00:00:00
You're on a bit of a heater, huh? >> It's It's the best time to be building >> and revenue has surged, but you face
00:00:10
really intense competition. Let's go right at that to start. >> If you were building a global financial
00:00:17
system from First Principles today, you wouldn't build it on 50-year-old legacy rails. You'd build Airwalks, one AI
00:00:24
native platform for global accounts, cards, and payments. is designed to make the entire world feel like a local
00:00:29
market. Others are bolting AI onto broken infrastructure, but Airwall was built for the intelligent era from day
00:00:36
one. Stop paying the legacy tax and start building the future at airwalls.com/allin.
00:00:41
Airwalls built for the future. >> 350 million in what 2 or 3 years and I'm hearing numbers five or 600 million now.
00:00:51
Tell us about the revenue ramp of the company from the moment you released the software to today. the the product's
00:00:57
been in market for 40 months, 50 months, you tell me. >> Spot on. We started company 2022. First
00:01:03
year was all about building the the research and the product to really kickstart the work. We built the first
00:01:10
texttospech model that finally could sound human. Released it in 2023, beginning of 2023. Then it took us
00:01:17
roughly 20 months to get to the first 100 million in ARR. Roughly 10 months to get to 200, 5 months to get to 300. Um
00:01:25
and um and that's how we closed end of the last year and now we are at 600. >> You're at $600 million in revenue. This
00:01:34
is just extraordinary. Um how many employees now? Because the company's obviously hit incredible valuations,
00:01:42
but you have to fill in that valuation and you you're competing at a very high level for talent. So, so tell us about
00:01:49
how many employees you have now and how you maintain the culture of the company when revenue is ripping, investors are
00:01:56
throwing money at you, showing up at your doorstep. I mean, quite literally. Um, but you've got to run the company.
00:02:03
You've got to build a culture. So, how many employees now and how are you dealing with these competing
00:02:10
uh priorities? Yeah, that's the that's the key element of how you all for us the the element of
00:02:16
like how we can maintain the culture despite the quick growth is is is kind of critical and how we optimize both the
00:02:22
interview cycle how we are bringing people on board how we on board them with 600 people today um so also very
00:02:28
quick growth on that [snorts] people's side and as a company we combine research and product so we we we are
00:02:34
building a communication platform for AI on the research side this includes everything across audio generating
00:02:40
speech transcribing speech, orchestrating speech for interactions on the product. This is how we can complete
00:02:45
the entirety of the customer journey from marketing and creating assets and localizing them internationally through
00:02:51
customer support with voice agents to proactive enablement of how voice agents can help in operations, training and
00:02:58
sales. So this requires a lot of different talent um and and and and a part of that revenue growth is actually
00:03:04
reflection of the functions we've grown over time. So from the original team very research very engineering heavy
00:03:11
from the first 10 people we had zero attrition everybody is uh uh still at the company from those core research and
00:03:17
engineering talent building together with us so so far been able to out compete and I think the common
00:03:23
Fred and and then credit to my co-founder who is incredible researcher himself we've been able to assemble the
00:03:29
team that is truly excited about solving audio solving interaction and building that research and if they if they are
00:03:36
looking for an opportunity out there and looking for a company to join and solve
00:03:40
that we are we are one of the the the leading if not the leading place to do that
00:03:45
>> and you started before AI was so impactful at making software >> right >> so when you were starting four years ago
00:03:54
5 years ago and working on this building software was limited to low percentage uh of the population of planet earth you
00:04:03
know the number of people could write code and now here we are, you know, went from vibe, we had a no code moment, then
00:04:09
vibe coding, and now we actually have people building production code who are not developers. You have developers
00:04:16
going 10x and token maxing. How has building software changed internally? And how do you deal with
00:04:24
making sure that the code is really high quality? because people are paying you this money
00:04:31
>> uh but they're going to demand really high quality product since they're spending so much money with you.
00:04:36
>> Yeah, that's it's also true that 2022 was still the year where topics of the day were crypto and um and metaverse. So
00:04:42
with the building there was also the best time to start because we could actually take a take a bit of time to
00:04:48
focus on what we thought is the future. Um but the the way we are structured is a lot of small teams especially across
00:04:55
the product engineering but also in how we think about go to market optimized for specific industries telco financial
00:05:01
services healthcare. So every unit is very tightly knit together. Um and we do that across the company. Uh so it's
00:05:08
usually five to 10 people teams that that that run ahead. And inside of each of those teams the decision we took
00:05:14
which is slightly different than how it's usually structured. We embedded engineers in in in every place and even
00:05:20
in the places which aren't engineering. So our talent team will have an engineer, our legal team will have an
00:05:25
engineer, our revenue engineering or go to market engineering have engineers embedded all across and those people
00:05:32
will have two roles. One is of course creating automations and bringing the software inside of that team, but second
00:05:39
is actually helping everybody else do what you said, which is make sure that people are adopting AI, but also there's
00:05:46
a security check for everything they deploy because ultimately if if you're not using a lot of the coding software,
00:05:51
a lot of the co-working software, then you're probably in the wrong spot. If you're using too much of it, that is
00:05:57
also a flag because you maybe are not doing that in the right way. Um and of course as you start bringing that into
00:06:04
the sites of the organizations that never were exposed they frequently can create but not necessarily review
00:06:10
whether that's actually doing behind the scenes all the secure ways or anything.
00:06:14
So that's an essential role in the in the in the company. >> Yeah. We the it's fantastic that
00:06:19
everyone can build software until you put it into production and you have a leak.
00:06:24
>> Yeah. uh or that person leaves the company and people forget they built that software and it's just deprecating
00:06:30
uh on its own. The other thing that seems to have changed is management. When you had 10 developers in your pod
00:06:38
or six, you had a UX designer, you might have a pure graphic designer, you'd have
00:06:43
a product manager, they rolled up and then suddenly, you know, we watched over the past 3 years. Oh, hey, this is
00:06:51
pretty good at um summarizing what happened on the call. Oh, it is actually creating action items and it's telling
00:06:58
us what to do next. Oh, and it's, you know, doing uh all the different stories in our conbon board. And now, how do you
00:07:05
think about product managers and management as the CEO and as the co-founder? >> Yeah, we we So, we
00:07:12
>> you fired them all, right? >> We don't we don't have any PMs, >> right? >> Did you ever or did you have you never
00:07:17
did? Okay. >> Never did. thought it's it's a little bit of what you mentioned was also
00:07:22
before the true AI impact started which was ideal ideal person in that role can code can understand the customer can
00:07:30
understand design of course that's very hard to find there's no truly that many people that are expert in any all all of
00:07:36
those fields at the same time so we optimize for profiles that are experts in at least one of those fields but
00:07:41
understand at least one other field really well to your point what we are seeing now there's if you can do a
00:07:48
little bit of all with AI you can maybe step change from being an amateur to being a advanced level maybe not an
00:07:56
expert level so suddenly you are not bottlenecked on all the other functions to do your work in u in growth
00:08:02
phenomenal part growth engineering a person can design experiment ship an experiment it's working and bring it
00:08:07
back um we also have the privilege where we are using a lot of our product ourselves so to be able to do that
00:08:14
ultimately to help everybody else create voice agents we ourselves need to create
00:08:18
voice agents too. Um so we are seeing that also in the non-traditional functions even in go to market like you
00:08:24
need to be able to create a version of that if we are offering that to the customers too and and we do we created
00:08:30
our inbound AI SDR agent that in addition to the form that you fill on the website you have an agent that you
00:08:39
can call >> and uh and people are um of course e can give all the information in a much
00:08:45
easier and quicker way but the second thing that happen is people also leave a lot more information so you can get
00:08:50
connected to the right problem and right person a lot a lot quicker. Um so we are
00:08:54
seeing that kind of phenomena all the time where actually using a lot of tooling makes you yourself better in
00:09:00
your job overall and in 11lapse in our specific tooling that we are solving for customers.
00:09:04
>> Yeah, it seems like the use case of calling on the phone and talking to a computer or previously going through
00:09:12
voice jail and and it was incredibly arduous and painful and annoying. made you just say operator and hit the
00:09:20
zero button like as fast as possible. But now it seems to have turned a corner where
00:09:25
>> talking to a human you I almost feel bad talking to a human where I'm like I am
00:09:29
so sorry I'm wasting your time with this and the AI is just so much more precise
00:09:34
and the fidelity is so great that when you tell them what you're looking to do and you cut them off you don't feel bad
00:09:40
you don't have to make small talk. Is that what you're seeing in your customer base in terms of the ability in real
00:09:47
time to interrupt the agent to interrupt the uh you know conversation and just move faster has made consumers and
00:09:56
companies basically embrace the technology. >> Yeah, it's it's slowly becoming that you
00:10:00
will be asking for a give me an AI agent effective call AI operator. But uh we we
00:10:05
are seeing a transition. We're suddenly and that's you know the the the biggest fuel of the recent growth for us is
00:10:11
enterprises sales team just doing incredible work but then finally the product combines the reliability that's
00:10:19
core with the orchestration for a lot of the AI models but also the knowledge and the
00:10:26
integrations to provide you the right experience. And yeah, I think it it was a step change in the last 12 months and
00:10:31
especially in the last six of of how good that experience became where it's like this golden era of consu for for
00:10:36
the consumers out there customers and and customers is coming where you're going to actually open a website um um
00:10:44
call an agent and have um the agent have information from your past interactions
00:10:48
and deliver that help. And I think we'll see this kind of interesting phenomena combining your previous question and
00:10:52
this where now of course you are reaching frequently when you have a problem and you're asking for help but
00:10:58
ultimately a the whole interface will change and morph depending on how you are operating with that inter interface
00:11:03
with voice being helping you in the background find that information will shift from reactive to proactive to help
00:11:09
you get uh that help before you potentially ask for it. Um and we are seeing those examples those examples too
00:11:15
>> seemed to me that speech to text um had a major blocker again in fidelity 10 years ago lawyers would put on dragon
00:11:24
dictate if you remember that terrible software they get a headset >> and it seemed like the the big blocker
00:11:31
was you felt like an idiot talking to a computer in an office, right? And so people who did it quietly in their
00:11:39
office, you know, they they kind of got away with it. Um, but now we see something very different. The whisper in
00:11:45
the office. People very quietly talking to their computer, giving it a prompt, you know, and talking to their agents.
00:11:52
And now there's a ring out. You can press it. And uh I use a really cool product called Whisper Flow. I don't
00:11:58
know if they use 11 Labs on the back end. I >> they they use they use and and a few
00:12:03
others uh as well. And they they are doing phenomenal work too. >> Whisper flow is just a tremendous
00:12:07
product. And then I got a pedal. Does anybody here use a pedal on their computer? Raise your hand if you're a
00:12:14
com There's one dork, two dorks. [laughter] Any others? Raise it high. Oh, she's
00:12:21
half dork. Okay, so there's about three and a half dorks here. Next year, this is going to be Do you have a pedal?
00:12:27
>> I don't. I I have I >> Have you considered a pedal? >> I I should consider a pedal. I love the
00:12:32
devices that you can wear and the >> I have the Plaude. It's incredible. >> Plaude Pocket. Phenomenal. Like so good.
00:12:38
And especially in events like this, I feel >> if you pre pre preempted that you're
00:12:44
recording, of course, >> but how incredible would it be that all the signal on the conversations that
00:12:49
otherwise disappear, you maybe tap tap few notes here and there to try to get signal afterwards if you can just have
00:12:54
that automatically fill your specific notes and make sure you do your follow-ups. Phenomenal.
00:12:58
>> All right, so let me make the case for the pedal. Okay, I have a I have three
00:13:04
pedals under the desk and I I think I'm trying to figure out what the company is, but with Whisper Flow, you press
00:13:11
down, it turns on, and uh you talk and then you let it go. And one of the annoying parts of working with an LLM is
00:13:20
typing. And you're kind of like exhausted when you're giving it the prompt. So, you stop prompting. But if
00:13:26
you're a professional artist like me and a talker, this is like incredible because when I press the
00:13:32
pedal down, I just give a stream of consciousness now. And it turns out what these LLMs actually do really well with
00:13:39
is taking a massive stream of consciousness where you just keep talking and talking and talking. So I'll
00:13:45
give it a one to twominut prompt. Then I let go and it has changed everything. Everything. It's it's you know like the
00:13:56
whole experience is changing so much a similar version of what we see happen is you know how you have a you want to say
00:14:03
a thought and then you're like okay I actually want to change and say something else
00:14:07
>> now you have those two context combined and the experience you get as an answer
00:14:11
is so much better uh so we already see that as an experience but even the the previous example of like people are
00:14:17
adjusting how they speak to AI versus how they speak to human people are asking >> how so yeah what what how should you
00:14:23
speak to the LLM. We saw Sergey Brin say threaten it with bodily harm. It's a very effective technique if you haven't
00:14:30
tried it, but what are the things that are different when you're talking to the LLM?
00:14:34
>> The the the a specific emotional example, we we work with a lot of um financial services companies uh uh uh
00:14:42
Revolute, Clara, Pogbank. And some of the frequent case not in all of them is of course how you remind people about
00:14:48
payment or that you collect the de that from the people that aren't answering and frequently people would naturally
00:14:57
feel ashamed of telling the real situation with AI people are much more open to share what actually happened
00:15:04
give the information and suddenly this emotional block of like in front of other human I don't want to be able to
00:15:08
to say all of that is is is very different so that's different um usually people are more snappy with AI voice
00:15:16
agent. It's like, you know, like quick responses. >> Yeah. You don't mind cutting it off.
00:15:20
>> Exactly. So, you can like kind of go through to to the point you want much quicker. Um, which you needed to like
00:15:26
change a little bit of the interaction model too, which which is working. Uh, but we'll work on the pedal and whether
00:15:31
whether we should we should do >> a little bit about celebrities on the platform. You have some celebrities who
00:15:39
are on there. Um, you also have um an issue with impersonation. Um, I know this because somebody was
00:15:48
like, "Oh my god, I love your bulldog videos." Many people know I'm a big fan of bulldogs. I currently have three. Um,
00:15:55
and I said, "I'm sorry, I don't know what you're talking about." And they sent me a channel where somebody had
00:16:01
created a bunch of dogs telling jokes and they made one and I guess they were looking for a podcaster. So they used
00:16:08
the this weekend startups archive and 11 lab to create my voice and do this huge
00:16:15
channel. And I contacted them and I said, "Oh my god, it's very flattering. How how did you do this? This is like a
00:16:20
year or two ago." And they said, "Oh, I used 11 Labs." So I I think I emailed you about it. I'm like, "H how do you
00:16:25
protect against this in advertising in the law in the United States? I'm not sure about here in France. I'm sure they
00:16:33
have 17 laws for this. We have one. Um, you guys are great at regulations unless
00:16:38
no offense. And the French guy over here is like, "Oh, Mond shal um the that's my French angry developer
00:16:48
guy." Um, I cannot smoke in the Lou. Um, [laughter] this is crazy. Um, and so [laughter]
00:16:58
is super like interesting. um with this right to privacy and I think you've got a quick education on this because you've
00:17:06
had a couple people I'm sure write you a legal letter what it basically means is
00:17:10
you you can't take somebody's voice and use it to you know do commerce in the world you can use it for parody there is
00:17:17
fair use I can do a Donald Trump impersonation up here if I like we're going to take about 5% of 11 lab
00:17:25
stock is okay with you put them in Trump accounts [laughter] sounds good okay and for that you get to
00:17:30
come to the White Okay, thank you. Um, [applause] nasty guy. Wouldn't give 5%. Loves socialism, but not America. It's
00:17:40
the problem with the Nordics. Um, nasty, nasty socialism. Then I noticed when my guys wanted to clone my voice so
00:17:49
that they could fix the ads where I mispronounce something or I do the wrong promo code, use the code jal 20. They
00:17:56
like were like, "It's 25, dummy." And I'm like, "Okay, I have dyslexia." and then they redid it and it was like I'm
00:18:01
sorry you cannot uh clone Jason's voice and then it's like I have to go in there
00:18:06
and do it and you put a bunch of protections in there. So explain what's happening in that regard in terms of
00:18:13
people's you know concerns around this and then the other side which is the opportunity because I think you got
00:18:18
Jamie Fox and some other folks actually that you paid for their voices. >> Yeah. No, the the the the voice is
00:18:26
identity and IP. It's like you know when you when you speak a certain way people
00:18:30
recognize it can feel that emotion and you know to some extent it was it it was a a um it could be a problem could be
00:18:38
opportunity before I mean as you did impersonation of of of the President Trump it's of course similarly uh uh u
00:18:46
something that is possible even of a human uh not specifically AI but um for us on the safeguard side you know over
00:18:55
over last years We took the role as we are leading on our development development. We also need to lead on a
00:19:00
lot of the safeguards. So that's like a critical element. We do three things. One, trace everything that's generated
00:19:05
so we can take action when needed. Two, now we moderate both on the voice and text level. So if you were to input
00:19:11
something that would be commercial in nature or uh would try to scam someone that gets flagged, we can block it. And
00:19:17
now free because over last years we've seen the development of those models more broadly. how how can we create
00:19:23
systems for the wider world so people can upload a sample and get information whether it's AI or not immediately um
00:19:30
and we do it for 11 Labs but we also do it for other open source models the interesting part given that it's such a
00:19:36
good um IP um and and part of your your your your element it opens up new opportunities so we partnered with
00:19:43
Matthew McConna on creating a world >> all right all right >> and across languages and it's the first
00:19:48
>> I haven't got paid a lot of money for these independent films but oh 11 lab stock is juicy.
00:19:53
>> The >> yum yum. >> Could you do it? Could you do it in Spanish? >> It's a it's it's a fugazia fugazi.
00:20:00
>> Yeah. But the the crazy thing with the with AI technology open is that now the
00:20:04
voice can be carried not only English but also in Spanish and Italian and Portuguese and you can still have
00:20:10
exactly that element of emotions coming through. Um so that that's kind of a a good example there but we've seen that
00:20:17
with master. >> What do you pay these guys? What does it cost to get Matthew McConn? Is this like
00:20:20
an eight figure deal, seven figure deal? You give him a little equity? >> Always depends. Uh so like you know the
00:20:26
masterclass for example is a good example where they worked with talent directly. And here you have u previously
00:20:32
a static content that you would learn from. Now um you have interactive content. So you have Gordon Ramsey
00:20:38
teaching you how to how to cook in the kitchen. He can scream at you if you're not doing
00:20:43
>> raw scallops raw. So, so, so that is that is definitely a >> So, they're doing characters now or or
00:20:50
AI instances using 11 Labs. So, you can interact with them as part of your subscription.
00:20:55
>> Exactly. But as a company, what we now do and this from the beginning, we created a marketplace where people can
00:21:00
create their voice. We authenticated, you can share it, and you earn money. Today, we paid back over $22 million
00:21:06
back to the community of of of talent. >> Really? uh which so those voiceover actors now who got paid as hourly
00:21:15
workers sometimes they get a little backend if they were doing a commercial or something now they can spend an hour
00:21:21
reading create an 11 labs voice and then license it out >> 100% and then >> do they get to pick their price or you
00:21:27
pick the price >> the depends on the model we do we do both so you can you can either give it a
00:21:32
default that lets us distribute that slightly more optimally or you can pick yours and the use case is going to be
00:21:37
going to be different and like you opens up a set of incredible opportunities in
00:21:41
the dynamic context in other languages. Uh but maybe a last one on that like voice is such a big part of identity and
00:21:48
probably probably our most important work was actually working with people that lost their voice due to ALS due to
00:21:55
throat cancer and working on bringing that voice back. So he worked with congresswoman in the US Jennifer
00:22:00
Jennifer Wexon who lost it and wanted to continue inspire others that you can do
00:22:05
incredible work despite that and uh and was the first speech delivered in in in in in Congress or more recently I think
00:22:13
this was my the most uh like heart uh warming story. There was um there's a woman that um that's wanted to get
00:22:20
married lost her voice before she could get married. >> Oh wow. >> And then they decided to redo the
00:22:26
marriage together. do the vows again >> and do the vows >> and you could see the whole family just
00:22:31
>> for the first time hearing the vows. It was just uh you could you could feel the
00:22:36
emotions that you can see in any other any other way because the voice is such a connecting thing.
00:22:41
>> Yeah. And and you've done it for some iconic voices. My understanding is the estate of James Earl Jones. I'm not sure
00:22:48
if they did he pass? Is James Earl Jones alive? Can somebody pass? >> He he passed, right?
00:22:52
>> Yes. But before he passed, I think he did a deal with Disney and he said, "Listen, for my family, I would like to
00:23:02
license the Darth Vader voice for all time to Disney." They gave him some incredible deal. And then they were left
00:23:09
with, "Well, how do we actually do this? Do we get a voice impersonator, but instead they went to you?" Talk a little
00:23:14
bit about that deal and how it went down. And is that what they used recently, you know, in in uh some of the
00:23:20
new films with Darth Vader? There's a new Darth Maul series um where they have Darth Vader and did
00:23:26
you power that? >> I don't know what I can say about the new things but definitely the big use
00:23:31
case that that big big completely new experience [clears throat] uh was in the gaming space where yes
00:23:37
>> uh Fortnite so Epic Games game Fortnite launched um Darth Vader which people and players could interact
00:23:46
with live in partnership with the state in partnership with Disney. So every player after reaching a certain stage
00:23:52
could have a Darve interact and help you solve the missions. And we are seeing that kind of mode coming up more and
00:23:59
more often of how you can effectively extend extend your likeness your like you said publicity into interactive use
00:24:06
cases bring it across the world uh together. Uh so that was exactly that model and now we are working on on a on
00:24:13
a one of the public one is headsp space. So Headspace has a great meditation. This is the uh second uh greatest
00:24:20
meditation app right behind Comm >> which you are an investor of. I am I didn't realize you're right. I did
00:24:26
[laughter] and it was a $4 million company >> is incredible. I think their their team
00:24:31
>> but anyway you were you were working with the the um the second place. >> Exactly. So they not exactly the second
00:24:36
place but exactly to the working part. They so they localize a lot the content and and calm I think is trying some of
00:24:43
the interactive elements. Could you have a meditation lesson that's personalized
00:24:47
to you? >> Which uh which we we would uh hopefully love to to to >> and like imagine just you know so many
00:24:55
voices. >> David Saxs is defending Trump. Take a deep breath in. Breathe out. Breathe in.
00:25:03
Breathe out. >> Maybe you should license the voice to come. >> I I mean that would be interesting.
00:25:08
Let's talk a little bit about um being up against some of the greatest entrepreneurs ever
00:25:17
who want to take your business from you. Specifically, Daario and Anthropic, Sam
00:25:22
from OpenAI. They they want your business. They've been pretty clear about it. Um, and I think you have used
00:25:29
the frontier models in your product, but you must be thinking, my lord, am I enabling my own demise by partnering
00:25:38
with them? And there's all these open- source models. So, um, how do you think about your partnerships with those type
00:25:46
of frontier models and the fact that they want to kill your company? >> [snorts]
00:25:52
>> So on on the on the first part the given we create a platform we try to provide
00:25:58
all alms out there. So our customers can pick entropic open AI open source uh Google models um and that agnostic being
00:26:06
agnostic to the specific model is actually helpful because customers can they make sure that they build a harness
00:26:12
build the agent orchestration create the voice element of how that agent interacts with the world how the
00:26:18
marketing interacts with the world but they are not dependent on any model. So for us that part is is is um is a is
00:26:24
actually good because we can provide that to the customers. On the on the kind of the second big part of like of
00:26:29
course the the space is overlapping increasingly models our platform platform are
00:26:35
application everything is becoming a little bit more more fuzzy for us. The still the defining piece was focusing on
00:26:41
that one layer of like how does interaction look like? How does communication look like? And we've been
00:26:46
able to out compete them on voice models um both on text to speech, speech to text, on the turn taking on music and um
00:26:53
and we've you know here our research team is is is is a a set of magicians that are able to continuously do it time
00:27:01
and time again. Um and I think part of the reason is it's on the research side. It's the architecture that matters, not
00:27:08
the scale. You really need to change how the model operates. Two, you need very specific data that there's of course a
00:27:15
wide set of data out there, but it's unlabelled data and where we spend a lot of time. So we build a internal team of
00:27:20
over thousand contractors that label all those audio assets to make them to make
00:27:24
them good. So that's on the research side. And then as you think about the rest of product stack, we want to create
00:27:28
a fully verticalized solution for that communication angle. The product understanding the right workflow in
00:27:33
financial services is very different to healthcare, very different to telos. We spend all of our product team to figure
00:27:40
out how that works and those companies don't. And then ultimately last piece is the ecosystem. Can you build the wider
00:27:46
set of integrations voices that you use templates for the agent authentication that you can benefit from instead of
00:27:52
starting from scratch? And so far we've been we've been able to create a new model for that.
00:27:56
>> Certainly though you must be concerned about hey the reinforcement learning the
00:28:02
data leakage. They say they're not using your data, but they're kind of using your data. And so, do you have an
00:28:09
open-source project internally as the like in case of Glass, we got to break this? And when do you think
00:28:17
>> you'll be able to discontin working with them if you had to? >> We we we we know that some companies are
00:28:28
continuously trying to figure out how to distill and use the data. So that is uh
00:28:32
that is a existing problem and we have few mechanism to to to stop it. Um or slow it down not stop it. Um but um but
00:28:41
on the open source question or like creating our own um um uh versions we are we are looking a little bit closer
00:28:48
on like how we could use our expertise of how does like you know we won't focus on
00:28:54
knowledge work. we won't focus on coding but any interaction and how you can combine all those pieces together and
00:29:00
make sure this is this is great. Yeah, we want to own. So we are spending more time there. Uh but it's also just great
00:29:06
to be in the arena and compete with those guys and uh and and every so often show that we can do it and do it better.
00:29:12
Yeah, it's pretty clear in my estimation that that's where you you will wind up and the ability to make your own
00:29:21
language model today, especially with all these great models out there that are now open- sourced. It's going to be
00:29:27
a pretty easy for a company with your level of resources. So, why wouldn't you at least offering it as an option? And
00:29:34
then I guess there's cost. I mean, you must be shipping tens of millions of dollars to the
00:29:39
Frontier models every year. ship good amount. Um we are good partners good partners with them. Uh um but but it's
00:29:48
um it's ultimately you know showing up in the value we can create too. So like a lot of what we spoke at the beginning
00:29:55
of how we can elevate ourselves as organization too is is definitely helpful. So I think they they done
00:30:01
tremendous work on on building uh it's it's almost crazy that each of us has like a a a Turing uh like you know if
00:30:08
you if you were to chat with an agent now it feels like um the Turing test will be complete it's as smart as
00:30:16
another human. Uh and we hope this year we'll do that same thing for voice uh where any conversation feels like you
00:30:22
are speaking with another human. It will be >> yeah I think you're there. It just
00:30:25
depends on the application and like what question you ask. But it definitely passes it. I mean it if we were to look
00:30:33
at the tests that were created to define artificial general intelligence or just
00:30:37
to define artificial intelligence, we passed all of those. These were tests that were created 30 or 40 years ago. We
00:30:43
need a new set of tests right now. I think the new test is like can this be more intelligent than every single
00:30:51
person on the planet times 10. And if we get anything less than that, we we're kind of like, oh yeah, it's not smart. I
00:30:57
mean, these things, we're kind of there on AGI, don't you think that we've kind of achieved it, we just haven't deployed
00:31:03
it. I >> I am I there are definitely places where where we did achieve it. >> Yeah, for sure. All right. Continued
00:31:09
success. Let's give it up for Mati from 11 Labs. Well done. Thanks for coming out. I'm going all in.
00:31:17
>> The AI companies building the future run on Oracle cloud infrastructure, training
00:31:21
and deploying at scale on one of the world's largest AI infrastructures. The same Oracle AI platform gives
00:31:27
enterprises access to leading models AI grounded in their own data and the security to move from pilot to
00:31:34
production. Learn more at oracle.com/ai or experience it live at oracle experience live. I'm going all in.
00:31:44
>> You're growing also at a very significant clip. >> Exponentially. >> Is it exponential? No, it's not
00:31:50
exponential. >> Oh, it's a 50 sustained 50% quarter over quarter for the last seven quarters.
00:31:56
>> 50% quarter over quarter. Last seven quarters. Yeah, that's that's pretty darn fast. So I think we actually just
00:32:02
became as of the close last week on Tuesday one of the fastest enterprise company with a direct sales motion to go
00:32:08
from one to 100 150 beating Sierra with one quarter. >> Amazing. Um and so um people I mean
00:32:18
there's a couple of things in life that people really hate and paying lawyers is
00:32:23
like way up on the top of the list. um with your tools obviously you got your contemporary and Harvey and people
00:32:33
>> it's just a it's a small company um in the in the states um and then you also
00:32:37
have I guess Claude and other folks also want to be in your business so this is a
00:32:43
big prize um to take I don't know 80% of what we pay lawyers for and compress it by 90% like
00:32:53
what what is the realistic power law here in terms of making for startups in the audience,
00:33:02
your legal bills dramatically drop in costs. Yeah. And I'm seeing it already in the startup space. I had an I had one
00:33:10
firm, one startup that hit a million in revenue. >> Yeah. >> They had closed multiple rounds of
00:33:16
funding. Um, multiple obviously large number of employees, a decent couple dozen employees. They didn't have a
00:33:22
corporate lawyer. >> No. And I said, "Whoa, whoa, whoa, whoa, whoa, whoa. You think I had a million
00:33:26
dollars in revenue? Like, somebody should review the contracts?" And they're like, "Chat CPT, bro." And I'm
00:33:32
like, "What about the cap table?" They're like, "Chp, bro." And I was like, "Okay." And HR. And they're like,
00:33:38
"Same thing, bro." And I'm like, >> "Okay, >> it's going to be a fun diligence target
00:33:42
one day." >> Well, that's what I said. I said, "Hey, you know, when you do the series A,
00:33:48
they're going to ask that like some of this stuff be reviewed. like, "Do you guys have like IP assignments?" They're
00:33:54
like, "Yeah." I'm like, "How did you know to do IP assignments?" First- time founders are like, "We asked chat GPT."
00:33:58
And I'm like, "Okay, wow. I just turned into unk like I guess." Yeah. [snorts] So, so take us through what you you
00:34:05
think is happening out there. This is not uncommon, right? What I scenario >> but but a a seedstage startup operates
00:34:12
very differently from, you know, one of the biggest banks in the US. And so the way to think about the market or at
00:34:18
least the way that we like to is you have this enormous bucket of legal services which today is being done
00:34:24
manually. It's a trillion dollars every year into legal services which is very fragmented.
00:34:30
>> But the software spend into legal technology is about 40 billion. So it means there's 4% software 96% service
00:34:39
which is bananas. The software piece should be much bigger than that. And [snorts] so the software piece naturally
00:34:45
will grow into um the service revenue but also legal is a very supply constrained market. The demand for legal
00:34:56
services is much larger than what there are lawyers or legal services available.
00:35:02
And so many of the legal service providers are now using technology to serve new use cases, new market segments
00:35:11
and to actually package new products and you will not make >> what's an example of that like
00:35:16
>> so so an example of that is Kulie actually they started serving startup founders directly with a sort of
00:35:25
software uh platform that you just log onto the platform they pumped it full with their material and their precedent
00:35:31
and then you have the startup material there and they've embedded workflows that reviews the contracts.
00:35:37
>> Um, and what I think is interesting by that is it it starts to um break this model where you charge out associates
00:35:46
for very high hourly rates and you have a billable hour model. And actually, if you look in law firms, the way that that
00:35:54
business model works is you overcharge for the associates and you actually undercharge for the partners. I don't
00:36:01
know if they're under charge. I mean, I got a bill recently and it was 1,800 an hour, right? But
00:36:06
>> for a senior person, I think the associates were 800. >> Well, you know, at Kirkland, it can go
00:36:09
up to 4,000 an hour. But the but the thing is when a when a Kirkland if so, let's say, you know, 30 minutes of a
00:36:17
Kirkland partner's time when it really matters can be worth a lot more than that.
00:36:22
>> Like a lot more than that. If it's bet the company litigation or you avoid a pitfall that would have costed the
00:36:28
company tens of millions of dollars, >> well worth it. Yeah. >> Right. Exactly. And but the only way
00:36:32
they know how to price that is to overcharge for the associates. >> But as you're saying, the enterprises
00:36:38
are looking at this and they're going, "Huh, we're spending a lot of dollars on legal services. Let's take this
00:36:43
in-house." And >> Oh, really? >> Absolutely. I mean, we're doing this partly at Lora. We acquired four
00:36:49
businesses so far this year. We did the diligence inhouse the with our own tool and the fastest transaction we did was
00:36:57
12 days from LOI to closing >> because your motivation as the founder is to get the deal done,
00:37:04
>> right? >> The motivation of the lawyer is to not have you sue them if they up the deal,
00:37:10
>> right? >> And to make as much money as possible, >> which means to drag it out,
00:37:15
>> which means their incentive is to even if they don't say it explicitly, it is
00:37:19
to drag it out. your incentive is to close it as quick as possible. Yeah. >> Yeah. And so, you know, I think a lot of
00:37:24
law firms are also experimenting with different pricing models where you do a fixed fee for a transaction or for a
00:37:30
fund raise. Um, in litigation, you can take a part of the success fee when you win the deal.
00:37:36
>> Yeah. >> Uh or win the case. And so, I think it's just very interesting how, you know, one
00:37:41
of the biggest industries in the world now is being completely transformed and reshapen as a consequence of the tech.
00:37:47
And are those law firms feeling like they're being disrupted or this is a huge opportunity
00:37:54
and and and did that switch at a certain point in time or has it switched for them?
00:37:58
>> Um there's a lot of anxiety and a lot of fear and you know these law firms are
00:38:07
enormously profitable and big businesses. Kirkland Ellis turns around $10 billion a year.
00:38:14
>> How many lawyers did it have? >> Four or 5,000. Wow. >> I mean, per partner, they make it
00:38:18
between 5 and 10 million every year in profits. And so, when something like AI comes along, that poses um existential
00:38:29
threats and existential opportunity. And I that's actually a big part of my job to help articulate with the leadership
00:38:37
teams that we work with because we will only be as successful as our customers are. And so [snorts] we actually have a
00:38:43
very unique role at Legora as well which is called the legal engineer. So in the
00:38:48
same way that Palanteer has forward deployed engineers, we have forward deployed lawyers and their job is to sit
00:38:55
down with the Kirkland partners and help them transform their business >> from a preAI to a post AI world. A and
00:39:03
it's sort of like document management and PCs were but 20 or 30 years ago when they were printing out and keeping
00:39:11
drafts in a in a library in in a storage facility and they had to sort of walk them through and handhold that.
00:39:18
>> Absolutely. But I think the difference >> the difference is those were you know
00:39:23
mild productivity gains. >> Yeah. this can do a lot of the work and so it's really reshaping what it also
00:39:32
means to be a junior lawyer going into this occupation. >> What does it mean? Are those jobs going
00:39:38
to still exist or a lot of the lawyers who are coming out of school going, "Oh my god, was this a good idea or a bad
00:39:44
idea?" >> The job will exist. The tasks will be different, >> right? Um, in order to have a
00:39:52
partner-driven model, you need to bring people up the ranks, right? In [snorts] the same way as you do with software
00:39:58
engineers. You need to have junior engineers so that one day you can have senior engineers who know what they're
00:40:03
doing. Um, but the way to get there is very different. The way of getting there today will not be lock yourself in the
00:40:10
physical data room, read through every single document, mark the errors, and you know, go fax it, right? It's and
00:40:19
it's also no longer just look in the virtual data room and control F. It's orchestrating the agent that will be
00:40:25
doing that work. >> And and when you look at that work, you you have a global backdrop
00:40:34
attorneys obviously very famously localized, right? And is this going to create attorneys who can operate across
00:40:43
borders in a way that didn't exist? And you're starting to see that and is that something that's built into the product?
00:40:48
So when you're doing even in the United States, it's a it's state level certification obviously. Um and doing a
00:40:55
non-compete in the Northeast is very different than doing it in California. >> They're not very enforceable or
00:41:01
enforceable at all in California as people know, but they're quite enforceable if you're in Boston.
00:41:06
>> Yeah. >> So, so talk about that. Uh because that seems to be a place where there could be
00:41:11
massive gains from AI. >> 100%. And it it's really two things. I mean the data that Legora sits on top of
00:41:18
is on one hand side the firms and enterprises own data their precedent their organizational data and secondly
00:41:26
we do the hard work of gathering all the cases all the legislation all the regulatory updates for every
00:41:33
jurisdiction in the world and that is very painful but once you start to do that at scale it builds a real data mo
00:41:41
>> and so in the system if you are the GC of a company in California and you just
00:41:47
landed your first customer in South Africa, right? Legora can be adapted to the local legislation in South Africa.
00:41:55
And we actually had a case of this where, you know, instead of having to call a lawyer who then knows a lawyer in
00:42:01
that region who will respond to the query, they can get an 80% accurate response immediately that they can start
00:42:08
working off out of. And the better that gets um the more um the more interesting
00:42:16
things I I I believe you can do because this data has really never been structured before and there are so many
00:42:23
people who are working with setting policy and building regulation and and this is a enormous inefficiency in
00:42:30
society >> and Lexus Nexus has been a juggernaut and the legacy player in you know all
00:42:38
the case law and regulations. They have a massive data moat. They but they only make a couple of billion dollars a year.
00:42:46
And if you put your revenue and Harvey's revenue together, you guys are probably
00:42:51
already just that the two of you uh you're both making hundreds of millions of dollars. So, they must be looking in
00:42:59
the rearview mirror at you like the Tyrannosaurus Rex in Jurassic Park and going, "Holy shit." Like, are they
00:43:06
coming for our business? And then here you are on stage saying, "Hey, we're doing all the manual hard work of
00:43:12
getting that information into our what I assume is a proprietary language model."
00:43:16
We'll get to that in a second. Um, are you going to just try and buy Lexus Nexus? I know it's part of a larger
00:43:22
enterprise, or are you just going to kill it? Well, I think that some of the existing
00:43:29
uh providers and the sort of legacy players have a really hard time pivoting into becoming AI native businesses.
00:43:37
>> Sure. >> And they have a really hard time meeting and and catching up to the tempo that we
00:43:43
run at. They can't get the talent. They don't work our hours. And they're so political in their organizations that
00:43:51
it's just hard to move. Um, I think at the outset of AI, many believed and made a bet that those organizations who had
00:43:59
all the data was going to be the winners. As we're starting to see in the market, that's no longer the case. I
00:44:05
think there's a real opportunity for us to partner with content providers. And we're already doing this in many of the
00:44:13
smaller jurisdictions like in Germany, in France, in Spain. The US is is peculiar because it's such a duopoly on
00:44:21
legal research as a West Law is the other one. >> West law and Lexus Nexus. Exactly. But
00:44:26
um yeah, if you look at how their stock is doing, um I think >> or are they getting priced in with the
00:44:33
AI uh certainty? >> Yeah. >> Yeah, that's one way of putting it. >> Yeah, they're getting crushed. Uh and I
00:44:39
I would assume there's some power law here. you know, they might have an incredible breath of, you know, old case
00:44:47
law that they scanned in and went to the courouses and did all that work on, sent
00:44:52
to India to be double blind typed in. Like they literally >> You're right. That's what you have to
00:44:56
do. >> Yeah. They literally had two different people type in the cases or OCR them,
00:45:01
then check them, look for the differences. I mean, cuz you can't get it wrong. >> Nope.
00:45:06
>> But today with the AI tools, the AI tools are really good at doing what they did manually. Yes. Um, you still have to
00:45:13
ship the books because you have to physically scan. This is very strange in the US, but uh, West Law basically has a
00:45:20
monopoly with the American government to report on the cases. So, they're not owned by the public in a way. They're
00:45:28
owned by >> That's crazy. >> The company. You guys are very good at capitalism. >> Sometimes too good. But sometimes too
00:45:34
good. But I mean, Harvard has a project. There's the court law court listener. >> They're trying. They're trying.
00:45:40
>> Yeah, it's it doesn't work. Um or rather put it this way, you cannot build a legal research solution that doesn't
00:45:49
have all of the data. >> Because if you go to WCTEL and a litigator at WCLE, the best law firm in
00:45:56
the world, says, "I'm going to use this to to, you know, go after Elon or or do a billion dollar case." You better make
00:46:04
sure you have all the cases. >> So So it's the opposite of the power law. You don't just need the top 80%.
00:46:10
You actually need all of it. >> All of it. >> Which means you have to go to courouses
00:46:14
and ask them for a copy to print it out and pay them 10 cents a page. >> Well, there's other ways of getting it,
00:46:20
but in in practice, yes. Um, you have to physically get the books all the way to
00:46:26
India. You need to open them. You need to scan them cuz you need to get what's called page citations. I never thought
00:46:33
in in college I would get this nerdy about legal data, but here we are. And what's interesting is that these
00:46:42
previous generation of databases were very much search in the database, find the case, and then the lawyer, you know,
00:46:49
does their work, >> right? >> What's really interesting about especially the agents following the
00:46:54
release of Opus 4.5 and 4.6 Six is they can now start to do really intelligent case strategy and they can actually
00:47:03
start to combine the witness statements, the cases and they can really do end to
00:47:10
end work which is I think moving us from a world where AI is just augmenting to AI is actually really doing things
00:47:20
and your job becomes to orchestrate and to manage those agents as we're seeing in coding.
00:47:26
And so you have partnerships with I'm assuming anthropic and open AI. Yes. To and you spend millions or tens of
00:47:36
millions of dollars on tokens. >> Absolutely. >> And they are also competing with you on
00:47:41
the margins. >> Um they are not competing in our product category at all at at all. Um from
00:47:48
>> for now the well you know from the outside uh you know Claude has a legal offering. Yeah. which is basically a
00:47:56
bundling of markdown skills files and a couple of integrations. >> And so I think what's really helpful
00:48:03
about that is that it illustrates to everyone how applicable AI is in law. >> Um what it also does is it drives a lot
00:48:12
of initial usage there and then you hit the ceiling or you know you understand how shallow it is and then you call us
00:48:19
right >> and so it's actually a big pipeline generator uh for us. >> Got it. So they start experimenting.
00:48:23
Boom. And we were just talking with the CEO of um 11 Labs about hey building your own models is you know pretty um uh
00:48:36
pretty doable these days and every 6 months it gets easier and easier. So are you working on your own models using
00:48:45
open source to then fork it and and make your own models? Is that the future for
00:48:48
your firm? So um I don't believe in fine-tuning or building any general intelligence models. I think that's uh
00:48:57
total waste of time and money. I do believe in very narrow models for narrow use cases that you also drive a lot of
00:49:05
scaling. So you can drive both cost and latency down. An example of this for us is we have a big feature called tabular
00:49:12
review which is basically the number of documents times the number of prompts. So 100 documents, 100 prompts, 10,000
00:49:19
API calls. >> If you [snorts] make a fine-tuned model at extracting contract data, it's very
00:49:25
applicable there. But it doesn't make sense to build a general legal intelligence model like some of our
00:49:33
competitors are attempting. >> Yeah. Um, and how do you mitigate against the data leakage issue with your
00:49:43
customers? these, you know, are highly regulated industries with a lot at stake.
00:49:50
So, putting in, you know, this recent case you're working on in a litigation, if [snorts] any of that [clears throat]
00:49:56
were to seep into a language model and then come out the other end, I mean, you this is disastrous. You have a higher
00:50:02
level of >> responsibility >> and uh compliance is our currency. And so um it's actually one of the reasons
00:50:12
why it's really hard to sell into law. There's a lot of legal AI companies and very few are making it through
00:50:21
>> and not because it's hard to build stuff. It's actually quite easy to understand where you can build value,
00:50:27
but getting it to the customer is very hard. But that's something we cracked pretty early on. And
00:50:33
>> once you're in, it's much easier to expand. So that's al also one of the driving forces behind our M&A strategy.
00:50:40
[snorts] But yeah, I mean we're hosting national secrets weapons manufacturers with their contracts on Lora and we work
00:50:47
with governments and >> does that mean you have to put it on prem as well or >> we don't do on prem. Um I
00:50:53
>> that's on the road map or >> no I mean you know deploying in a VPC is very time consuming and it creates a lot of
00:51:03
dependencies which slow down your road map and the execution forward. >> All right continuous success max thanks
00:51:10
for taking some time for us. [music] >> [music] [music] >> I'm going all in.