Great. Okay, we're gonna get started. I'm Laura Martin. I'm the Senior Media and Internet Analyst at Needham & Company, and I'm here with Adam Singolda. Let me just introduce him. He's our next speaker. Adam is the Founder & CEO of Taboola, where he has built a global platform that serves billions of content recommendations daily across the open web. Prior to launching Taboola in 2007, he honed his deep technical and leadership skills over seven years as an officer in the elite mathematical and computational science unit of the Israeli National Security Agency. Let me see. That might be it. And that's it. And we're gonna start with a question about leadership. Generative AI is a transformational technology. I think that's a consensus view.
What kind of pressures does that put on you as a leader? What changes when you have transformational disruption underlying the tech stack of your business?
Good to be here. Thank you for having me, always, and I look forward to dinner tonight. I think, you know, we've been in a state of mind of change for quite some time now. I mean, if you think about, you know, from the pandemic and recession and tariffs and wars, We haven't had a boring day in a long time, you know? I think it already in many way kind of pushed, you know, companies and leaderships to change and evolve so that they can deal with all this madness that's happening around the world globally. You know, I think we talked about that last time, about, you know, transparency and how do you speak to people and what they expect of you, during these days.
We've been in this war mode for a long time now, one. Two, I think this is so big, you know, the change of AI.
that you have to almost reimagine who you would be if you were born now, and then ask yourself what are the gaps between that version of yourself to how you look right now and build bridges towards that new future. You have to imagine if you started Taboola now, and we'd be a $2 billion revenue business, open web, all the things that make Taboola special. What, you know, today, native AI company, what would we look like?
That's the exercise that, you know, we're going through with the team. We're constantly asking ourself what do we need to do, not to be better version of ourself, but almost to reimagine a new Taboola and then b, have the courage to speak about it, to build processes to get there. We just had a board meeting, and our COO and President, you know, who owns that project, this internal AI rebirth.
presented to the board all the things we're doing internally. This is not product. I'm just talking about identity, culture. You know, how do you look like? What organizations go away and change? What you don't have? How do you communicate with each other? What type of information people have access to. It's amazing. I mean, we now have over 80% of our code has something drafted by AI in it.
We have now hundreds of apps. There's a marketplace.
of apps people build using Claude internally so they can do things way better. I can now build an app that can query different things in the database and present it to professional services, and now I just saved so much time, and we are all better. Hundreds of apps. Employees are so curious. I think to me, it's mainly about, as always, about people and culture.
You know, especially when you go through massive transformations like we do with AI, you have to be even more transparent.
even more honest with yourself about what works and what's not.
Have the courage to reimagine yourself and build a plan to become that company.
One of the most interesting things that's going on in the workforce organization is there's this work being done, and it's theoretical, because it for an academic to study it, somebody has to have done it, and then they look at the data. It hasn't happened yet. The idea is that maybe functions are degrading, that you can't have separate sales from product, from engineering, from marketing the way almost all companies are set up, that instead we're gonna have these SWAT teams of like five people with one product guy, one engineer, one marketing person, one salesperson, and they're gonna have a team, and there'll be a deliverable for that five-person group or maybe 50-person group, but the point is in six months you need to have this product up and running.
These teams work together with deliverables, and they rise and fall together. Like if they don't hit the deliverable, they're all out of a job, and if they do hit the deliverable, they get the next project. Functions need to break down in a way, like not think of themselves as sales versus marketing, where somebody says, "That's not my job.
Right
These jobs become redefined by AI. Do you have a point of view on that?
I don't know if it's gonna be just that. I think it's a lot of that. I think it's so liberating to be a salesperson or an account manager and have so much power in your hands to do something you used to do, you know, manually.
That used to take you hours and not even wait for someone to do it for you. Just do it yourself.
You know, when I wanted to check DeeperDive, one of our AI, you know, product for publishers, I wanted to see how often does it happen that we ask a question, the answer is not to my liking, and how many times certain things happen. I didn't email anyone. I just asked Claude to take authorization over my browser, I said, "Here's usage today. By the time I wake up, I'd like you to test it 10,000 times, and here's what I wanna know." We had a daily meeting about DeeperDive. When I came into that meeting, I was a super version of myself. No one knew more than me. I knew the most. I had 10,000 people-
Claude had taught you?
Yeah, me and Claude. You know, it's like the Avengers, you know? We were just like Marvel through and through.
Iron Man.
Yes. Maybe. I don't know. It ended bad for him.
Yeah
It's so liberating if you, if you have the curiosity and courage to lean in. I cannot imagine not spending two, three hours a day coding
I was just with a CEO and I gave him that line. I said, "I have CEOs who are spending four hours a day on Claude." He goes, "I would love to do that. I have to manage people." I'm like, "Well, a lot of these CEOs-
I mean, I'm.
are really software guys.
Yeah, I mean, That's a very scary answer if I were him. I don't know who would has the courage to say that. That's a very. I'm not naming names. I don't know who that is. That person may not have a job in the future.
I won't let him know.
No, no, don't Yeah, no one Yeah, this is just between you and me. Nobody needs to know.
Okay, nobody has to know.
Yeah.
It's a secret.
even if that's what you think, don't say it loud. Just, you know, lie about it. Just, you know, make, like, something up.
Mr. Honesty during disruption.
Yeah, like what are-
You're out telling a lie.
What are you doing? What are you doing, you know, saying that, you know, your old self is better than your new, machine-fused version? You have to do it because I'll tell you why. If you don't build things yourself, not your teams telling you how it works, then you don't.
Know its capabilities or its limits.
You don't have a gut feeling.
Right.
You know, so much about leadership is about developing, you know, arts and science. You know, you wanna have data-driven, you know, kind of making decisions that matter, but you also wanna have a gut.
You know, something needs to feel right or something needs to feel wrong, and you can't have that if you haven't, you know, got dirty with it yourself.
Yeah. No, that's fair. Okay, I saw you 2 weeks ago at the POSSIBLE Conference in Miami. Let's begin where we left off. Can you talk about the Realize+ platform and what does it do for Taboola, you know, sort of in 2026, but as you look forward in the next maybe 24 months, how does Realize+ accelerate your growth rate?
Yeah, Realize+ is our version of Google Performance Max or our version of Meta Advantage+. It's an agentic AI.
Wow
bridge for advertisers who want to have less control and more just outcomes, you know, driven kind of system. I'll give an example. If you are a subscription service that's only relevant for New York, you want clients in New York City.
Yeah
and you know exactly who they are, and you know that you want them on, when they use their iPhone and when they read about certain things, and you have a very specific opinion about what audience you're trying to get, you should use Realize. 'Cause you know what you want. Open five campaigns. Open a mobile New York campaign. Open a strategy for if they read about this, show them the ad. Whatever you want. You know what you want, go realize your potential.
Right.
If you don't have that level of, you know, kind of definitions of what you want to get and your bidding strategy is not so defined in your head, you just wanna know, "I wanna sell this table.
I know that so long that it's $50 or less, I'll buy as many ads as possible from Taboola.
You can go to Realize+, give us that goal.
Realize+ will essentially open and shut down campaigns on your behalf in real time, 24/7, and will report to you know, what works and what's not. This is really exciting because a lot of clients wanted us to give them, like, a P Max option.
You know, they wanted us to say, "It's nice we're spending with you maybe $100,000, but just so you know, if you had a PMax, we would have another budget with you.
that goes on top of what we already do." To me, the exciting piece about it is that I look at it as a whole net new budget opportunity with advertisers.
Okay.
It's a way to get money we never got from people who just want results, and they don't want that level of control. That's, you know, it's new.
As always, I will measure that like I measure all products, you know, in terms of will it reduce churn rates, will it increase spend, and will it result in more scaled advertisers working with us.
which should impact our financials. You know, we've seen an accelerated growth so far this year, and the more we get budgets from advertisers, the more we can continue our pathway to double-digit growth rates.
Why is it we only track scaled advertisers? Why wouldn't you care? Why do we care about whether more scaled advertisers work with you versus just more advertisers?
Well, scaled advertiser, which is basically anyone spending more than $100,000, just means that that advertiser found market fit with us. They spend enough, so the likelihood they'll churn is lower significantly.
I see.
The likelihood they'll increase spend is also higher. For us And also by the way, it's the vast majority of our revenue.
For us, we look at that number, and the more it goes up, the more it's a good proxy that our revenue is stable, can continue to go up. We do, and we share it with investors because we know they wanna look at the same thing as well.
Mm-hmm. Yeah, and lower churn is definitely a big deal. Okay. Okay, what have you learned? Realize was a year ago at Possible. You had just launched Realize, right? This year was-
Right
Realize+. What have you learned in the year about Realize? I know that we were trying to open up the display market and get a bigger share of wallet. What have you learned in the 12 months since launching the original Realize that you didn't know, like last time we were on this stage 1 year ago?
Yeah. Well, I think it's working. It's progressing. You know, there's a lot of work ahead of us, but it's progressing in the right direction, and there are many things on the Realize kind of strategy that needed to happen. One of them was Taboola to become a performance advertising platform, like AppLovin for apps, Taboola for the open web.
outside of search and social. That was a brand progress, and I think we're progressing. You know, we're going to the right places. We're talking to the right people. There was sales kind of restructuring to make sure that we're going after segments of the market where we know the chances for success are higher. We call that ICP.
We've hired Krishan from Amazon to kind of lead our revenue initiatives globally, as part of that kind of motion. As you said, we made it much easier to give us any format you may have, and this was a supply and demand kind of effort. On the supply side, it meant going to publishers and saying, "Give us that display inventory too.
Give us on your apps more." Basically, "Give us all the inventory that you have.
Okay.
Let us participate in all of that.
Okay.
On the demand side, it meant that no longer did we force you to change for us. If you had display, you can give us display. If you have native, give us native. If you have social, give us social. If you have vertical, give us vert. Whatever you have, we'll take it.
Okay.
The last one was a new feature that was part of Realize called Predictive Audiences.
Okay.
Which essentially said, if you spent some money with us.
You've got some conversions.
Yeah
we can predict how much money do you have to spend further.
to get more conversions, which advertisers love.
Okay.
I guess much like investors.
advertisers like predictability and growth.
Yeah.
When you tell an advertiser like, "You're spending $50,000 with us. If you give us $10,000 more.
Here's how much it's gonna cost you per conversion.
before you spend the money.
Yeah.
They like it, that became a very well-adopted new kind of capability for us, yeah. I would say overall, we're seeing progress across, you know, utilization of the features, sales restructuring, market brand and perception kind of improvement, and with Realize+, it's another step forward, making it even easier. I know we're gonna talk about Claude at some point during this Fireside Chat about our work with Claude. I will say in general, overall, all we're trying to do is make it really easy for advertisers
To use. Mm-hmm.
to succeed, much like it is on Meta and Google.
Okay. Good. Let's go to Claude now. I know you said at Possible that you've created agentic tie-ins, and I think there's a definitely a difference of opinion among CEOs in ad tech about whether we're gonna be on new rails for agentic or on the old programmatic rails. Why don't you tell people what you've been doing with Claude and agents, and also I think explain your thesis about programmatic rails are gonna be different and replaced by agentic rails.
First, let me just say I think I'm right on this one. I'm gonna tell my opinion.
Okay.
I just want full disclosure. This is one of those obviously moments. You ready for this?
What?
I'm about to say, like what's gonna happen. I'm gonna give a prediction here at Needham.
Okay.
Okay, a year from now we're gonna look back and say We're gonna replay this, just this part of our conversation.
Okay.
Okay. Here's what I think is happening.
Okay.
It will happen really fast.
Okay.
With Claude now, essentially anyone, advertisers or even more importantly, holding companies.
agencies can basically build buying agents.
that can interact with search Google directly.
via an MCP, Google just released one, or with Meta to buy social.
Tell them what an MCP is.
MCP is a protocol that allows agents to talk to each other.
Right.
Basically Google says, "If you wanna buy from us, you can open PMax and buy your campaigns.
We now have a way for your agent to talk to us. Here's a protocol available for you.
As of now, you can talk to us through this protocol so you don't, you never have to come here. Your agent can talk to my agent.
Google just made one available. Taboola made one available, and you can go even a step further and say, "Actually, if you're using Claude, I'm gonna build a skill for you that if you turn it on.
you can talk to Claude in your own language, and Claude will understand how to do it." We're gonna make Claude really smart as relates to buying from us. Google did it. We did it. At that point, Claude essentially can buy search directly, social directly, Taboola directly, and when it buys it's not buying it programmatically. CPMs, splash, you know, spray and pray, hope for the best. It's buying the specific platform in an intimate way. Claude says to Taboola, "I wanna sell tables. Talk to me." Taboola will say, "Well, do you have a goal in mind?" Claude will say, "$50." Taboola will say, "Do you need it in any specific way, or should I just go and do the best I can?" Claude will say, "No, no, it has to be desktop users. I want these demographics.
It has to be East Coast, that's what I want. By the way, they have to have read something about this topic over the last 30 days." Taboola MCP will say to Claude, "Great, you need Realize. I recommend you spend $10,000 test campaign." Claude will negotiate. They negotiate, they negotiate, they negotiate. The campaign is live, we're on. There's another alternative.
Who made the ad? What ad are they using?
Taboola will ask Claude, "Do you have creative, or do you want to make one for you?" Claude will say, "My client doesn't care. Do it for me," or, "My client has very specific needs. Here's the creative." All of this will happen from a Claude software. The client or the agency have never left. At that point, you can build an agent that just bought Google, just bought Meta, and just bought Taboola open web, and in fact, you know, Paramount has API self-service. You can even buy TV. At that point, your agent, which costs you $200 subscription a month, can buy search, social, open web, and TV, bypassing the entire thing. I'm not even talking about saving cost.
I'm talking about the fact that when you use the platform native APIs and conversation versus programmatic, you're using all the goodies that the platform has to offer you. Let's simplify this further. If you buy from Taboola programmatically, Taboola, which is live on ESPN, Fox Sports, CBS Sports, Yahoo Sports, we know sports. If you like the Knicks, which I am, Taboola knows about it. You've read someplace about the Knicks. Taboola knows it. DSPs don't know it. When they buy from us programmatically, programmatic doesn't know what I know.
They're just saying, "Give me something at that price.
They can match cookies maybe.
They don't know what I know.
When you buy from me directly, you can say, "If someone read about the Knicks, offer them NBA subscription.
I know it. Only I know it. When you buy from Claude, the agent is not just doing it perhaps more affordably than a very expensive programmatic buying.
We're talking billions of dollars of saving. Billions, maybe tens of billions. It also buys it much more intimately. It asks Taboola, "Do you know someone who read about the Knicks on a sports site today? If so, I want them to see the ad." Only Claude can do it. Programmatic is-
Why couldn't OpenAI, ChatGPT, and Gemini?
They can also do it. Any one of those agents can do it.
Okay.
Programmatic cannot do it.
Can't talk.
It ca-
Can't interact.
There's no there It's a dumb pipe.
It's a dumb pipe.
It's a very efficient pipe.
Yeah.
It did a lot for our industry. It's made a lot of good things happen. It reached the end of that journey because it's not smart enough.
It's uniquely expensive. To my team, I refer to programmatic as an old knee, you know, that needs to be replaced with something to allow me to run a marathon, and that is kind of. Now, when you look at those things, which is the only thing that moves industries since the beginning of time, has always been greed and fear.
Otherwise, I'm not changing.
Yeah. That's true.
If you're an agency, and you just gave this ad tech industry $5 billion, that's greed.
If you think someone else is building an agent getting better results for the client because you're using agent versus programmatic, that's fear. Now you're looking at $100 billion a year agency business just in the U.S. with a lot of greed and a lot of fear.
They are really slow.
Greed and fear accelerate change. Billions is a great number, so I just think we may see really fast change. That's, in any case, I think that's where I think it's happening.
Yeah.
How fast, we'll see.
I think, agent to agent, I wanna own a big part of that market.
Okay.
You know. That's, to me, the open web needs a seat at the table. A good seat.
Well, I think you have unique supply, which is what you need in that world, right? 'Cause Claude isn't coming to you unless it can get something from you it can't get anywhere else.
It is so simple and so true. Unless you have supply no one else has.
Unless you I'm talking exclusive rights, not bidding on that supply.
Yeah.
Claude can bid, too. I'm talking it's mine.
Unless you have data that Claude cannot get.
Okay
Claude is getting it. Claude is bypassing you. It has to. It will. I'm so, you know, looking backwards, I guess we never knew of LLM, you know.
revolution. I mean, now I'm looking at Taboola with 11,000 publishers, 30-year Yahoo, decades of relationships with 600 million people every day reading about stuff on the web. It's a very unique sense, you know, intent data.
It's very unique supply relationship.
Realize is a very good engine.
The best performance engine on unique supply and data in the world, I think there's really something beautiful that can happen here.
You represent publishers. Why can't Claude just buy from the publisher? Why does he need the Taboola as an intermediary at all in the new rails?
Claude is going to try to make the client happy. The client wants performance and outcomes.
For performance and outcomes, we're just the best. You know, you can't call any one company and say, "No one has enough scale to offer what a client wants." If I'm a retailer, I want to spend a lot of money getting a lot of outcomes, so I'm gonna need someone that has a lot of reach, a lot of scale. When it comes to outcome, if it's a branding campaign, maybe I want a specific thing, like to be on the homepage of someone. If you want to get a million clients to buy something, you come to Google, Facebook, and you come to us, 'cause we're very good at it. You come to I mean, I'll give you an example. For us, AI is completely correlated to value.
As an example on Deeper Dive, which is our publisher answer engine, you can go to USA Today and ask a question and it uses AI. We host it. It uses AI to get an answer and things to read, and we show ads. We make money on that thing.
Yeah.
We give it for free. The publisher doesn't have to pay, you know, OpenAI APIs, which are very expensive.
Okay.
We host, we train, we build a RAG, we do the whole thing. We give it to you for free, we drive engagement through it.
Yeah
We even pay you for it because we put ads on it.
Right
like Gemini and Google.
Oh, right.
Right. In this case to me, yeah, I give it to you for free.
It costs me money, but I'm making more than I spend. In fact, I'm sharing that upside with you, and everybody gets excited. It's a differentiator in the marketplace. Same for Realize and Realize+. I'm using a lot of AI. I'm not charging directly for that, but if I do a good job, prices go up. You pay me more because I'm able to get you affordable conversions. You wanna spend more with me and increase your price because you really like it.
On the advertising side, it's all directed to value.
Okay.
I do a good job, you give me more money. I don't do a good job, you leave.
Okay. All right, you're gonna get it through the take rate. Yeah, you're trying to get higher either through the take rate or just more spending with the same take rate.
Well, it's not, the advertiser has no rate. It's, like, it's about CPC and spend.
Yeah.
Do they increase the price because they love what they get or no?
Do they spend more money with me or do they churn?
I look at a client and I'm saying over time, it's very simple: Are they staying?
Are they spending more?
Bigger part of their-
Yeah, that's it.
Yeah.
If the answer is they're not staying, bad.
If they're not increasing budget, they're not happier.
Yeah.
You know? That's all I wanna see.
Okay. All right. Fair enough. The bottom line is you're not up-charging for the AI. You're just trying to get people to churn less or spend more money.
Right. I just, I wanna grow my business-
Right
at the same, EBITDA margin which we're able to do, and we're using AI to get more value for the client.
through DeeperDive
Yeah
or the client through Realize and Realize Plus.
Right. Okay. Ad buyers are demanding more transparency, large publishers are gaining negotiating leverage. Does this mean there's gonna be take rate pressure on your business do you think over time, or not?
I mean, to me it's the opposite.
Okay
more we do a good job.
Okay
for advertisers Taboola is a two-sided marketplace.
Yeah.
We're very unique in that place. We're like, you know, we're like a walled garden outside of the walls.
Yeah
'cause we have supply and demand like a Google and Meta.
'Cause most companies in ad tech space are either supply focused or demand focused.
Okay.
We're both. For us, if I look at the demand side, the only thing my clients care about is was it working.
Yes
Is it stable and scalable? Can I do more of it so I spend more money with you?
Okay.
They, I have to provide value, and so long that I do it.
I get more budget. If you look at our financials, you're seeing more scaled advertisers.
Yes.
Average spend goes up
growth and our financials are improving.
Okay.
That's the only thing we focus as a team.
Okay.
The clients just wanna know, they know what they're trying to get 'cause they compare us to Meta and Google.
It doesn't matter. They don't care about our publisher rev share.
Right.
They're just like, "At the end of the day, how much does it cost me to get a client?
Yes.
Is it good or bad? Based on that, they make decisions. We have a publisher rev share which I think is fair. Like, I don't expect it to change, you know, in a material way over time. I think we wanna be a significant revenue to publishers. We're proud of that.
Okay.
We're not looking for ways to, you know, get another, you know, kind of up our game there. We, in fact, wanna be a bigger share of wallet, pay publishers even more.
Okay
over time. I think, you know, when I look at the business, if we do a good job implementing technology, we gain growth rates.
Yeah
Hopefully even, you know, leverage on our margins over time. It's the opposite in my mind, and a lot of it's thanks to being a performance advertising platform.
It's not a wine and dine business.
No, that's true.
It's not a please like me, try me, none of that. I mean, I hope they like us, but I mean, at the end of the day, they only stick around if it works.
Yeah. Usually that's the metric for small and medium businesses. Do you find I know you launched a platform I think a year and a half ago that was really targeted to small and medium businesses. How has that adoption gone, Abby?
It's still, I mean, it's our self-service, which Abby is mostly relevant. It's still small. It's growing. It's nice. I love that business.
I will tell you, even bigger advertisers, look, they need to trust you. They need to know, you know, a lot of time there's a management kind of like interaction. You know, I can tell you now advertisers, big ones and agencies, they really like our AI vision.
Okay.
They like-
Oh
they like how disruptive this can be.
Okay
to their existing way of doing business, and they wanna get close to us.
Okay.
That's great to see because it means they give us a shot.
Okay.
However, that shot is all it is at that stage.
Okay.
From that point on, it has to be good for them.
Okay.
That's, it's what I like about Taboola where we have to be good.
Okay.
You know, we have to deliver real value to advertisers and publishers, they wanna stick around and grow with us. I do think that bigger advertisers, to your point, at times they do wanna connect to the company vision and, you know, disruption and, you know, being on the right side of growth.
Eventually it has to work.
Do you think that walled gardens benefit when we think about the gen AI disruption? The walled gardens have been really fast adopters of sort of these LLM tools to lower their costs and to accelerate sort of their revenue growth by faster time to market for products. Do you think that gen AI benefits the open web, or do you think the walled gardens are the big beneficiary of gen AI tools?
I think it depends. gen AI can benefit everyone if depends where you sit on the spectrum. If you're an open web, a bigger brand, a known brand. You have low search exposure, and you're now at the beginning of AI adoption.
Which means you can make a lot more money and you can engage consumers a lot more.
Yeah.
You're gonna benefit from AI.
Okay.
If you're a small publisher with a, you know, with a high search exposure, you probably carry more risk.
Okay.
If you're an advertiser or an agency, if you think it's not important to develop agents yourself.
If you don't try to disrupt yourself by leaning into LLM and this agent economy.
I know.
You're, you better be right.
You know, because your competitors will definitely try something else.
I think it's a cultural question. I think everyone has something to gain, so long that they have ideally unique data-
distribution access
Okay
The right culture to fail fast and be okay changing.
Well, I have companies that are doing these integrations, these AI integrations, trying to bring these companies up to 'Cause I don't think the senior management has any kind of technological expertise, so they have to hire companies to bring them in to make agents.
That's fine.
Yeah.
I mean, I think you should do whatever you need to do. I mean, if you, if that's something, you know, you. Some people are closer to the tech side and they don't need it.
Yeah
They need sales help.
'cause it's gonna impact everything and so long that you're not waiting to see what happens.
Yeah
I think you're probably doing the right thing. Yeah, I do think you have to lean in to get really to dream big as relates to who would you be if you were born now.
Yeah, build the bridge to that.
You have to.
You know, start over. Because you know what? There's guys dropping out of Harvard that are making something from scratch with 5 guys.
Exactly.
You know, and letting agents do all the work.
I had a meeting with my team before this, and we talked about the future job of QA, quality assurance.
Okay
someone who's testing a product.
Yeah.
In the past, which it's such an important job. Someone has to test a product before it goes to the client.
Yeah.
In the past, that person did a lot of things. With Claude, if that person does anything more than once, Claude should do it because it's repetitive, it's deterministic. We know what we expect. The future job of a QA is almost like an artist. It's someone that needs to have a vibe and you know, emotions about your product and give you how does it feel to use your product.
As relates to testing things and see how, what happens in a repetitive way, Claude should do it.
Everything is changing. If you start your company now-
what I just said would be obvious to you. You're like, "Why would I have a QA person.
checking the product works?
That's a Claude job.
Claude can only predict the past. It can predict things that have happened.
Yeah.
It cannot predict the future because it never happened.
Right.
a human-
Yeah
can.
Yeah.
This is where-
if it feels clunky.
Yeah
feels too many steps or it.
This is it. This is where humans shine.
Yeah.
It's the feeling, it's the gut, it's the emotion, it's the art.
I think if you understand that.
you know, I told my team the threshold.
Yeah
for a human is really high.
Otherwise, if Claude can do it, Claude should. It's very obvious to new native.
Mm-hmm. I wonder if what you're saying, though, is you're gonna employ less STEM people and more liberal arts people? If we're moving towards art and emotion.
I don't know.
It feels like we're moving out of computer science and data analytics skills and because Claude's gonna do those, and we're moving into things like, "I don't like this pink. It makes me feel stressed.
I think.
I don't like this combination of
I think there's a lot of value in people that are architect of things because they can have more critical Even what to tell Claude to do and to negotiate with Claude about, yeah. There's value in that as well.
Okay.
The composition is changing. We're having people we didn't have before.
Yeah.
I think this is exactly Look, I came from Israel. English to me is a second language.
You know? I will never.
You're pretty good at it.
I will never say R in a perfect way.
Yeah.
You can tell I'm not from here. That is basically all companies. We all are second, you know, we're all expats.
We're all expats.
'Cause we-
To gen AI.
To gen AI.
You're right.
We're all.
is it gonna be natives that are competing against.
We are all not from this country.
Yeah.
None of us speaks English the right way.
The gen AI, the gen AI English.
Yeah. Another question is who's practicing their English? You know, I went to Starbucks to buy coffee 10 hours a day so I can feel comfortable talking English when I move to this country.
Right.
You know? Who is doing those things so that AI can be somewhat native to them?
Yeah
because you're not from this country.
Right, right.
We all know you're an expat.
I'm not that.
Yeah. This is the reality.
Yeah. That's what's very funny. I think about how, I think it's really interesting how this gen AI affects labor forces, labor force organization, labor force demands on what is required out of the ecosystem. The more Claude can do in the STEM world, the more you're gonna rely on liberal arts majors, which was not foreseeable 3 years ago.
Yeah.
You know. I think that's really interesting.
No, this is.
Maybe more engineers, which are problem-solving kinds of mindsets.
Yeah. Look, there's growth. I think there's growth in jobs and hiring.
Really?
Yeah. I think because you're gonna want, if the same person can create much more value to the client, there's more growth rates and there's more EBITDA and there's more cash flow. At the end of the day, we're all here to build companies that gain market share, growing faster, and generate profit.
Yeah.
You know? I think this can be If your employee base is evolving-
Yeah
into this native AI, you know, better R speakers.
Yeah
They're all awesome, and you could get more of them. Those who don't, like again, if I Don't tell me working for Taboola you're not building, you're not building agents. Lie to me. Don't tell me this to my face.
My favorite line from Possible was your line where I said, "When are you cutting people? Why don't I see your FTEs going down?" You said, "Laura, when people come to Taboola, they don't come for three months. They come for 10 years.
Yeah.
I'm gonna find stuff for them to do.
There's a lot of work to do.
There's a lot of work to do.
Taboola is a very hyper-transparent company. We speak about all of these uncomfortable topics all the time at all levels. I have an offsite with my team in June. I am telling 2,000 employees what is my agenda.
Today.
Tell me what I'm missing?
when you leave in May.
Tell me what am I not doing? I think this is a great time to celebrate culture and people.
Yeah. Okay. Great. Any questions from the audience for Adi?
I hope you're building Claude agents.
Yeah. Don't tell him you're not spending four hours a day on Claude.
Lie to me in any case. Yeah.
Lie to him, please. Lie to me. Okay. Any questions for Okay. My last question there, since we have 2 minutes, what is the thing that you think Wall Street or investors are missing about the Taboola story that we should be focusing them on as you walk out the door?
I mean, I honestly think they want us to grow faster.
Yeah.
We're working hard on it. Some people like us more, some people like us less.
There's a lot of growth in advertising. It's a growing market. It's an exciting market. Everyone is getting into it. OpenAI is getting into it. Netflix got in. Everybody wants an advertising business.
Yeah.
everybody needs a friend, and Taboola is a better friend than Google.
Cause we're always gonna be on your side.
Okay.
We only grow if you grow. Google is like, you know, they grow if they grow. I think we have, you know, we sit in a good place. However, we're currently, you know, our guidance now is 8%. You know, that's not why I'm here, and that's not why they want to buy the stock.
Tell Claude I'm mad at him. He needs to grow faster.
I know. I am.
We're yelling at Claude.
I think as we get to nine% and 10% and 11% and 12% and, you know, 15% and 20% consistently organically at our EBITDA margin.
Yeah
technically it's a very exciting upside, if you believe.
I think Wall Street mainly wanna, you know, those who are early, you know, have higher risk profile getting now.
Going the way up, and those who are less, they talk to us, they track us. I think what they want is faster growth rates, and that's what I want too.
Okay. All right. That's it. Okay, I will call it there. Thank you very much.
Thank you.
Thanks, everybody. Thanks. That was great.
You're the best.
You're the best.