it andexperimenting.
(01:34):
We use for work like knowledge,work, productivity, tooling.
All of that is going to getreinvented on top of AI, on top
of these large language models.
So my co-founder, sam, and I wealways envisioned okay, what is
(01:56):
the AI native tool of thefuture to help us do work, and
that's still what we're tryingto build with Granola.
We just decided to start withone very specific use case,
which is helping you take notesduring meetings.
Speaker 1 (02:08):
Fantastic.
And before Granola, you foundedSocratic, a great name, and
sold it to Google.
Tell us about your Googlejourney and how did that help
you build Granola?
Speaker 2 (02:20):
So it's just so different to do a startup than
to be in a big company.
You know, I first I joinedGoogle out of college.
I was it's like my first joband then I quit it after four
years and went to go to Socratic.
And I did that for five yearsand then when I came back to
(02:40):
Google, it had just become amuch larger company on one hand
and I had gotten used to workingon a very small team on the
other hand.
So it's a bit of a cultureshock.
I think when you're buildingsomething from scratch, the
types of problems you have arejust very, very different than
the types of problems like aGoogle or Apple or Microsoft
(03:00):
have.
When you're building somethingfrom scratch, the hard thing is
to get people to care about whatyou're doing right, and I think
the the problems with the thegoogle's have to deal with is
like okay, you know we alreadyare getting a lot of stuff right
, a lot of people trust us, alot of people use us.
How can we innovate or makechanges without breaking what we
already have or breakingpeople's expectations?
(03:23):
It's just a very differentworld to navigate.
Speaker 1 (03:26):
I bet.
So a lot of AI tools come andgo.
I know I use a dozen or sodifferent tools and I have my
favorite tech stack any giventime.
Granola seems to be stickingaround.
People are using it.
Congratulations on that.
So what's been a secret or thesecret for keeping users happy
and giving them what theyactually want and value?
Speaker 2 (03:50):
yeah, I, I think in ai it's very easy to build an
incredible demo and I think wesee these all the time.
It's like, oh my god, all of asudden I can do this crazy thing
for me and, and that's greatand that's incredible.
It's really hard to buildsomething that's actually very
(04:15):
useful day in and day out, and Ithink doing that has very
little to do with AI.
It's a lot about taking all thelittle paper cuts, all the
little annoying things inpeople's workflows, and trying
to fix those or to streamlinethose, and I think we've tried
to do that with granola.
(04:35):
We spent a lot of time on that.
So it's a little bit when Idescribe granola to someone,
it's like, oh, it doesn't soundthat different from anything
else out there and it's kind ofhard to articulate why they
should use it.
But the moment someone usesgranola they tend to say, oh my
God, it was just so effortless,it just fit into my workflow, I
didn't have to think about itand kind of did exactly what I
wanted.
(04:55):
And of course you know we tryto stay up with the latest and
greatest AI, which is changingday in and day out.
But a lot of the things thatmake.
I think granola have stickingpower.
Are things about quitethoughtful, by the the way
notifications show up right andhow when you use granola it's
(05:16):
not invasive, and all the kindof small details that we had to
do along the way.
Speaker 1 (05:20):
Brilliant, and you've talked about giving people a
steering wheel for AI.
What do you mean by that?
Why is control important?
Speaker 2 (05:30):
Yeah, great question.
I think there's two main ideasthere.
One is, I think there's avision of AI where the AI kind
of replaces you, in terms of youdelegate a task and just the AI
go and does it.
And then there's a vision forAI where the AI augments you, so
instead of replacing you, itjust gives you superpowers, it
(05:51):
lets you do the task way betterthan you would have been able to
do it without AI.
And that's the vision of AIproducts that we believe in,
that we're excited about, thatwe're building towards products
that we believe in, that we'reexcited about, that we're
building towards um and and tobe able to do that effectively,
you need to give people theright controls and how to steer
(06:12):
the ai in the way you want.
And I think we are in the we'rein like the caveman days of
that in as an industry.
Or maybe a better metaphorwould be um, like like under
Kirk Pothy, I think, used thisthe other day where he said like
chatting with AI today feelslike using the terminal, felt in
like the seventies or earlyeighties, right Before like the
(06:35):
graphical user interface camealong.
And I think as an industry,we've been giving this
incredible technology and westill haven't figured out the
most human uh way to let peopleinteract and make the most of it
, and it's one of the things wewe spent a lot of time working
on at granola.
Speaker 1 (06:54):
Brilliant.
So many of us in business orenterprises are using team
collaboration tools and haveSlack or Zoom or Workspace etc.
Etc.
And you've come out with teamfeatures.
I guess you call it Granola 2.0.
What's changed?
How does that work and how arebusinesses using your service?
Speaker 2 (07:16):
Yeah, when we launched Granola, it was very
much a tool just for you as anindividual and you're in control
and the notes are bespoke foryou.
For example, if you use Granolaand I use Granola on the same
meeting, granola would generatecompletely different notes for
you and for me, because I wouldthink about what's most
(07:37):
important for me to take awayfrom the meeting versus what's
most important for you to takeaway from the meeting.
But the thing about meetings isthat oftentimes information
shared in meetings is inherentlysocial and inherently relevant
to, like, a group of yourcoworkers, of the people you
work with, and we kept hearingfrom users that they'd take the
(07:57):
granola notes and they'd shareit with their team over slack.
That works fine, um, but themoment you so that like
something that's really powerfulin granola is if you take a set
of meetings and you startchatting with our ai across the
set of meetings, like granola,like the underlying models are
really good at pulling outthemes or ideas across that set
(08:18):
of meetings and, uh, well, youcould, you couldn't, you
couldn't do that in Granola 1.0.
So we basically built theability to share a folder, Like
you can create a folder whereyou can put a bunch of meetings
or files, whatever context isrelevant to a specific project.
Let's say and share that withinyour company and then everybody
can chat with that set ofmeetings.
(08:38):
I guess an example for us wouldbe every sales meeting someone
like Granola does goes into oursales folder that's public to
the whole company and you cankind of go in there and be like,
okay, what are customers askingfor this week that they weren't
asking for last week?
Or whatever question you mayhave.
Speaker 1 (08:56):
Very cool, so I use a number of transcription, um,
and they use bots to joinmeetings on my behalf.
I recently showed up to ameeting and there were more bots
sitting.
People, yeah, and people it wasa little strange.
Speaker 2 (09:17):
You don't use bots.
Why?
Two reasons.
One is we want, like, granolais a tool, and great tools are
very consistent in the way theywork and they always work.
So I want Granola to be as easyand consistent to use as a
pencil and a pad of paper, right, and when you have a pencil and
(09:40):
a pad of paper, you don't thinkto yourself, oh, is this
meeting on Zoom and thereforecan my bot join Zoom?
But it doesn't work.
If it's, you know, on a Slackhuddle or you know, an iMessage
call, or maybe it's just anin-person meeting, and there you
know, there's no.
Basically, I just wantedGranola to always work, and so
(10:00):
what that means is we don't.
There's no bot that joins themeeting.
Granola uses your microphoneand it listens to the audio on
your computer, which means youcan have an in-person meeting
right, there's no Zoom link oranything like that and use
Granola and it'll still work thesame.
So that's why we originallydesigned it that way.
Uh, we found that, um, bots,bots, just they're in that
(10:25):
uncanny valley, I think oftechnology, like you're, you're,
you know if you and I arehaving a one-on-one conversation
over zoom, but now there's likethis black square on the screen
and it's neither providing anyvalue in the moment, but it's
kind of making you feel likeyou're you're watched and
recorded.
Um, it just changes the, thedynamic of the conversation.
(10:46):
Also, the, the ai note takersthat had bots that, as far as I
know, most of them record theaudio and the video and they
store the audio and the videofrom the, from the calls, which
granola does not do.
Granola does not store theaudio, it only generates notes.
Yeah, and that's by design,because I think what people want
most of the time is they justwant to have great notes, like
the equivalent of if you justsat there and were furiously
(11:09):
typing notes through the wholemeeting, like that's kind of
what you want.
The audio that could be usefulfor sure, but again, that's an
invasive act to record audio.
That's like a different.
It changes the dynamic of theconversation.
Speaker 1 (11:24):
Yeah, well done.
I guess the other concern mightbe that you know note takers
that are AI driven will justmake us less engaged, or you
know tendency to zone out.
How do you think about that,balancing, you know, ai and
automation with keeping peopleengaged and present?
Speaker 2 (11:43):
Yeah, I think it's.
I think it's a really importantquestion and I think I think
our competencies as humans willwill change here.
I'll give you a great example.
So, um, uh, there are a fewcities I lived in before Google
Maps came out on a cell phone.
(12:04):
I'm old enough where I canremember this.
You know there's a life prebeing able to look up where the
heck you are at any moment onyour phone and post that and the
cities I lived in before thatyou could.
You could drop me in thosecities today and I'd probably be
able to navigate them fine,because I had to really learn my
way around those cities and thecities I lived in after Google
Maps on your phone came out.
(12:24):
I don't have nearly as strongof a mental model of the city
because whenever I was lost Icould just look it up on my
phone.
So I do think there's like anthere's an atrophy of skill
whenever a new technology kindof comes out, and I do think
that will happen with AI.
I think the important questionis are you outsourcing the busy
(12:48):
work or are you outsourcing thethinking right?
And I do think the way wedesign AI tools have a big
impact on this.
Speaker 1 (12:57):
And.
Speaker 2 (12:57):
I think there's a real danger that we design AI
tools where people by default,start outsourcing their thinking
to AI, and I think that's it'sprobably not the future society
or vision that we want to livein.
Right and sure, if you'reoutsourcing busy work, like if
it's, for example, maybe todaypeople are really good at typing
very quickly during a meetingright, because that's just been
(13:19):
a necessity to be able to takenotes, and now in the future,
people aren't able to type sowell during a meeting because
they don't need to, becausethere are things like granola
that take notes.
That's probably a fine tradeoff, right, because now people can
be more present.
If people stop paying attentionwhen they're talking to each
other, right, or thinkingcritically about the decision
(13:40):
because they think AI will catchit for them, I think that's
really problematic.
Speaker 1 (13:45):
Yeah, what a balancing act.
Speaker 2 (13:48):
And just one point on that like we did design and
again, this is just one step,but like a big difference
between Granola and other notetakers is when we launched is
that Granola is still a notepad,it's still a place where you
write your notes, so and andwhat the ai does is that it
takes your notes and makes thembetter.
So we're designing the product.
(14:08):
So it's still a place, it'sstill a tool for thought for you
, so you're still supposed toput your thoughts in there and
the ai adjusts and um andimproves what you do, as opposed
to some of these othernote-taking services where it's
just the AI dissolving right andyou just kind of trust that the
AI gets it right, and then yououtsource it and you don't think
about it.
(14:29):
And I do think that small detail.
Actually, it really changes theway Granola feels and it really
changes the way the notes thatGranola generates feel.
Speaker 1 (14:38):
Brilliant.
The way the notes that thatkernel generates feel brilliant.
Speaking of other ai notetakers, you have lots of
competition which has pluses andand minuses.
Other note takers you know bigtech companies.
What do you see as yourcompetition now?
In the future, what does thelandscape look like?
Speaker 2 (14:56):
yeah, when.
So when we, when we entered themeeting note-taking landscape,
it was already incrediblysaturated.
Companies like otter andfireflies I think they're eight
or nine years old at this point.
We've been at it for a long time, and there are lots and lots of
other companies out there doingthat, as well as the googles,
the zooms and the microsoft's.
(15:17):
Um, that's not when I thinkabout competition.
That's not what I, what I thinkabout.
I think about so that, and toexplain that, I probably just
need to talk about the future ofgranola a little bit.
So earlier I mentioned thatwith granola, we're starting off
with meeting notes, but ourvision had always been to be
(15:39):
something much larger, which isthe place where you do work with
the help of ai.
So we want Granola to be theplace that has a lot of context
about you knows about all thedifferent things that you're
doing, and we have that fromyour meetings.
And then now, how can we use AIto help you actually do your
work and your thinking?
And I think that's the futureof Granola.
(16:01):
And I think when I think aboutcompetition, I think about who's
, who's going to be buildingthat, or who's gonna building
tooling towards in thatdirection.
Um, and from that lens, I thinkit's much more the anthropics
and the open ais of the world,uh, that are going to be like a
longer term competition for us,or or a brand new startup that
(16:24):
hasn't launched yet.
Right, the space is moving soquickly, um and and you see,
like these really innovativestartups pop up, so like those
are the companies that I kind ofthink about because I think
about the future vision.
I'm less worried aboutnote-taking, specifically, I
think I think yeah, I know it'sgoing to be commoditized, Like I
(16:48):
think we have the best notesright now and hopefully that
continues, but I think the thingthat's really going to be
powerful for people is thisfuture tool that lets you do
work five times better and fivetimes faster than you'd be able
to do without it.
Speaker 1 (16:57):
Wow, very thoughtful.
You've integrated with many ofthe top AI models.
How do you decide which AImodel to use and do you see
yourself building your own modelin the future or leveraging
open source or some of theseamazing emerging?
Speaker 2 (17:12):
Yeah, we, we use whatever is the best model for
our use case at that time wealways try to use so, so one is
the best model at any, giventhat changes frequently right
within a matter of months.
Sometimes you have one of thebig labs launch one model and we
switch to that and then anotherlab launches another model, so
(17:34):
we switch to that and we alwaystry to use the highest end uh
model.
So if you're using granola uh,you should be getting the the
best possible premium experience, like on the market.
That's possible today.
Um, in terms of uh doing ourown model, it's possible.
We'll do it in the future.
The pace of innovation of thefoundation model layer has just
(17:57):
been so incredible over the lastfew years.
I expect that to continue.
It's possible at some somepoint that fine-tuning our own
model on.
We have a lot of, obviously wehave a lot of user data on like
what are good notes, what arenot good notes, like what's
actually the most importantthing for meetings, so like we
could do that in the future.
We found that that hasn't been,that hasn't been like the best
(18:20):
use of our time.
So far We've been kind ofletting the models get better
and therefore, granola gettingbetter and we're spending our
time on the user experience, thedesign, the actual product
layer, because there's a lot todo there.
Speaker 1 (18:32):
A lot to do, indeed.
And for other founders outthere, builders just getting
started or thinking of gettingstarted in the AI space, what's
your advice for them?
You've seen it all.
At this point, what would yougive?
Them for guidance or persuasion.
Speaker 2 (18:52):
Yeah, I think this space is changing faster than
any other space that I canremember since I've been alive,
(19:14):
and I think that has a fewimplications.
I think one um, a lot of thehistorical advice or best
practices might not hold true inthis.
In this, like the laws it'slike the laws of physics have
kind of changed.
We're still all trying tofigure out the new laws.
So if you're a new builder,that's a great, that's an
advantage in a way, becausewe're all kind of trying to
figure it out together.
So I think that's anopportunity.
And because the space is movingso quickly, once you launch a
(19:45):
product, once you're in the ratrace, you just kind of have to
be sprinting just to keep upright.
It's like, oh, now this companyhas launched this feature, now
you have to launch it.
So everyone's sprinting.
But when you're sprinting youdon't have a whole lot of time
to really explore and thinkdeeply about a problem.
So if you're like starting offright now, you might feel like,
(20:06):
oh no, I'm being left behind.
I need to immediately justchoose something, work on it.
I would say, actually take abreath and maybe choose a
problem you care about andcherish that time and just
really go deep and really try tounderstand it well and build
your intuitions about what'simportant in that problem area
(20:26):
or not, because once you launcha product, you probably won't
have time to do that and there'sso much value in being able to
have that deep thought on atopic.
So that'd be my second piece ofadvice.
Speaker 1 (20:39):
Wonderful advice.
So what are you looking forwardto the rest of the summer?
What's on your radar?
Any travel events or otherplans heading into the fall?
Speaker 2 (20:51):
So we're based in London and summer in London is
like a wonderful time of year.
It's kind of like that yeah,it's the time of year, the
winter I wouldn't recommend somuch much, but the summer is
really fantastic.
So, uh, no, we just have a, wehave a great set of.
It's just a fun time.
Like we have an office, we havea rooftop, like the team goes
out there, we eat lunch there,we do happy hours there, um, so,
(21:15):
yeah, just lots to build, um,and yeah, we'll be here
wonderful.
Speaker 1 (21:20):
well, thanks so much for joining.
Can't wait to try granola.
Personally, maybe ditch mycurrent transcription service
and onwards and upwardsCongratulations on all the
success.
Thank you so much.
Have a great day and thankseveryone for listening, watching
.
Be sure to check out our new TVshow, tech Impact TV, now on
Bloomberg and Fox Business.
(21:41):
Thanks everyone.
Thanks, chris.
Popular Podcasts
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.