WEBVTT
00:00:00.160 --> 00:00:12.720
This is like an area where we can not only help build the infrastructure for the next way of how the world communication is going to go through and run through, but also really help influence it.
00:00:13.039 --> 00:00:20.399
Welcome to First Commit, a podcast from Notable Capital where we talk to the builders solving hard problems before the industry catches up.
00:00:22.480 --> 00:00:34.719
Today we are joined by Aviv Shamney, co-founder of Lime AI, a platform building the connection layer between brands and AI agents, helping global companies understand, measure, and influence how agents discover, fetch, and convert on their products across the web.
00:00:34.880 --> 00:00:35.679
Welcome, Aviv.
00:00:36.000 --> 00:00:36.799
Hey, how's it going?
00:00:37.039 --> 00:00:37.759
Good to be here.
00:00:38.320 --> 00:00:39.520
Excited to get rolling.
00:00:39.759 --> 00:00:44.799
So you have an extremely impressive background across engineering, sales, and large-scale data systems.
00:00:44.960 --> 00:00:49.119
We would love to touch on kind of how that background informed what you're building today.
00:00:49.520 --> 00:00:52.880
Yeah, so first, really excited to be here, and it's good to meet the both of you.
00:00:53.119 --> 00:01:07.920
Basically, kind of like when when we started, when we saw this whole shift happening with AI, we understood that this is much more than just like a little tweak of yeah, we can now talk to data, but it's actually going to influence essentially how we're even communicating with kind of like the digital world.
00:01:08.239 --> 00:01:25.200
And when we deep dive into really understanding the fundamentals of how these things work, you know, that required essentially, you know, our technical expertise and really taking a very, very close look into how these AI agents are interacting with essentially the internet, with digital assets.
00:01:25.280 --> 00:01:27.359
What is the net what is the new networking layer?
00:01:27.519 --> 00:01:33.920
And so that really got us to a point where we were able to uncover a lot of very interesting things.
00:01:34.159 --> 00:01:37.760
And, you know, some people say that you don't really need technical backgrounds as a co-founder.
00:01:37.920 --> 00:01:44.879
I might agree, but I think that was like a especially now with how everything is changing, that is something that we needed to have in front of us.
00:01:45.120 --> 00:01:45.599
So yeah.
00:01:45.920 --> 00:01:46.239
For sure.
00:01:46.319 --> 00:01:54.000
And I feel like most of the Israeli founders we talk to and work with are operating in the cybersecurity domain, and you're quite literally at the opposite end of that spectrum.
00:01:54.239 --> 00:01:58.239
So what was it about the idea with Limey or where did you see it firsthand?
00:01:58.640 --> 00:02:02.400
Yeah, so we didn't come from, like at least I didn't.
00:02:02.480 --> 00:02:07.200
My CTO did, he's from 8200, also not specifically from the cybersecurity department.
00:02:07.439 --> 00:02:16.319
But I for sure wasn't a part of like the whole cybersecurity clan in Israel, which is also extremely unique and has a beautiful glow and aura behind it.
00:02:16.639 --> 00:02:22.560
But my background actually comes from like a little bit different areas in the army, more special forces.
00:02:22.719 --> 00:02:24.479
So that was a little bit of a different route.
00:02:24.639 --> 00:02:37.280
And I think there's a lot of very interesting things that we are now learning in this new digital world around how AI agents are interacting essentially with digital assets, whether it's websites or whether it's content that is being put out there.
00:02:37.680 --> 00:02:53.360
And the fact that, you know, we were able to kind of like deviate from the classic Israeli story of cybersecurity to something that is a little bit more unique, I think gave us a lot of momentum too, and also a lot of new perspective around the world and what we want to do.
00:02:53.759 --> 00:02:58.159
Building your company in Israel, obviously most of the investors are cyber.
00:02:58.479 --> 00:02:59.520
Very, very cyber.
00:03:00.960 --> 00:03:02.560
You're a bit of an offshoot there.
00:03:02.719 --> 00:03:04.639
Maybe walk us through how that went for you.
00:03:04.719 --> 00:03:08.080
How was that experience explaining why you guys are not doing cybersecurity?
00:03:08.560 --> 00:03:11.360
I think no one expected us to do cybersecurity from the beginning.
00:03:11.520 --> 00:03:17.759
Like uh, you know, we're three co-founders, myself, Ido, who's used to be actually my commander in training in Special Forces.
00:03:17.919 --> 00:03:20.639
Um, he's a COO and our CTO who's from 8200.
00:03:20.800 --> 00:03:22.800
So he was like, Yeah, why aren't you doing cyber?
00:03:22.879 --> 00:03:27.759
He also consulted to like Wiz and Sierra, like, you know, that he worked a lot around these fields.
00:03:28.080 --> 00:03:35.360
But when we started, like we knew we didn't want to do, you know, the typical Israeli story.
00:03:35.520 --> 00:03:47.520
We wanted something that is a little bit different, something that's more unique, something that's more rooted in kind of like the infrastructure of like how communication is going to go about in the in the new world, uh, the new internet essentially.
00:03:47.840 --> 00:03:52.479
And I think cybersecurity, as interesting as it is, one, we don't have expertise in it.
00:03:52.639 --> 00:03:55.840
Second, we were definitely not, you know, the persona that investors would see.
00:03:56.000 --> 00:04:05.520
And I think investors were actually intrigued to see a new kind of persona coming into this, you know, entrepreneurship in Israel.
00:04:05.759 --> 00:04:19.199
In Israel, you know, as you mentioned, a lot of entrepreneurs are like, you know, very technical, coming from 8200, 4.9, all the incredible units that have like really incredible engineers and and specialize in very unique things very early on in their life.
00:04:19.439 --> 00:04:31.600
You know, we came in with like after five years of you know being officers in the army with a little bit of a different background, more sales and go-to-market oriented, and I think that really refreshened a lot of our investors.
00:04:31.759 --> 00:04:36.800
And to be honest, we only have one Israeli VC, the rest of them are American VCs.
00:04:37.199 --> 00:04:43.920
I mean, you had a similar story, having been technical and served, and then building a less technical domain.
00:04:44.240 --> 00:04:44.560
Yes.
00:04:44.800 --> 00:04:46.160
Oh, still the thunder here.
00:04:46.319 --> 00:04:48.079
But we have a very, a very similar background.
00:04:48.160 --> 00:04:49.519
I also grew up in special forces.
00:04:49.680 --> 00:04:54.240
And as we figured out right before starting recording, we also grew up in the same little town.
00:04:54.319 --> 00:04:56.000
So I know his brother very well.
00:04:56.319 --> 00:05:03.360
Yeah, he's uh he knows my brother, and it seems to be something in this series because we've we're already three people from the same little town in Israel in this.
00:05:03.920 --> 00:05:04.800
The only source family.
00:05:05.040 --> 00:05:06.160
The only source family.
00:05:07.759 --> 00:05:11.040
That's a good transition to what exactly Limey is doing today.
00:05:11.279 --> 00:05:16.399
One of the core theses is that AI agents will dominate the web, which I think we're also on that wavelength.
00:05:16.560 --> 00:05:20.720
And so what does that future look like and how does Limey play in it?
00:05:21.279 --> 00:05:52.639
So I think one of the things that happened kind of like in the past three years is that we understood that ChatGPT is not, and of course, like the equivalents, Cloud and Grok and everything, you know, all the AI search engines and essentially AI in general is not just us talking to a chatbot, but it's actually going to fundamentally impact how we see, communicate, get information, consume information, execute some tasks, and not only from a consumer perspective, but also from a B2B perspective, from businesses perspective, they're transitioning tremendously.
00:05:53.040 --> 00:05:56.800
And at the beginning, it was just like, yeah, how do I show up on chat?
00:05:56.959 --> 00:05:58.639
Like that was kind of like you know our focus.
00:05:58.720 --> 00:06:02.000
Like, how do I make sure that I can make brands, products show up on chat?
00:06:02.079 --> 00:06:19.439
But as we kind of like got even deeper and deeper and started working with Fortune 10s and Fortune 100s, and we started understanding what they actually need, got even deeper into their into the layer of where these AI agents communicate, we understood that this is fundamentally a much bigger change than just, you know, let's talk with a chat bot.
00:06:19.519 --> 00:06:22.720
But it's actually a complete different way of how we're getting information.
00:06:22.959 --> 00:06:35.600
And if you think of it just like a quick scenario, is that you go out, you go to ChatGPT asking for the best running shoes for women, running a marathon is an example, just goes out to websites, reinterprets everything, and writes it down for you, and you get the answer.
00:06:35.680 --> 00:06:36.639
You might trust it, you might not.
00:06:36.720 --> 00:06:37.600
You might go to the link.
00:06:37.759 --> 00:06:41.439
About 40% of searches end now with no clicks.
00:06:41.519 --> 00:06:43.600
So people are just starting to stay within.
00:06:43.839 --> 00:06:47.279
60% of the web traffic is AI agents.
00:06:47.519 --> 00:06:54.240
So really starting to see that these AI agents are starting to control essentially how we maybe think a little bit too.
00:06:54.560 --> 00:07:16.480
And when we came about it, we understood that we need to be and need to build something around that layer, that new networking layer that helps brands essentially control their communication with their end consumers, control how their brand is being represented, control what these AI agents are saying about them, and also building the infrastructure for these agents to communicate with brands.
00:07:16.560 --> 00:07:23.680
So think about it as like kind of like accessibility for agents, being able to access information in the best way possible and then to reinterpret it.
00:07:24.079 --> 00:07:32.000
So if I I don't want to read too much between the lines here, but walk me through a bit what's the the package for theoretical customer reviewers.
00:07:32.160 --> 00:07:39.439
Do they they probably care about how they're perceived, probably decision-making support if it's a B2B use case?
00:07:39.839 --> 00:07:47.279
And then you do that by somehow putting a layer that the agent can interact with, the customer's layer can interact with.
00:07:47.759 --> 00:07:52.240
Yeah, so kind of like how our how we work with our customers on is usually on three layers.
00:07:52.399 --> 00:08:02.480
One of the main things that every CMO and every marketer wants to know is essentially what he needs to do, what he needs to optimize for to drive the most revenue to the company.
00:08:02.639 --> 00:08:09.759
It can be clicks or users coming in or Reddit links or GitHub stars, whatever, right?
00:08:09.839 --> 00:08:18.240
But he needs, you know, all these CMOs and marketers need to make sure that they're the first, you kind of like the shining star on the keywords and prompts that people are searching.
00:08:18.399 --> 00:08:24.959
So the first layer that we give them is the ability to actually know what people are searching about them, literally like know what their target audience is searching.
00:08:25.040 --> 00:08:28.800
And this is like at its core, it's like the most valuable thing for them right now.
00:08:28.959 --> 00:08:38.639
On top of that, we're giving them the ability to see what are the sources that are being referred to by these LLMs once someone asks a question, like you know, what their target audience says.
00:08:38.799 --> 00:08:47.440
And the third layer is that we give them the ability to make enhancements and recommendations and implement, sorry, recommendations but AI agents autonomously.
00:08:47.600 --> 00:08:49.679
So everything just happens on the go.
00:08:49.759 --> 00:08:54.240
You know, once they get a little bit of confidence in the product, they can enhance their content on their website.
00:08:54.320 --> 00:09:03.039
Everything can be changed constantly based on the data, but also on any third-party websites they would need to publish in, everything is very much executed automatically.
00:09:03.440 --> 00:09:08.639
But with the premises that it will influence the perception of the customer side agents, okay.
00:09:08.960 --> 00:09:10.559
With the premise that they're going to be shown.
00:09:10.639 --> 00:09:15.360
And one of the things that we've added recently is also the ability to tie that back into revenue attribution.
00:09:15.519 --> 00:09:19.200
So say, okay, we've optimized for, I don't know, 50 new prompts.
00:09:19.360 --> 00:09:26.960
Now we're showing up on these supposedly target audience prompts that we needed to focus on.
00:09:27.120 --> 00:09:28.799
How much has that resulted in sales?
00:09:28.879 --> 00:09:30.639
And how much does that really have an impact?
00:09:30.720 --> 00:09:38.080
And some of the brands that we work with have been seeing 8 to 12% of the revenue coming in from chat from due to optimizations they did with us.
00:09:38.159 --> 00:09:40.720
So it's been kind of like a wild ride recently.
00:09:42.240 --> 00:09:56.480
So how do you maintain control while not having control essentially over one, the agents, maybe two the LLMs, and three the competition, meaning there's other folks that are also tackling this problem that are trying to gain control.
00:09:56.559 --> 00:09:58.879
And it's it can be a little bit ambiguous right now.
00:09:59.360 --> 00:10:03.440
Yeah, I think ambiguous is a wonderful world for what's going on here now.
00:10:03.679 --> 00:10:05.840
So there were in terms of the control.
00:10:05.919 --> 00:10:11.279
So I think the shift is so big that they're, you know, everyone's trying to take a little bite, you know, everyone wants a slice of the pie.
00:10:11.679 --> 00:10:23.759
And one of the things that a lot of the brands, you know, what some of our competitors did, and you know, they tried to, you know, kind of like give something in the beginning was they're using what's called clickstream data.
00:10:23.919 --> 00:10:34.320
And that clickstream data is essentially how me and you or anyone in the world is interacting on their browser, client-side browser, meaning, you know, the code that is loaded on our Chrome or Safari or whatever.
00:10:34.559 --> 00:10:39.360
And that clickstream data is referring to how humans are interacting essentially with the web.
00:10:39.519 --> 00:10:50.799
So they took all that data, they put it there, they tried to overlay it on agents' decisions and agents' logics on how to surface things, and then they just gave kind of like a similar what I said, but based on different data.
00:10:50.879 --> 00:10:54.480
So analytics, recommendations, and potentially also the execution.
00:10:54.720 --> 00:11:09.039
In order to gain back the control and the chaos, we understood that because us as humans and AI agents have a complete different logic, and we go to the browser client side, but the agent has no access to open a browser, it's actually going to the server side, communicating through the CDN.
00:11:09.200 --> 00:11:10.480
We should actually be right there.
00:11:10.639 --> 00:11:16.480
You know, we should actually be where these AI agents are actually actually interacting with websites and content and digital assets.
00:11:16.720 --> 00:11:24.480
And the way that we gave them back the control to these, to the CMOs and marketers that we work with is being able to understand the agent's logic.
00:11:24.639 --> 00:11:26.159
Like, why am I going to this website?
00:11:26.320 --> 00:11:27.759
Why am I not going to the other one?
00:11:27.919 --> 00:11:29.600
What am I fetching from your content?
00:11:29.919 --> 00:11:31.759
What piece is intriguing me?
00:11:31.919 --> 00:11:34.159
What is interesting within your content?
00:11:34.240 --> 00:11:37.600
What am I attracted to on your website or any third-party websites?
00:11:37.759 --> 00:11:42.559
And all of that knowledge is being essentially converted into actions they can take.
00:11:42.720 --> 00:11:47.120
So there's a clear roadmap to any marketer to say, okay, this is what's happening.
00:11:47.200 --> 00:11:54.799
This is what AI agents in my sector are attracted to based on the prompts that my target audience is searching, and this is what I need to do in order to make sure that I'm top performing.
00:11:55.039 --> 00:12:04.159
So this kind of like, you know, gets them the ability to be a little bit more relaxed, come to their board meetings and say, hey, we're now showing up better.
00:12:04.559 --> 00:12:09.279
How do you find historically marketeers in in B2B love a lot of them?
00:12:09.519 --> 00:12:12.639
They're not the most sophisticated marketeers in the market.
00:12:13.279 --> 00:12:19.840
How do you find these conversations around introducing the idea of agentic search, agentic positioning?
00:12:20.000 --> 00:12:21.200
How do they react to that?
00:12:21.519 --> 00:12:26.240
That's a really good question because that's something that we constantly battle within our marketing team.
00:12:26.399 --> 00:12:27.679
Like, what are we going to say?
00:12:27.840 --> 00:12:30.960
Are we gonna say business to agent or agent-led growth?
00:12:31.120 --> 00:12:33.200
Or are we just gonna say show up on ChatGPT?
00:12:33.360 --> 00:12:35.200
You know, it's like a constant battle that we have.
00:12:35.600 --> 00:12:36.399
How simplified do you mean?
00:12:36.639 --> 00:12:38.159
Yeah, like how simplified it is, right?
00:12:38.240 --> 00:12:43.519
Because like, you know, when we talk to investors, it needs to be these big headlines like agent-led growth and business to agents.
00:12:43.600 --> 00:12:46.559
But you know, when you go up to a marketer, like loves a buzzword.
00:12:46.720 --> 00:12:47.440
Yeah, right.
00:12:47.679 --> 00:12:48.399
You guys love it.
00:12:48.639 --> 00:12:54.240
But like when you go out to customers, you know, they don't care about agent-led growth, right?
00:12:54.320 --> 00:12:57.039
They care about the bottom line revenue, and it's exactly what you're saying.
00:12:57.279 --> 00:13:01.919
One of the things that we understood is there's there's levels of understanding, also in different markets, by the way.
00:13:02.080 --> 00:13:07.600
Like we sell in AIPAC, in the United States, in Europe, and Israel, and each market understands completely differently.
00:13:07.679 --> 00:13:10.879
Sometimes in the US, it's prone to different regions, like you know, San Francisco.
00:13:11.120 --> 00:13:13.759
Agent-led growth is like the best term you can put out there.
00:13:13.919 --> 00:13:17.759
But like in different places, you know, it's just like, hey, let's make yourself on Chat GPT.
00:13:18.240 --> 00:13:29.840
And I think there are a lot of marketers recently understood that they have to make the transition to have a much more sophisticated tech stack, and they need to know more.
00:13:30.000 --> 00:13:39.200
And they cannot just go to traditional places because if you think of it, you're if you're a marketer, right now, someone's searching on Google, it's getting AI overviews, right?
00:13:39.279 --> 00:13:43.279
So it's pushing the ads down, it's pushing the SEO down.
00:13:43.519 --> 00:13:46.559
So ads are not getting clicked on, so they're costing more.
00:13:46.720 --> 00:13:48.559
SEO is not as relevant as possible.
00:13:48.639 --> 00:13:52.080
So people are seeing incredible drops in their organic traffic.
00:13:52.320 --> 00:13:53.840
So keck is going up.
00:13:54.080 --> 00:13:58.000
Organic is not showing as much as they want to, like not performing as much as they want to.
00:13:58.159 --> 00:13:59.519
They need to like do something.
00:13:59.679 --> 00:14:01.200
So the I I guess you guys see it too.
00:14:01.279 --> 00:14:08.000
A lot of guerrilla marketing that's been happening recently, a lot of like use of like AI agents, growth hacking, their things they're doing.
00:14:08.480 --> 00:14:09.600
Kind of like a reinventing.
00:14:09.679 --> 00:14:13.519
They reinvent what organic means is just evolving.
00:14:13.759 --> 00:14:18.559
You can see it's word of mouth used to be human word of mouth, now it's like some hybrid.
00:14:19.120 --> 00:14:19.840
Google word of mouth.
00:14:20.080 --> 00:14:23.039
Yeah, or like agent-human combo word of mouth.
00:14:23.279 --> 00:14:23.679
Absolutely.
00:14:23.840 --> 00:14:29.600
I mean, if agents are making the decisions of what to surface, you know, you you want them, you know, you want to be in their mouth in a way.
00:14:29.679 --> 00:14:29.919
Exactly.
00:14:30.159 --> 00:14:37.759
Have there been any hacks or tricks that you found that have mapsites or products more desirable for agents that you're advising your customers on?
00:14:38.000 --> 00:14:42.480
So that's a question like we're getting a lot right now is how do we make our website or our product agent ready?
00:14:42.559 --> 00:14:49.600
And if agents are surfing the web, we want to not only come up first, but we want to make sure that there's a click, there's a pr, it's it's the follow-through.
00:14:49.759 --> 00:14:57.039
And I feel like this whole GEO space started off to your point of just like a dashboard, or not just a dashboard, like this is how you're showing up at ChatGBT, this is what you need to do better.
00:14:57.200 --> 00:14:59.120
And I think that got commoditized quickly.
00:14:59.279 --> 00:15:03.840
And then I think where all of these companies are going, it's a few different directions right now.
00:15:04.240 --> 00:15:05.360
Yeah, a hundred percent.
00:15:05.519 --> 00:15:09.279
I think there's two things that kind of like really impact.
00:15:09.440 --> 00:15:12.240
One is like first understanding like what are people searching, right?
00:15:12.320 --> 00:15:15.279
Like you really want to know essentially what your target audience is searching.
00:15:15.600 --> 00:15:24.960
But that has also some limitations because whenever I'm searching on ChatGPT, what happens is that ChatGPT translates that into something that's called a query fan out.
00:15:25.200 --> 00:15:30.159
So I write best running shoes for men running a marathon with knee pain in Boston next year, right?
00:15:30.240 --> 00:15:31.600
It's such a long prompt.
00:15:31.759 --> 00:15:36.960
So, you know, so chat doesn't really take all that prompt and and looks for for specific answer to that.
00:15:37.039 --> 00:15:39.200
So what it does, it kind of like reinterprets it.
00:15:39.360 --> 00:15:41.120
And the way that that's called a query fan out.
00:15:41.200 --> 00:15:43.759
So it might say running shoes men reviews 2026.
00:15:43.919 --> 00:15:44.720
You know, simple as that.
00:15:44.879 --> 00:15:46.480
Marathon, running shoes reviews.
00:15:46.639 --> 00:15:49.679
And that's what it's going out and sending the agents to go and find sources on.
00:15:49.840 --> 00:15:55.600
And the way that these agents are actually finding sources is that they're translating that word into a vector and looking for similar vectors.
00:15:55.679 --> 00:15:57.360
So they're doing vector searches.
00:15:57.759 --> 00:16:05.519
And I would say that the main that that you know, when I always get asked, like, what what is the one thing we need to do?
00:16:05.679 --> 00:16:16.639
So I always say, like, it unfortunately there's no silver bullet, but what the first place I would start at is like first understanding what my target audience is searching and what vector it's being translated into.
00:16:16.799 --> 00:16:26.799
Because if keywords were a very simplified process in Google of like, okay, you need to have these keywords, they need to have some amplifications, they need to have backlinks or anything that SEO was oriented.
00:16:26.960 --> 00:16:35.679
Right now, attracting the agents requires to have a very similar vector space to what is being searched and to what is on your website, so they would be attracted to it.
00:16:35.919 --> 00:16:41.120
I know it's very kind of like a vague answer in a way because you know it's it's a complicated task.
00:16:41.360 --> 00:16:43.919
So that would be my best advice.
00:16:44.480 --> 00:16:52.639
Is the future world where it looks like implementing that almost on the fly in a dynamic way once you have trust with the marketing teams?
00:16:53.120 --> 00:16:55.759
Yeah, so today we work with over 200 brands.
00:16:55.840 --> 00:17:07.519
I think about 20% of them are Fortune 100s, then all these brands are, I think maybe 10% of them are automating, like everything's completely automated, you know, brands that have like very, very strict brand guidelines.
00:17:07.599 --> 00:17:15.200
We work with pharmaceutical manufacturers, you know, it's like for them to say something that's out of balance or out of guardrails is very, very complicated.
00:17:15.279 --> 00:17:18.480
And they're feeling very comfortable to automate every process.
00:17:18.720 --> 00:17:25.519
And I think in addition to that, if you really think of the way, even like how you browse around the web, right?
00:17:25.599 --> 00:17:29.039
Like you like, how often do you go to websites today?
00:17:29.200 --> 00:17:34.799
Like, you know, you get an answer in Chat GPT, you might go to to the link to to vet it.
00:17:34.880 --> 00:17:36.400
I don't think anyone's doing it anymore.
00:17:36.559 --> 00:17:42.799
You know, people are just looking, you know, we're pretty much staying within these LMs, they're becoming centralized hubs of information.
00:17:43.119 --> 00:17:48.160
And I think the future of the internet is not us just having beautiful websites.
00:17:48.240 --> 00:17:58.240
I mean, these websites are still gonna exist, but there's just gonna be some content databases in the world that they're just going to be, you know, information's gonna be taken from them.
00:17:58.480 --> 00:18:11.039
And if marketers were like, yeah, I'm not uploading this blog post, I'm not writing this, this doesn't fit my brand's guidelines or brand kit or whatever, right now that doesn't really matter because AI agents are taking it, reinterpreting it, and putting it out in front of your audience.
00:18:11.359 --> 00:18:13.039
So I think it's very much changing.
00:18:13.440 --> 00:18:22.640
I'm curious, do your customers talk to you about a lot of what we're hearing about the next step will be this new infrastructure for ads in generative formats.
00:18:22.880 --> 00:18:30.079
So chat is the most obvious ones, but also you have generative media where you can insert logos or ads or hints.
00:18:30.319 --> 00:18:32.319
Is that something that's coming up from your customers?
00:18:32.640 --> 00:18:34.000
So yeah, it's coming up a lot.
00:18:34.160 --> 00:18:40.400
Like they're definitely, you know, they want to see how they're they can monetize, you know, these LLMs.
00:18:40.640 --> 00:18:43.759
Chat Perplexity looked at it like November 24.
00:18:43.839 --> 00:18:45.519
They were like checking, testing ads.
00:18:45.599 --> 00:18:47.359
They like they took a step back.
00:18:47.519 --> 00:18:51.359
OpenAI, as you guys know, recently released uh, you know, their ad play.
00:18:51.599 --> 00:19:00.480
They haven't accepted a lot of people yet into their kind of like the equivalent of Google ads, you know, to give people the ability to actually perform ads.
00:19:00.799 --> 00:19:02.319
But yeah, we're getting that all the time.
00:19:02.400 --> 00:19:09.839
Like everyone is understanding they need to be there, everyone's understanding that you know they need to monetize that channel somehow.
00:19:10.240 --> 00:19:11.920
The costs are pretty significant.
00:19:12.079 --> 00:19:21.599
I don't know if you saw, but it's about sixty dollars for impressions, which is like three times more than LinkedIn, which is right now the cossiest platform.
00:19:21.680 --> 00:19:23.039
So it's quite significant.
00:19:23.200 --> 00:19:27.119
And it's impressions, not click, which is a kind of like the Google model.
00:19:27.759 --> 00:19:32.559
How do you think you and your approach are gonna be positioned in that world?
00:19:33.119 --> 00:19:37.119
The ad world, there is two ways to to go about it.
00:19:37.279 --> 00:19:43.359
There is the closed systems, like meta, that's like, what do you want to target?
00:19:43.599 --> 00:19:49.039
We'll find it, we're gonna give us the copy, the algorithm runs, you know, we do our thing.
00:19:49.119 --> 00:19:52.799
You know, you just fine-tune your c your content to make sure it's getting picked up.
00:19:52.960 --> 00:20:04.480
And there is the second kind of like approach that more of like the Google approach and you know, the publisher's approach, which is like this is the this is the ability, this is the keys to the Bugatti.
00:20:04.799 --> 00:20:06.400
Take it and drive wherever you want.
00:20:06.480 --> 00:20:14.000
Like you can go wherever you want, you can build the content, you can decide what to bid on, you can decide how much to bid, like everything is very open.
00:20:14.319 --> 00:20:24.079
From what I'm seeing, you know, we also have a couple of customers that are in this, like kind of like POC with OpenAI around the ads, it's going to be a bit more closer to Meta.
00:20:24.240 --> 00:20:31.920
So we won't have the ability to run the bidding because they're going to do the bidding, and the bidding is kind of like where, you know, it's the most interesting place.
00:20:32.000 --> 00:20:33.279
Like, what prompts am I bidding on?
00:20:33.359 --> 00:20:36.079
What prompts I'm putting more money on be based on their performance.
00:20:36.240 --> 00:20:37.519
They're not going to tell you exactly.
00:20:37.599 --> 00:20:40.720
It's they're just going to run that for you, and you're going to get the end results.
00:20:40.960 --> 00:21:05.039
But when you're running a campaign, one of the most significant things is not actually the campaign itself, but it's the research before the campaign, what's happening throughout the campaign, to know what to, what campaigns to take down, what campaigns to keep, what campaigns to put more efforts on, to create more and duplicate them, and also the post, like kind of like the post-mortem of the campaign, like you know, researching about it, what happened, what we can do.
00:21:05.200 --> 00:21:07.759
We come in kind of like throughout all these phases.
00:21:08.000 --> 00:21:13.680
We cannot do the bidding, but we can help our marketers make the right decisions on on what to do.
00:21:13.759 --> 00:21:16.799
And that's exactly where the money is actually really lies in.
00:21:17.200 --> 00:21:21.119
Being a founder, we we sometimes get lost in the problem we're solving.
00:21:21.359 --> 00:21:24.319
I find that there's always a really interesting story around the motivation.
00:21:24.559 --> 00:21:33.039
Why are you so obsessed with the marketeers and how people are gonna be doing decision making and buying in in this new age of agentic search and web.
00:21:33.440 --> 00:21:39.359
Maybe walk us through a bit what influenced you into dedicating some of your prime years to this problem.
00:21:39.920 --> 00:21:41.119
Prime years, I love that.
00:21:41.200 --> 00:21:42.079
That's good.
00:21:43.359 --> 00:21:47.759
I think there's two things that kind of like came to mind when we started the company.
00:21:48.079 --> 00:21:55.759
First is that we want to work with a target audience that suits us, like that we're we're fitted in, like we're welcomed in, I would say.
00:21:55.920 --> 00:21:59.839
Like if I go to different markets, you know, insurance markets, like I don't know.
00:22:00.240 --> 00:22:01.599
I'm not gonna be a great fit there.
00:22:01.839 --> 00:22:02.880
You don't give the insurance vibes?
00:22:03.119 --> 00:22:04.799
I don't give the insurance vibes at all, you know.
00:22:04.960 --> 00:22:07.680
Like I would love to, but I don't give my the insurance vibes.
00:22:07.839 --> 00:22:09.519
I think marketers accept me much more.
00:22:09.759 --> 00:22:13.599
So first we wanted to go to a place we'd be welcomed in with the people that would be enjoying working.
00:22:13.839 --> 00:22:16.400
And I think marketing is like so much creativity.
00:22:16.480 --> 00:22:20.079
It's like incredible people, things are just flowing, all the money's there too.
00:22:20.160 --> 00:22:25.279
So like it's quite an incredible department to work in at any company, I think.
00:22:25.599 --> 00:22:29.839
And you know, at the beginning it was it's very much that, okay, let's let's help marketers.
00:22:30.240 --> 00:22:40.720
But as kind of like we we started deep diving more and we understood that it's actually much more deep into how information is being communicated to people, we understood this is like a fundamental shift.
00:22:40.799 --> 00:22:54.000
And this is like an area where we can not only help build the infrastructure for the next way of how the world communication is going to go through and run through, but also really help influence it.
00:22:54.400 --> 00:23:02.079
And being at that junction of where content and information is being delivered through, I think is one of the most exciting things in the world.
00:23:02.240 --> 00:23:05.039
So I would say this makes my years prime, right?
00:23:05.119 --> 00:23:11.119
Like not sitting at the beach or doing like the fact that I'm doing this and dealing with these incredible problems.
00:23:11.279 --> 00:23:14.079
I think this is what makes my my life prime.
00:23:14.400 --> 00:23:19.440
What do you think is the biggest misunderstanding that current generation marketeers have about AI?
00:23:19.680 --> 00:23:25.039
So I think the the younger folks are like they're understanding everything.
00:23:25.200 --> 00:23:32.880
Like they're very much understanding, they're very much out there, they know what they want to do, they see AI, they see it evolving, they love playing with the products.
00:23:33.039 --> 00:23:40.880
I think like some of the marketers that I work with, I'm 29, you know, some of them are like 21, they're much more advanced than me, and either incredible, they're doing super cool things.
00:23:41.039 --> 00:23:51.759
I think sometimes where it gets stuck is is either kind of like higher management or you know, kind of like the companies that are having a hard time going through that rewiring.
00:23:52.319 --> 00:23:55.759
Applying old frameworks of budgeting to new methods.
00:23:56.079 --> 00:23:56.480
Exactly.
00:23:56.640 --> 00:23:59.759
Like, you know, SEO is like irrelevant now.
00:23:59.839 --> 00:24:02.640
Like what are you still like we're not trying to focus on something else?
00:24:02.720 --> 00:24:07.119
You know, it's like it's these things that you have to like try to walk them through in understanding.
00:24:07.200 --> 00:24:15.759
And you know, sometimes we see it, but you would think that such executive, experienced people would have like the ability to make these transitions really quickly.
00:24:15.839 --> 00:24:25.119
Sometimes it's not only up to them, but we are seeing a lot of the time that there's a lot of like hesitation, stagnation around like things that we find hard to understand why it's taking so much time.
00:24:25.359 --> 00:24:43.200
And I think sometimes a lot of AI adoption is is complicated, especially around marketing, because for one, it's potentially the first or the second largest expense in the company, which is you know, you don't want to shuffle everything around, you know, space at a place where you know money is like so impactful.
00:24:43.359 --> 00:24:57.839
And the second is that this is what drives impact and shifting and changing things around that place, you know, around the revenue, the leads, what are we we we're getting a consistent, repetitive funnel that's working beautifully.
00:24:58.079 --> 00:24:59.200
Why should we change it?
00:24:59.440 --> 00:25:02.880
You know, it's like it's hard to make the decision, it's hard to rip off the band-aid, I think.
00:25:03.119 --> 00:25:06.160
What do you see as the moment typically for Fortune 100?
00:25:06.960 --> 00:25:09.279
Once it monetizes for like three or four months.
00:25:09.440 --> 00:25:11.359
Like once they see it, can we actually monetize?
00:25:11.440 --> 00:25:14.559
And this is not just like, oh yeah, like it's it's just showing up on chat.
00:25:14.640 --> 00:25:15.119
Like, who cares?
00:25:15.200 --> 00:25:16.319
It's a branding thing.
00:25:16.480 --> 00:25:27.359
I don't know if you know, but like one of the things that le that companies least invested in was like branding technologies, like social intelligence technologies, like don't really care about it.
00:25:27.519 --> 00:25:31.519
Just like outsourcing it and using like services for it, or yeah, they never really used any software.
00:25:31.680 --> 00:25:32.160
Like who would pay?
00:25:32.480 --> 00:25:33.119
It's the first to go.
00:25:33.200 --> 00:25:34.880
It's hard to justify, it's not it's in downtown.
00:25:35.200 --> 00:25:35.599
Exactly.
00:25:35.759 --> 00:25:37.680
And right now it's like the most tangible thing.
00:25:37.839 --> 00:25:39.680
Right now, that's exactly what you want to do.
00:25:39.920 --> 00:25:44.160
It's funny, but like things kind of like really shifted on 180 degrees.
00:25:44.480 --> 00:25:53.279
In full transparency, before starting my my company, I used to run uh global operations for a public company, so all of marketing operations was under my responsibility.
00:25:53.359 --> 00:26:01.200
And I always thought of the marketing space as like an this weird approximation, impressions, clicks.
00:26:01.440 --> 00:26:03.599
Really, what we're trying to get to is just intent.
00:26:03.839 --> 00:26:10.880
If you could tap into someone's brain and understand what's going on there, what are their fears, what are they looking for, what are they trying to accomplish?
00:26:11.359 --> 00:26:13.440
That's the pure thing.
00:26:14.079 --> 00:26:22.880
I think fundamentally when the medium is more conversational, you essentially bypass a lot of that, you know, bullshit that we used to engage with in the past.
00:26:23.359 --> 00:26:31.839
Do you think these modern platforms that you're also building in are gonna like just transition to that pure intent signal?
00:26:32.240 --> 00:26:34.960
I think the word approximation is perfect.
00:26:35.279 --> 00:26:54.640
I think that there's a lot of approximations that have been done in the previous world around the previous Google world around revenue and attribution and clicks, and and you know, you go to your HubSpot, you see one number, you go to your Google Analytics, you see other number, like everything is like no one really knows.
00:26:54.799 --> 00:26:59.039
So an approximation was a big part of like of ops essentially in marketing.
00:26:59.279 --> 00:27:04.640
I think that right now we're not trying to find human intent, but we're trying to find machine intent.
00:27:04.799 --> 00:27:09.440
And I think finding machine intent is becoming much more crucial in that world.
00:27:09.599 --> 00:27:19.200
So machine intent is much more logical, much more comprehensive, much more structured than human intent.
00:27:19.359 --> 00:27:28.559
And that's why it makes it a little bit easier to decrease approximations and to have literally rooted accurate numbers.
00:27:28.880 --> 00:27:35.680
Just for the fun of it, I do recommend you one time read through the developer traces of clot code.
00:27:36.240 --> 00:27:37.680
It feels very irrational.
00:27:37.920 --> 00:27:38.160
Really?
00:27:38.400 --> 00:27:41.200
The traces sometimes go like, I should have not said that word.
00:27:41.279 --> 00:27:43.279
I wonder what the human thinks about me.
00:27:43.519 --> 00:27:45.519
Maybe I'll go back and do it again.
00:27:45.759 --> 00:27:50.240
Or maybe the human will think I'm indecisive, so I'll randomly reverse us not.
00:27:50.559 --> 00:27:53.839
Yeah, but like just it's actually it sounds freakishly human.
00:27:54.160 --> 00:27:55.599
Right, but that's the LLM, right?
00:27:55.759 --> 00:27:55.920
Yeah.
00:27:56.319 --> 00:28:03.759
And the way what we're actually trying to influence is not the LLM itself, which is freakishly weird thinking about weird things.
00:28:03.920 --> 00:28:05.359
It's actually what's called a retriever.
00:28:05.440 --> 00:28:12.240
That's what goes out and finds these sources and hands off to that freakish beast that you just mentioned to write the answer.
00:28:12.799 --> 00:28:17.599
Are you building Limee in New York in person today, or is it still a hybrid in Israel?
00:28:18.079 --> 00:28:19.519
So we have two offices.
00:28:19.759 --> 00:28:22.319
Actually, potentially opening a third one very soon.
00:28:22.480 --> 00:28:23.279
We have two offices.
00:28:23.359 --> 00:28:26.079
One is in Tel V where product engineering and design is.
00:28:26.240 --> 00:28:32.640
You know, our CTO is from 8200, our incredible team is all from there, designers, the developers.
00:28:32.720 --> 00:28:35.440
It's really like the top of the top of engineers.
00:28:35.519 --> 00:28:36.559
We couldn't wish for better people.
00:28:36.799 --> 00:28:37.839
And the shekel is so attractive.
00:28:38.079 --> 00:28:39.599
And the shekel is attractive, right?
00:28:39.920 --> 00:28:40.960
The main reason.
00:28:42.240 --> 00:28:44.960
But it there were times that it wasn't so attractive.
00:28:45.200 --> 00:28:49.519
And our go-to-market is here in New York, but we also have a couple of West Coast hires.
00:28:49.599 --> 00:28:58.559
So we are a little bit hybrid in a way, but I think like once you want to take over a market, you have to be in the market.
00:28:58.640 --> 00:29:00.160
You can't just, you know, be up far.
00:29:00.400 --> 00:29:02.640
That's why you're also opening an APAC office soon.
00:29:02.799 --> 00:29:06.559
But you know, really kind of like to make sure that we're in the place where our customers are.
00:29:06.640 --> 00:29:10.000
You know, we want to understand when they're send their needs, especially around the marketing.
00:29:10.240 --> 00:29:12.319
And you've relocated recently to New York.
00:29:12.559 --> 00:29:12.799
Yeah.
00:29:12.960 --> 00:29:14.079
How's that process been for you?
00:29:14.240 --> 00:29:15.119
I love Williamsburg.
00:29:15.200 --> 00:29:15.359
Yeah.
00:29:16.079 --> 00:29:17.200
I thought I will never like it.
00:29:17.359 --> 00:29:19.039
Honestly, as a visitor, I never liked it.
00:29:19.279 --> 00:29:21.279
All Israeli founders end up in Williamsburg.
00:29:21.359 --> 00:29:21.519
Yeah.
00:29:21.680 --> 00:29:22.799
Or Upper West Side.
00:29:23.039 --> 00:29:26.000
Or Upper West, but I think it's just like Williamsburg is very Tel Aviv.
00:29:26.400 --> 00:29:26.559
Yes.
00:29:26.720 --> 00:29:26.880
Yeah.
00:29:26.960 --> 00:29:27.519
Very Tel Aviv.
00:29:27.680 --> 00:29:27.759
Yes.
00:29:28.000 --> 00:29:29.519
Like short buildings, you know.
00:29:29.839 --> 00:29:31.839
Do you want to wrap with a few questions before we close out?
00:29:32.079 --> 00:29:33.119
Let's do it.
00:29:33.519 --> 00:29:36.960
Does brands still matter when agents are intermediating everything?
00:29:37.279 --> 00:29:39.279
Yes, like matters most, I think.
00:29:39.599 --> 00:29:44.400
What is a KPI that a CMO should potentially stop tracking next year?
00:29:45.200 --> 00:29:45.759
Clicks.
00:29:46.240 --> 00:29:49.200
If Limey failed tomorrow, what do you think would be the reason?
00:29:49.440 --> 00:29:50.480
Us founders.
00:29:50.720 --> 00:29:51.359
The founders?
00:29:51.519 --> 00:29:51.759
Yeah.
00:29:51.920 --> 00:29:52.160
Okay.
00:29:52.480 --> 00:29:53.680
I think full accountability.
00:29:53.920 --> 00:29:54.960
Always founders.
00:29:55.279 --> 00:29:56.480
This was an awesome episode.
00:29:56.640 --> 00:29:57.359
Thank you for coming on.
00:29:57.680 --> 00:29:58.400
Thank you so much, guys.
00:29:58.559 --> 00:29:58.720
Thank you.
00:29:59.039 --> 00:29:59.519
So much fun.
00:29:59.759 --> 00:30:00.960
Really appreciate it.
00:00:00.160 --> 00:00:12.720
This is like an area where we can not only help build the infrastructure for the next way of how the world communication is going to go through and run through, but also really help influence it.
00:00:13.039 --> 00:00:20.399
Welcome to First Commit, a podcast from Notable Capital where we talk to the builders solving hard problems before the industry catches up.
00:00:22.480 --> 00:00:34.719
Today we are joined by Aviv Shamney, co-founder of Lime AI, a platform building the connection layer between brands and AI agents, helping global companies understand, measure, and influence how agents discover, fetch, and convert on their products across the web.
00:00:34.880 --> 00:00:35.679
Welcome, Aviv.
00:00:36.000 --> 00:00:36.799
Hey, how's it going?
00:00:37.039 --> 00:00:37.759
Good to be here.
00:00:38.320 --> 00:00:39.520
Excited to get rolling.
00:00:39.759 --> 00:00:44.799
So you have an extremely impressive background across engineering, sales, and large-scale data systems.
00:00:44.960 --> 00:00:49.119
We would love to touch on kind of how that background informed what you're building today.
00:00:49.520 --> 00:00:52.880
Yeah, so first, really excited to be here, and it's good to meet the both of you.
00:00:53.119 --> 00:01:07.920
Basically, kind of like when when we started, when we saw this whole shift happening with AI, we understood that this is much more than just like a little tweak of yeah, we can now talk to data, but it's actually going to influence essentially how we're even communicating with kind of like the digital world.
00:01:08.239 --> 00:01:25.200
And when we deep dive into really understanding the fundamentals of how these things work, you know, that required essentially, you know, our technical expertise and really taking a very, very close look into how these AI agents are interacting with essentially the internet, with digital assets.
00:01:25.280 --> 00:01:27.359
What is the net what is the new networking layer?
00:01:27.519 --> 00:01:33.920
And so that really got us to a point where we were able to uncover a lot of very interesting things.
00:01:34.159 --> 00:01:37.760
And, you know, some people say that you don't really need technical backgrounds as a co-founder.
00:01:37.920 --> 00:01:44.879
I might agree, but I think that was like a especially now with how everything is changing, that is something that we needed to have in front of us.
00:01:45.120 --> 00:01:45.599
So yeah.
00:01:45.920 --> 00:01:46.239
For sure.
00:01:46.319 --> 00:01:54.000
And I feel like most of the Israeli founders we talk to and work with are operating in the cybersecurity domain, and you're quite literally at the opposite end of that spectrum.
00:01:54.239 --> 00:01:58.239
So what was it about the idea with Limey or where did you see it firsthand?
00:01:58.640 --> 00:02:02.400
Yeah, so we didn't come from, like at least I didn't.
00:02:02.480 --> 00:02:07.200
My CTO did, he's from 8200, also not specifically from the cybersecurity department.
00:02:07.439 --> 00:02:16.319
But I for sure wasn't a part of like the whole cybersecurity clan in Israel, which is also extremely unique and has a beautiful glow and aura behind it.
00:02:16.639 --> 00:02:22.560
But my background actually comes from like a little bit different areas in the army, more special forces.
00:02:22.719 --> 00:02:24.479
So that was a little bit of a different route.
00:02:24.639 --> 00:02:37.280
And I think there's a lot of very interesting things that we are now learning in this new digital world around how AI agents are interacting essentially with digital assets, whether it's websites or whether it's content that is being put out there.
00:02:37.680 --> 00:02:53.360
And the fact that, you know, we were able to kind of like deviate from the classic Israeli story of cybersecurity to something that is a little bit more unique, I think gave us a lot of momentum too, and also a lot of new perspective around the world and what we want to do.
00:02:53.759 --> 00:02:58.159
Building your company in Israel, obviously most of the investors are cyber.
00:02:58.479 --> 00:02:59.520
Very, very cyber.
00:03:00.960 --> 00:03:02.560
You're a bit of an offshoot there.
00:03:02.719 --> 00:03:04.639
Maybe walk us through how that went for you.
00:03:04.719 --> 00:03:08.080
How was that experience explaining why you guys are not doing cybersecurity?
00:03:08.560 --> 00:03:11.360
I think no one expected us to do cybersecurity from the beginning.
00:03:11.520 --> 00:03:17.759
Like uh, you know, we're three co-founders, myself, Ido, who's used to be actually my commander in training in Special Forces.
00:03:17.919 --> 00:03:20.639
Um, he's a COO and our CTO who's from 8200.
00:03:20.800 --> 00:03:22.800
So he was like, Yeah, why aren't you doing cyber?
00:03:22.879 --> 00:03:27.759
He also consulted to like Wiz and Sierra, like, you know, that he worked a lot around these fields.
00:03:28.080 --> 00:03:35.360
But when we started, like we knew we didn't want to do, you know, the typical Israeli story.
00:03:35.520 --> 00:03:47.520
We wanted something that is a little bit different, something that's more unique, something that's more rooted in kind of like the infrastructure of like how communication is going to go about in the in the new world, uh, the new internet essentially.
00:03:47.840 --> 00:03:52.479
And I think cybersecurity, as interesting as it is, one, we don't have expertise in it.
00:03:52.639 --> 00:03:55.840
Second, we were definitely not, you know, the persona that investors would see.
00:03:56.000 --> 00:04:05.520
And I think investors were actually intrigued to see a new kind of persona coming into this, you know, entrepreneurship in Israel.
00:04:05.759 --> 00:04:19.199
In Israel, you know, as you mentioned, a lot of entrepreneurs are like, you know, very technical, coming from 8200, 4.9, all the incredible units that have like really incredible engineers and and specialize in very unique things very early on in their life.
00:04:19.439 --> 00:04:31.600
You know, we came in with like after five years of you know being officers in the army with a little bit of a different background, more sales and go-to-market oriented, and I think that really refreshened a lot of our investors.
00:04:31.759 --> 00:04:36.800
And to be honest, we only have one Israeli VC, the rest of them are American VCs.
00:04:37.199 --> 00:04:43.920
I mean, you had a similar story, having been technical and served, and then building a less technical domain.
00:04:44.240 --> 00:04:44.560
Yes.
00:04:44.800 --> 00:04:46.160
Oh, still the thunder here.
00:04:46.319 --> 00:04:48.079
But we have a very, a very similar background.
00:04:48.160 --> 00:04:49.519
I also grew up in special forces.
00:04:49.680 --> 00:04:54.240
And as we figured out right before starting recording, we also grew up in the same little town.
00:04:54.319 --> 00:04:56.000
So I know his brother very well.
00:04:56.319 --> 00:05:03.360
Yeah, he's uh he knows my brother, and it seems to be something in this series because we've we're already three people from the same little town in Israel in this.
00:05:03.920 --> 00:05:04.800
The only source family.
00:05:05.040 --> 00:05:06.160
The only source family.
00:05:07.759 --> 00:05:11.040
That's a good transition to what exactly Limey is doing today.
00:05:11.279 --> 00:05:16.399
One of the core theses is that AI agents will dominate the web, which I think we're also on that wavelength.
00:05:16.560 --> 00:05:20.720
And so what does that future look like and how does Limey play in it?
00:05:21.279 --> 00:05:52.639
So I think one of the things that happened kind of like in the past three years is that we understood that ChatGPT is not, and of course, like the equivalents, Cloud and Grok and everything, you know, all the AI search engines and essentially AI in general is not just us talking to a chatbot, but it's actually going to fundamentally impact how we see, communicate, get information, consume information, execute some tasks, and not only from a consumer perspective, but also from a B2B perspective, from businesses perspective, they're transitioning tremendously.
00:05:53.040 --> 00:05:56.800
And at the beginning, it was just like, yeah, how do I show up on chat?
00:05:56.959 --> 00:05:58.639
Like that was kind of like you know our focus.
00:05:58.720 --> 00:06:02.000
Like, how do I make sure that I can make brands, products show up on chat?
00:06:02.079 --> 00:06:19.439
But as we kind of like got even deeper and deeper and started working with Fortune 10s and Fortune 100s, and we started understanding what they actually need, got even deeper into their into the layer of where these AI agents communicate, we understood that this is fundamentally a much bigger change than just, you know, let's talk with a chat bot.
00:06:19.519 --> 00:06:22.720
But it's actually a complete different way of how we're getting information.
00:06:22.959 --> 00:06:35.600
And if you think of it just like a quick scenario, is that you go out, you go to ChatGPT asking for the best running shoes for women, running a marathon is an example, just goes out to websites, reinterprets everything, and writes it down for you, and you get the answer.
00:06:35.680 --> 00:06:36.639
You might trust it, you might not.
00:06:36.720 --> 00:06:37.600
You might go to the link.
00:06:37.759 --> 00:06:41.439
About 40% of searches end now with no clicks.
00:06:41.519 --> 00:06:43.600
So people are just starting to stay within.
00:06:43.839 --> 00:06:47.279
60% of the web traffic is AI agents.
00:06:47.519 --> 00:06:54.240
So really starting to see that these AI agents are starting to control essentially how we maybe think a little bit too.
00:06:54.560 --> 00:07:16.480
And when we came about it, we understood that we need to be and need to build something around that layer, that new networking layer that helps brands essentially control their communication with their end consumers, control how their brand is being represented, control what these AI agents are saying about them, and also building the infrastructure for these agents to communicate with brands.
00:07:16.560 --> 00:07:23.680
So think about it as like kind of like accessibility for agents, being able to access information in the best way possible and then to reinterpret it.
00:07:24.079 --> 00:07:32.000
So if I I don't want to read too much between the lines here, but walk me through a bit what's the the package for theoretical customer reviewers.
00:07:32.160 --> 00:07:39.439
Do they they probably care about how they're perceived, probably decision-making support if it's a B2B use case?
00:07:39.839 --> 00:07:47.279
And then you do that by somehow putting a layer that the agent can interact with, the customer's layer can interact with.
00:07:47.759 --> 00:07:52.240
Yeah, so kind of like how our how we work with our customers on is usually on three layers.
00:07:52.399 --> 00:08:02.480
One of the main things that every CMO and every marketer wants to know is essentially what he needs to do, what he needs to optimize for to drive the most revenue to the company.
00:08:02.639 --> 00:08:09.759
It can be clicks or users coming in or Reddit links or GitHub stars, whatever, right?
00:08:09.839 --> 00:08:18.240
But he needs, you know, all these CMOs and marketers need to make sure that they're the first, you kind of like the shining star on the keywords and prompts that people are searching.
00:08:18.399 --> 00:08:24.959
So the first layer that we give them is the ability to actually know what people are searching about them, literally like know what their target audience is searching.
00:08:25.040 --> 00:08:28.800
And this is like at its core, it's like the most valuable thing for them right now.
00:08:28.959 --> 00:08:38.639
On top of that, we're giving them the ability to see what are the sources that are being referred to by these LLMs once someone asks a question, like you know, what their target audience says.
00:08:38.799 --> 00:08:47.440
And the third layer is that we give them the ability to make enhancements and recommendations and implement, sorry, recommendations but AI agents autonomously.
00:08:47.600 --> 00:08:49.679
So everything just happens on the go.
00:08:49.759 --> 00:08:54.240
You know, once they get a little bit of confidence in the product, they can enhance their content on their website.
00:08:54.320 --> 00:09:03.039
Everything can be changed constantly based on the data, but also on any third-party websites they would need to publish in, everything is very much executed automatically.
00:09:03.440 --> 00:09:08.639
But with the premises that it will influence the perception of the customer side agents, okay.
00:09:08.960 --> 00:09:10.559
With the premise that they're going to be shown.
00:09:10.639 --> 00:09:15.360
And one of the things that we've added recently is also the ability to tie that back into revenue attribution.
00:09:15.519 --> 00:09:19.200
So say, okay, we've optimized for, I don't know, 50 new prompts.
00:09:19.360 --> 00:09:26.960
Now we're showing up on these supposedly target audience prompts that we needed to focus on.
00:09:27.120 --> 00:09:28.799
How much has that resulted in sales?
00:09:28.879 --> 00:09:30.639
And how much does that really have an impact?
00:09:30.720 --> 00:09:38.080
And some of the brands that we work with have been seeing 8 to 12% of the revenue coming in from chat from due to optimizations they did with us.
00:09:38.159 --> 00:09:40.720
So it's been kind of like a wild ride recently.
00:09:42.240 --> 00:09:56.480
So how do you maintain control while not having control essentially over one, the agents, maybe two the LLMs, and three the competition, meaning there's other folks that are also tackling this problem that are trying to gain control.
00:09:56.559 --> 00:09:58.879
And it's it can be a little bit ambiguous right now.
00:09:59.360 --> 00:10:03.440
Yeah, I think ambiguous is a wonderful world for what's going on here now.
00:10:03.679 --> 00:10:05.840
So there were in terms of the control.
00:10:05.919 --> 00:10:11.279
So I think the shift is so big that they're, you know, everyone's trying to take a little bite, you know, everyone wants a slice of the pie.
00:10:11.679 --> 00:10:23.759
And one of the things that a lot of the brands, you know, what some of our competitors did, and you know, they tried to, you know, kind of like give something in the beginning was they're using what's called clickstream data.
00:10:23.919 --> 00:10:34.320
And that clickstream data is essentially how me and you or anyone in the world is interacting on their browser, client-side browser, meaning, you know, the code that is loaded on our Chrome or Safari or whatever.
00:10:34.559 --> 00:10:39.360
And that clickstream data is referring to how humans are interacting essentially with the web.
00:10:39.519 --> 00:10:50.799
So they took all that data, they put it there, they tried to overlay it on agents' decisions and agents' logics on how to surface things, and then they just gave kind of like a similar what I said, but based on different data.
00:10:50.879 --> 00:10:54.480
So analytics, recommendations, and potentially also the execution.
00:10:54.720 --> 00:11:09.039
In order to gain back the control and the chaos, we understood that because us as humans and AI agents have a complete different logic, and we go to the browser client side, but the agent has no access to open a browser, it's actually going to the server side, communicating through the CDN.
00:11:09.200 --> 00:11:10.480
We should actually be right there.
00:11:10.639 --> 00:11:16.480
You know, we should actually be where these AI agents are actually actually interacting with websites and content and digital assets.
00:11:16.720 --> 00:11:24.480
And the way that we gave them back the control to these, to the CMOs and marketers that we work with is being able to understand the agent's logic.
00:11:24.639 --> 00:11:26.159
Like, why am I going to this website?
00:11:26.320 --> 00:11:27.759
Why am I not going to the other one?
00:11:27.919 --> 00:11:29.600
What am I fetching from your content?
00:11:29.919 --> 00:11:31.759
What piece is intriguing me?
00:11:31.919 --> 00:11:34.159
What is interesting within your content?
00:11:34.240 --> 00:11:37.600
What am I attracted to on your website or any third-party websites?
00:11:37.759 --> 00:11:42.559
And all of that knowledge is being essentially converted into actions they can take.
00:11:42.720 --> 00:11:47.120
So there's a clear roadmap to any marketer to say, okay, this is what's happening.
00:11:47.200 --> 00:11:54.799
This is what AI agents in my sector are attracted to based on the prompts that my target audience is searching, and this is what I need to do in order to make sure that I'm top performing.
00:11:55.039 --> 00:12:04.159
So this kind of like, you know, gets them the ability to be a little bit more relaxed, come to their board meetings and say, hey, we're now showing up better.
00:12:04.559 --> 00:12:09.279
How do you find historically marketeers in in B2B love a lot of them?
00:12:09.519 --> 00:12:12.639
They're not the most sophisticated marketeers in the market.
00:12:13.279 --> 00:12:19.840
How do you find these conversations around introducing the idea of agentic search, agentic positioning?
00:12:20.000 --> 00:12:21.200
How do they react to that?
00:12:21.519 --> 00:12:26.240
That's a really good question because that's something that we constantly battle within our marketing team.
00:12:26.399 --> 00:12:27.679
Like, what are we going to say?
00:12:27.840 --> 00:12:30.960
Are we gonna say business to agent or agent-led growth?
00:12:31.120 --> 00:12:33.200
Or are we just gonna say show up on ChatGPT?
00:12:33.360 --> 00:12:35.200
You know, it's like a constant battle that we have.
00:12:35.600 --> 00:12:36.399
How simplified do you mean?
00:12:36.639 --> 00:12:38.159
Yeah, like how simplified it is, right?
00:12:38.240 --> 00:12:43.519
Because like, you know, when we talk to investors, it needs to be these big headlines like agent-led growth and business to agents.
00:12:43.600 --> 00:12:46.559
But you know, when you go up to a marketer, like loves a buzzword.
00:12:46.720 --> 00:12:47.440
Yeah, right.
00:12:47.679 --> 00:12:48.399
You guys love it.
00:12:48.639 --> 00:12:54.240
But like when you go out to customers, you know, they don't care about agent-led growth, right?
00:12:54.320 --> 00:12:57.039
They care about the bottom line revenue, and it's exactly what you're saying.
00:12:57.279 --> 00:13:01.919
One of the things that we understood is there's there's levels of understanding, also in different markets, by the way.
00:13:02.080 --> 00:13:07.600
Like we sell in AIPAC, in the United States, in Europe, and Israel, and each market understands completely differently.
00:13:07.679 --> 00:13:10.879
Sometimes in the US, it's prone to different regions, like you know, San Francisco.
00:13:11.120 --> 00:13:13.759
Agent-led growth is like the best term you can put out there.
00:13:13.919 --> 00:13:17.759
But like in different places, you know, it's just like, hey, let's make yourself on Chat GPT.
00:13:18.240 --> 00:13:29.840
And I think there are a lot of marketers recently understood that they have to make the transition to have a much more sophisticated tech stack, and they need to know more.
00:13:30.000 --> 00:13:39.200
And they cannot just go to traditional places because if you think of it, you're if you're a marketer, right now, someone's searching on Google, it's getting AI overviews, right?
00:13:39.279 --> 00:13:43.279
So it's pushing the ads down, it's pushing the SEO down.
00:13:43.519 --> 00:13:46.559
So ads are not getting clicked on, so they're costing more.
00:13:46.720 --> 00:13:48.559
SEO is not as relevant as possible.
00:13:48.639 --> 00:13:52.080
So people are seeing incredible drops in their organic traffic.
00:13:52.320 --> 00:13:53.840
So keck is going up.
00:13:54.080 --> 00:13:58.000
Organic is not showing as much as they want to, like not performing as much as they want to.
00:13:58.159 --> 00:13:59.519
They need to like do something.
00:13:59.679 --> 00:14:01.200
So the I I guess you guys see it too.
00:14:01.279 --> 00:14:08.000
A lot of guerrilla marketing that's been happening recently, a lot of like use of like AI agents, growth hacking, their things they're doing.
00:14:08.480 --> 00:14:09.600
Kind of like a reinventing.
00:14:09.679 --> 00:14:13.519
They reinvent what organic means is just evolving.
00:14:13.759 --> 00:14:18.559
You can see it's word of mouth used to be human word of mouth, now it's like some hybrid.
00:14:19.120 --> 00:14:19.840
Google word of mouth.
00:14:20.080 --> 00:14:23.039
Yeah, or like agent-human combo word of mouth.
00:14:23.279 --> 00:14:23.679
Absolutely.
00:14:23.840 --> 00:14:29.600
I mean, if agents are making the decisions of what to surface, you know, you you want them, you know, you want to be in their mouth in a way.
00:14:29.679 --> 00:14:29.919
Exactly.
00:14:30.159 --> 00:14:37.759
Have there been any hacks or tricks that you found that have mapsites or products more desirable for agents that you're advising your customers on?
00:14:38.000 --> 00:14:42.480
So that's a question like we're getting a lot right now is how do we make our website or our product agent ready?
00:14:42.559 --> 00:14:49.600
And if agents are surfing the web, we want to not only come up first, but we want to make sure that there's a click, there's a pr, it's it's the follow-through.
00:14:49.759 --> 00:14:57.039
And I feel like this whole GEO space started off to your point of just like a dashboard, or not just a dashboard, like this is how you're showing up at ChatGBT, this is what you need to do better.
00:14:57.200 --> 00:14:59.120
And I think that got commoditized quickly.
00:14:59.279 --> 00:15:03.840
And then I think where all of these companies are going, it's a few different directions right now.
00:15:04.240 --> 00:15:05.360
Yeah, a hundred percent.
00:15:05.519 --> 00:15:09.279
I think there's two things that kind of like really impact.
00:15:09.440 --> 00:15:12.240
One is like first understanding like what are people searching, right?
00:15:12.320 --> 00:15:15.279
Like you really want to know essentially what your target audience is searching.
00:15:15.600 --> 00:15:24.960
But that has also some limitations because whenever I'm searching on ChatGPT, what happens is that ChatGPT translates that into something that's called a query fan out.
00:15:25.200 --> 00:15:30.159
So I write best running shoes for men running a marathon with knee pain in Boston next year, right?
00:15:30.240 --> 00:15:31.600
It's such a long prompt.
00:15:31.759 --> 00:15:36.960
So, you know, so chat doesn't really take all that prompt and and looks for for specific answer to that.
00:15:37.039 --> 00:15:39.200
So what it does, it kind of like reinterprets it.
00:15:39.360 --> 00:15:41.120
And the way that that's called a query fan out.
00:15:41.200 --> 00:15:43.759
So it might say running shoes men reviews 2026.
00:15:43.919 --> 00:15:44.720
You know, simple as that.
00:15:44.879 --> 00:15:46.480
Marathon, running shoes reviews.
00:15:46.639 --> 00:15:49.679
And that's what it's going out and sending the agents to go and find sources on.
00:15:49.840 --> 00:15:55.600
And the way that these agents are actually finding sources is that they're translating that word into a vector and looking for similar vectors.
00:15:55.679 --> 00:15:57.360
So they're doing vector searches.
00:15:57.759 --> 00:16:05.519
And I would say that the main that that you know, when I always get asked, like, what what is the one thing we need to do?
00:16:05.679 --> 00:16:16.639
So I always say, like, it unfortunately there's no silver bullet, but what the first place I would start at is like first understanding what my target audience is searching and what vector it's being translated into.
00:16:16.799 --> 00:16:26.799
Because if keywords were a very simplified process in Google of like, okay, you need to have these keywords, they need to have some amplifications, they need to have backlinks or anything that SEO was oriented.
00:16:26.960 --> 00:16:35.679
Right now, attracting the agents requires to have a very similar vector space to what is being searched and to what is on your website, so they would be attracted to it.
00:16:35.919 --> 00:16:41.120
I know it's very kind of like a vague answer in a way because you know it's it's a complicated task.
00:16:41.360 --> 00:16:43.919
So that would be my best advice.
00:16:44.480 --> 00:16:52.639
Is the future world where it looks like implementing that almost on the fly in a dynamic way once you have trust with the marketing teams?
00:16:53.120 --> 00:16:55.759
Yeah, so today we work with over 200 brands.
00:16:55.840 --> 00:17:07.519
I think about 20% of them are Fortune 100s, then all these brands are, I think maybe 10% of them are automating, like everything's completely automated, you know, brands that have like very, very strict brand guidelines.
00:17:07.599 --> 00:17:15.200
We work with pharmaceutical manufacturers, you know, it's like for them to say something that's out of balance or out of guardrails is very, very complicated.
00:17:15.279 --> 00:17:18.480
And they're feeling very comfortable to automate every process.
00:17:18.720 --> 00:17:25.519
And I think in addition to that, if you really think of the way, even like how you browse around the web, right?
00:17:25.599 --> 00:17:29.039
Like you like, how often do you go to websites today?
00:17:29.200 --> 00:17:34.799
Like, you know, you get an answer in Chat GPT, you might go to to the link to to vet it.
00:17:34.880 --> 00:17:36.400
I don't think anyone's doing it anymore.
00:17:36.559 --> 00:17:42.799
You know, people are just looking, you know, we're pretty much staying within these LMs, they're becoming centralized hubs of information.
00:17:43.119 --> 00:17:48.160
And I think the future of the internet is not us just having beautiful websites.
00:17:48.240 --> 00:17:58.240
I mean, these websites are still gonna exist, but there's just gonna be some content databases in the world that they're just going to be, you know, information's gonna be taken from them.
00:17:58.480 --> 00:18:11.039
And if marketers were like, yeah, I'm not uploading this blog post, I'm not writing this, this doesn't fit my brand's guidelines or brand kit or whatever, right now that doesn't really matter because AI agents are taking it, reinterpreting it, and putting it out in front of your audience.
00:18:11.359 --> 00:18:13.039
So I think it's very much changing.
00:18:13.440 --> 00:18:22.640
I'm curious, do your customers talk to you about a lot of what we're hearing about the next step will be this new infrastructure for ads in generative formats.
00:18:22.880 --> 00:18:30.079
So chat is the most obvious ones, but also you have generative media where you can insert logos or ads or hints.
00:18:30.319 --> 00:18:32.319
Is that something that's coming up from your customers?
00:18:32.640 --> 00:18:34.000
So yeah, it's coming up a lot.
00:18:34.160 --> 00:18:40.400
Like they're definitely, you know, they want to see how they're they can monetize, you know, these LLMs.
00:18:40.640 --> 00:18:43.759
Chat Perplexity looked at it like November 24.
00:18:43.839 --> 00:18:45.519
They were like checking, testing ads.
00:18:45.599 --> 00:18:47.359
They like they took a step back.
00:18:47.519 --> 00:18:51.359
OpenAI, as you guys know, recently released uh, you know, their ad play.
00:18:51.599 --> 00:19:00.480
They haven't accepted a lot of people yet into their kind of like the equivalent of Google ads, you know, to give people the ability to actually perform ads.
00:19:00.799 --> 00:19:02.319
But yeah, we're getting that all the time.
00:19:02.400 --> 00:19:09.839
Like everyone is understanding they need to be there, everyone's understanding that you know they need to monetize that channel somehow.
00:19:10.240 --> 00:19:11.920
The costs are pretty significant.
00:19:12.079 --> 00:19:21.599
I don't know if you saw, but it's about sixty dollars for impressions, which is like three times more than LinkedIn, which is right now the cossiest platform.
00:19:21.680 --> 00:19:23.039
So it's quite significant.
00:19:23.200 --> 00:19:27.119
And it's impressions, not click, which is a kind of like the Google model.
00:19:27.759 --> 00:19:32.559
How do you think you and your approach are gonna be positioned in that world?
00:19:33.119 --> 00:19:37.119
The ad world, there is two ways to to go about it.
00:19:37.279 --> 00:19:43.359
There is the closed systems, like meta, that's like, what do you want to target?
00:19:43.599 --> 00:19:49.039
We'll find it, we're gonna give us the copy, the algorithm runs, you know, we do our thing.
00:19:49.119 --> 00:19:52.799
You know, you just fine-tune your c your content to make sure it's getting picked up.
00:19:52.960 --> 00:20:04.480
And there is the second kind of like approach that more of like the Google approach and you know, the publisher's approach, which is like this is the this is the ability, this is the keys to the Bugatti.
00:20:04.799 --> 00:20:06.400
Take it and drive wherever you want.
00:20:06.480 --> 00:20:14.000
Like you can go wherever you want, you can build the content, you can decide what to bid on, you can decide how much to bid, like everything is very open.
00:20:14.319 --> 00:20:24.079
From what I'm seeing, you know, we also have a couple of customers that are in this, like kind of like POC with OpenAI around the ads, it's going to be a bit more closer to Meta.
00:20:24.240 --> 00:20:31.920
So we won't have the ability to run the bidding because they're going to do the bidding, and the bidding is kind of like where, you know, it's the most interesting place.
00:20:32.000 --> 00:20:33.279
Like, what prompts am I bidding on?
00:20:33.359 --> 00:20:36.079
What prompts I'm putting more money on be based on their performance.
00:20:36.240 --> 00:20:37.519
They're not going to tell you exactly.
00:20:37.599 --> 00:20:40.720
It's they're just going to run that for you, and you're going to get the end results.
00:20:40.960 --> 00:21:05.039
But when you're running a campaign, one of the most significant things is not actually the campaign itself, but it's the research before the campaign, what's happening throughout the campaign, to know what to, what campaigns to take down, what campaigns to keep, what campaigns to put more efforts on, to create more and duplicate them, and also the post, like kind of like the post-mortem of the campaign, like you know, researching about it, what happened, what we can do.
00:21:05.200 --> 00:21:07.759
We come in kind of like throughout all these phases.
00:21:08.000 --> 00:21:13.680
We cannot do the bidding, but we can help our marketers make the right decisions on on what to do.
00:21:13.759 --> 00:21:16.799
And that's exactly where the money is actually really lies in.
00:21:17.200 --> 00:21:21.119
Being a founder, we we sometimes get lost in the problem we're solving.
00:21:21.359 --> 00:21:24.319
I find that there's always a really interesting story around the motivation.
00:21:24.559 --> 00:21:33.039
Why are you so obsessed with the marketeers and how people are gonna be doing decision making and buying in in this new age of agentic search and web.
00:21:33.440 --> 00:21:39.359
Maybe walk us through a bit what influenced you into dedicating some of your prime years to this problem.
00:21:39.920 --> 00:21:41.119
Prime years, I love that.
00:21:41.200 --> 00:21:42.079
That's good.
00:21:43.359 --> 00:21:47.759
I think there's two things that kind of like came to mind when we started the company.
00:21:48.079 --> 00:21:55.759
First is that we want to work with a target audience that suits us, like that we're we're fitted in, like we're welcomed in, I would say.
00:21:55.920 --> 00:21:59.839
Like if I go to different markets, you know, insurance markets, like I don't know.
00:22:00.240 --> 00:22:01.599
I'm not gonna be a great fit there.
00:22:01.839 --> 00:22:02.880
You don't give the insurance vibes?
00:22:03.119 --> 00:22:04.799
I don't give the insurance vibes at all, you know.
00:22:04.960 --> 00:22:07.680
Like I would love to, but I don't give my the insurance vibes.
00:22:07.839 --> 00:22:09.519
I think marketers accept me much more.
00:22:09.759 --> 00:22:13.599
So first we wanted to go to a place we'd be welcomed in with the people that would be enjoying working.
00:22:13.839 --> 00:22:16.400
And I think marketing is like so much creativity.
00:22:16.480 --> 00:22:20.079
It's like incredible people, things are just flowing, all the money's there too.
00:22:20.160 --> 00:22:25.279
So like it's quite an incredible department to work in at any company, I think.
00:22:25.599 --> 00:22:29.839
And you know, at the beginning it was it's very much that, okay, let's let's help marketers.
00:22:30.240 --> 00:22:40.720
But as kind of like we we started deep diving more and we understood that it's actually much more deep into how information is being communicated to people, we understood this is like a fundamental shift.
00:22:40.799 --> 00:22:54.000
And this is like an area where we can not only help build the infrastructure for the next way of how the world communication is going to go through and run through, but also really help influence it.
00:22:54.400 --> 00:23:02.079
And being at that junction of where content and information is being delivered through, I think is one of the most exciting things in the world.
00:23:02.240 --> 00:23:05.039
So I would say this makes my years prime, right?
00:23:05.119 --> 00:23:11.119
Like not sitting at the beach or doing like the fact that I'm doing this and dealing with these incredible problems.
00:23:11.279 --> 00:23:14.079
I think this is what makes my my life prime.
00:23:14.400 --> 00:23:19.440
What do you think is the biggest misunderstanding that current generation marketeers have about AI?
00:23:19.680 --> 00:23:25.039
So I think the the younger folks are like they're understanding everything.
00:23:25.200 --> 00:23:32.880
Like they're very much understanding, they're very much out there, they know what they want to do, they see AI, they see it evolving, they love playing with the products.
00:23:33.039 --> 00:23:40.880
I think like some of the marketers that I work with, I'm 29, you know, some of them are like 21, they're much more advanced than me, and either incredible, they're doing super cool things.
00:23:41.039 --> 00:23:51.759
I think sometimes where it gets stuck is is either kind of like higher management or you know, kind of like the companies that are having a hard time going through that rewiring.
00:23:52.319 --> 00:23:55.759
Applying old frameworks of budgeting to new methods.
00:23:56.079 --> 00:23:56.480
Exactly.
00:23:56.640 --> 00:23:59.759
Like, you know, SEO is like irrelevant now.
00:23:59.839 --> 00:24:02.640
Like what are you still like we're not trying to focus on something else?
00:24:02.720 --> 00:24:07.119
You know, it's like it's these things that you have to like try to walk them through in understanding.
00:24:07.200 --> 00:24:15.759
And you know, sometimes we see it, but you would think that such executive, experienced people would have like the ability to make these transitions really quickly.
00:24:15.839 --> 00:24:25.119
Sometimes it's not only up to them, but we are seeing a lot of the time that there's a lot of like hesitation, stagnation around like things that we find hard to understand why it's taking so much time.
00:24:25.359 --> 00:24:43.200
And I think sometimes a lot of AI adoption is is complicated, especially around marketing, because for one, it's potentially the first or the second largest expense in the company, which is you know, you don't want to shuffle everything around, you know, space at a place where you know money is like so impactful.
00:24:43.359 --> 00:24:57.839
And the second is that this is what drives impact and shifting and changing things around that place, you know, around the revenue, the leads, what are we we we're getting a consistent, repetitive funnel that's working beautifully.
00:24:58.079 --> 00:24:59.200
Why should we change it?
00:24:59.440 --> 00:25:02.880
You know, it's like it's hard to make the decision, it's hard to rip off the band-aid, I think.
00:25:03.119 --> 00:25:06.160
What do you see as the moment typically for Fortune 100?
00:25:06.960 --> 00:25:09.279
Once it monetizes for like three or four months.
00:25:09.440 --> 00:25:11.359
Like once they see it, can we actually monetize?
00:25:11.440 --> 00:25:14.559
And this is not just like, oh yeah, like it's it's just showing up on chat.
00:25:14.640 --> 00:25:15.119
Like, who cares?
00:25:15.200 --> 00:25:16.319
It's a branding thing.
00:25:16.480 --> 00:25:27.359
I don't know if you know, but like one of the things that le that companies least invested in was like branding technologies, like social intelligence technologies, like don't really care about it.
00:25:27.519 --> 00:25:31.519
Just like outsourcing it and using like services for it, or yeah, they never really used any software.
00:25:31.680 --> 00:25:32.160
Like who would pay?
00:25:32.480 --> 00:25:33.119
It's the first to go.
00:25:33.200 --> 00:25:34.880
It's hard to justify, it's not it's in downtown.
00:25:35.200 --> 00:25:35.599
Exactly.
00:25:35.759 --> 00:25:37.680
And right now it's like the most tangible thing.
00:25:37.839 --> 00:25:39.680
Right now, that's exactly what you want to do.
00:25:39.920 --> 00:25:44.160
It's funny, but like things kind of like really shifted on 180 degrees.
00:25:44.480 --> 00:25:53.279
In full transparency, before starting my my company, I used to run uh global operations for a public company, so all of marketing operations was under my responsibility.
00:25:53.359 --> 00:26:01.200
And I always thought of the marketing space as like an this weird approximation, impressions, clicks.
00:26:01.440 --> 00:26:03.599
Really, what we're trying to get to is just intent.
00:26:03.839 --> 00:26:10.880
If you could tap into someone's brain and understand what's going on there, what are their fears, what are they looking for, what are they trying to accomplish?
00:26:11.359 --> 00:26:13.440
That's the pure thing.
00:26:14.079 --> 00:26:22.880
I think fundamentally when the medium is more conversational, you essentially bypass a lot of that, you know, bullshit that we used to engage with in the past.
00:26:23.359 --> 00:26:31.839
Do you think these modern platforms that you're also building in are gonna like just transition to that pure intent signal?
00:26:32.240 --> 00:26:34.960
I think the word approximation is perfect.
00:26:35.279 --> 00:26:54.640
I think that there's a lot of approximations that have been done in the previous world around the previous Google world around revenue and attribution and clicks, and and you know, you go to your HubSpot, you see one number, you go to your Google Analytics, you see other number, like everything is like no one really knows.
00:26:54.799 --> 00:26:59.039
So an approximation was a big part of like of ops essentially in marketing.
00:26:59.279 --> 00:27:04.640
I think that right now we're not trying to find human intent, but we're trying to find machine intent.
00:27:04.799 --> 00:27:09.440
And I think finding machine intent is becoming much more crucial in that world.
00:27:09.599 --> 00:27:19.200
So machine intent is much more logical, much more comprehensive, much more structured than human intent.
00:27:19.359 --> 00:27:28.559
And that's why it makes it a little bit easier to decrease approximations and to have literally rooted accurate numbers.
00:27:28.880 --> 00:27:35.680
Just for the fun of it, I do recommend you one time read through the developer traces of clot code.
00:27:36.240 --> 00:27:37.680
It feels very irrational.
00:27:37.920 --> 00:27:38.160
Really?
00:27:38.400 --> 00:27:41.200
The traces sometimes go like, I should have not said that word.
00:27:41.279 --> 00:27:43.279
I wonder what the human thinks about me.
00:27:43.519 --> 00:27:45.519
Maybe I'll go back and do it again.
00:27:45.759 --> 00:27:50.240
Or maybe the human will think I'm indecisive, so I'll randomly reverse us not.
00:27:50.559 --> 00:27:53.839
Yeah, but like just it's actually it sounds freakishly human.
00:27:54.160 --> 00:27:55.599
Right, but that's the LLM, right?
00:27:55.759 --> 00:27:55.920
Yeah.
00:27:56.319 --> 00:28:03.759
And the way what we're actually trying to influence is not the LLM itself, which is freakishly weird thinking about weird things.
00:28:03.920 --> 00:28:05.359
It's actually what's called a retriever.
00:28:05.440 --> 00:28:12.240
That's what goes out and finds these sources and hands off to that freakish beast that you just mentioned to write the answer.
00:28:12.799 --> 00:28:17.599
Are you building Limee in New York in person today, or is it still a hybrid in Israel?
00:28:18.079 --> 00:28:19.519
So we have two offices.
00:28:19.759 --> 00:28:22.319
Actually, potentially opening a third one very soon.
00:28:22.480 --> 00:28:23.279
We have two offices.
00:28:23.359 --> 00:28:26.079
One is in Tel V where product engineering and design is.
00:28:26.240 --> 00:28:32.640
You know, our CTO is from 8200, our incredible team is all from there, designers, the developers.
00:28:32.720 --> 00:28:35.440
It's really like the top of the top of engineers.
00:28:35.519 --> 00:28:36.559
We couldn't wish for better people.
00:28:36.799 --> 00:28:37.839
And the shekel is so attractive.
00:28:38.079 --> 00:28:39.599
And the shekel is attractive, right?
00:28:39.920 --> 00:28:40.960
The main reason.
00:28:42.240 --> 00:28:44.960
But it there were times that it wasn't so attractive.
00:28:45.200 --> 00:28:49.519
And our go-to-market is here in New York, but we also have a couple of West Coast hires.
00:28:49.599 --> 00:28:58.559
So we are a little bit hybrid in a way, but I think like once you want to take over a market, you have to be in the market.
00:28:58.640 --> 00:29:00.160
You can't just, you know, be up far.
00:29:00.400 --> 00:29:02.640
That's why you're also opening an APAC office soon.
00:29:02.799 --> 00:29:06.559
But you know, really kind of like to make sure that we're in the place where our customers are.
00:29:06.640 --> 00:29:10.000
You know, we want to understand when they're send their needs, especially around the marketing.
00:29:10.240 --> 00:29:12.319
And you've relocated recently to New York.
00:29:12.559 --> 00:29:12.799
Yeah.
00:29:12.960 --> 00:29:14.079
How's that process been for you?
00:29:14.240 --> 00:29:15.119
I love Williamsburg.
00:29:15.200 --> 00:29:15.359
Yeah.
00:29:16.079 --> 00:29:17.200
I thought I will never like it.
00:29:17.359 --> 00:29:19.039
Honestly, as a visitor, I never liked it.
00:29:19.279 --> 00:29:21.279
All Israeli founders end up in Williamsburg.
00:29:21.359 --> 00:29:21.519
Yeah.
00:29:21.680 --> 00:29:22.799
Or Upper West Side.
00:29:23.039 --> 00:29:26.000
Or Upper West, but I think it's just like Williamsburg is very Tel Aviv.
00:29:26.400 --> 00:29:26.559
Yes.
00:29:26.720 --> 00:29:26.880
Yeah.
00:29:26.960 --> 00:29:27.519
Very Tel Aviv.
00:29:27.680 --> 00:29:27.759
Yes.
00:29:28.000 --> 00:29:29.519
Like short buildings, you know.
00:29:29.839 --> 00:29:31.839
Do you want to wrap with a few questions before we close out?
00:29:32.079 --> 00:29:33.119
Let's do it.
00:29:33.519 --> 00:29:36.960
Does brands still matter when agents are intermediating everything?
00:29:37.279 --> 00:29:39.279
Yes, like matters most, I think.
00:29:39.599 --> 00:29:44.400
What is a KPI that a CMO should potentially stop tracking next year?
00:29:45.200 --> 00:29:45.759
Clicks.
00:29:46.240 --> 00:29:49.200
If Limey failed tomorrow, what do you think would be the reason?
00:29:49.440 --> 00:29:50.480
Us founders.
00:29:50.720 --> 00:29:51.359
The founders?
00:29:51.519 --> 00:29:51.759
Yeah.
00:29:51.920 --> 00:29:52.160
Okay.
00:29:52.480 --> 00:29:53.680
I think full accountability.
00:29:53.920 --> 00:29:54.960
Always founders.
00:29:55.279 --> 00:29:56.480
This was an awesome episode.
00:29:56.640 --> 00:29:57.359
Thank you for coming on.
00:29:57.680 --> 00:29:58.400
Thank you so much, guys.
00:29:58.559 --> 00:29:58.720
Thank you.
00:29:59.039 --> 00:29:59.519
So much fun.
00:29:59.759 --> 00:30:00.960
Really appreciate it.
