Taboola's chatbot ad network

Ad technology firm Taboola issued its latest quarterly financial report yesterday showing slightly higher revenue but lower gross profit in Q4 2025.

Snippet: “Revenues in the fourth quarter were $522.3 million and $1.9 billion for the full year, an increase of 6.4% and 8.3%, respectively.” Read more.

Post-earnings report, the company’s market cap hovered near $860 million as the stock traded lower by ~5%.

Regarding the results, Taboola CEO Adam Singolda summarized them on LinkedIn saying, in part, “Our deep distribution, exclusive publisher relationships, and intent signals are becoming structural advantages in an AI-first world.” Read more.

The automation of ad platforms

On the earnings call with analysts, the company’s AI-enabled Realize performance ad platform was the centerpiece as the company plans on moving toward a more automated advertising approach similar to Google’s Performance Max and Meta’s Advantage+.

Singolda said, “…the biggest thing that we will see from Realize later [this year], for those who should succeed with us, it will be more about automation and making it even easier to drive success in an effort to compete.

Chatbot monetization on the open web

Mr. Singolda also spoke about Deeper Dive which launched in June – it’s both a chatbot product and an ad network:

“…we have a product called Deeper Dive, essentially bringing ChatGPT-type technology to those bigger publishers so that consumers can converse and talk to publishers. If you go to USA Today, you can check it out. I think there is a significant ARPU growth, a significant revenue generation opportunity for publishers when they actually adopt AI on their own sites.

The risk, I think, is more on the smaller sites, which we do not have exposure to, or for those who are very dependent on search — also not a publisher that we work with. I think there is a very bright future for the trusted publishers, especially when they adopt AI in a bigger way.”

Regarding the chatbot itself, Singolda noted the increased visitor engagement that the chatbot drives for publishers. And, he saw something else:

“The second thing we are seeing—and we will share more data about that later in the year—is that when advertisers show up in LLM experiences, the opportunity for them to drive conversion and the CPMs we are seeing are something that I can tell you in fifteen years of doing this, I have never seen before.”

Open Web traffic is growing

On the call, Mr. Singolda said traffic on the open web was going up:

“On the open web, we basically have a very structural advantage in where we sit in the open web. We are seeing traffic going up. We are seeing search traffic going down. But overall, through primarily direct traffic to publishers and then onboarding more publishers, traffic is overall going up.

The exposure we have to search traffic, which I think is the main risk that investors are tracking, for us it is in the single digits. A lot of it is because we work with massive platforms like Microsoft and Yahoo and Apple News. A third of our traffic is in-app.”

Read: Taboola Q4 2025 earnings call transcript (February 25) – Motley Fool

From tipsheet: The chatbot ad network competitive set is still in its early stages. In yesterday’s email, tipsheet covered Koah, a startup focused on enabling “native advertising” in-app — i.e., chatbots.

Google and “Adsense for chatbots” or “Adsense with Gemini” would likely be the largest eventual competitor in the category once it formalizes its approach. OpenAI would seem to be a logical player as well.

Perhaps Meta and Amazon could break out beyond their walled gardens with a chatbot network appended with “native ads,” too

LLMs & CHATBOTS

Developments

“Introducing Perplexity Computer” (February 25) – Perplexity on LinkedIn
Let Gemini handle your multi-step daily tasks on Android (February 25) – Google’s The Keyword blog
Anthropic acquires Vercept to advance Claude’s computer use capabilities (February 25) – Anthropic

AGENCIES

Pitching “proprietary” in an AI world

Touting agentic workflows and orchestration, Adobe and WPP announced a new partnership on agents that will combine Adobe’s AI capabilities with WPP Open, the ad holding company’s AI-enabled marketing platform.

On LinkedIn, WPP CTO Stefan Pretorius broke down the partnership in 4 ways:

“1 – standard product integration patterns between WPP Open and Adobe software

2 – training and deployment of hundreds of creative technology forward-deployed engineers (FDEs) that clients can tap into for innovation

3 – scaling up our business transformation and systems integration capacity to support clients with the adoption and value creation with Adobe software

4 – joint GTM and development of custom brand models using Adobe’s Foundry AI platform”

TECH

Apple’s lurking AI revenue strategy

For all the consternation regarding Apple’s supposed lack of an AI strategy — remember ‘Apple must buy Perplexity’ last summer? — the company’s de minimis ‘capex’ now looks prudent compared to its “big tech” competitors in the public markets.

Taking into consideration Apple’s obsession with privacy on the consumer device, perhaps Apple is waiting for a future where AI — the AI chatbot or LLM — is accessed locally on the device rather than by using a third-party such as OpenAI or Google.

Mobile Dev Memo analyst Eric Seufert sees the seeds of an Apple AI strategy bubbling, as he shared on X yesterday:

“In November, Apple updated its App Store developer guidelines to require developers to obtain affirmative, opt-in consent from users before sending ‘personal data’ to third-party AI services.

A pertinent question is: if Apple is able to dissuade developers from utilizing third-party AI services through the consent requirement, but accessing models hosted via [Personal Cloud Compute] does not require consent, will Apple be able to charge a revenue share for ‘default AI status’ at some point, as it does with default search engine status in Safari?”

Read Mr. Seufert on X. (February 25)

Doubling down in a post on his Mobile Dev Memo blog titled, “AI Tracking Transparency” (intentionally echoing Apple’s App Tracking Transparency), Seufert explains the potential revenue strategy he sees:

“Put another way: if the only way to host a model in PCC is to enter into some sort of partnership with Apple, could Apple charge a revenue share on API-based token usage for the frontier models it allows into its PCC environment? Apple currently pays Google for the privilege of using Gemini, but I imagine that it’d prefer that money flow in the other direction. Apple may be setting the stage for a revenue opportunity in in-app AI usage by creating a consent distinction between models hosted via PCC through a partnership with Apple and those that aren’t.”

Read: “AI Tracking Transparency” (February 24) – Mobile Dev Memo (subscription)

From tipsheet: Great insights, as usual, from Seufert.

It would be a major win for Apple if this gets executed successfully

FINANCIALS

Earnings reports

The Trade Desk Reports Fourth Quarter and Fiscal Year 2025 Financial Results (February 25) – The Trade Desk
Magnite Reports Fourth Quarter and Full-Year 2025 Results (February 25) – Magnite

“Agentic” on the earnings calls:

Jeff Green, CEO, The Trade Desk: “We are convinced that agentic AI will ultimately accrete the most value to companies that already have deep customer trust, that have scaled, refined, and objective data sets, and that prioritize objectivity, not by companies with limited data, hoping an AI framework becomes their business model. In that sense, agentic AI is an evolution of outcome-based platforms, not a shortcut around them.” (February 25) – transcript on Investing[dot]com
Magnite CEO Michael Barrett: “There has been speculation that generative AI and agent-based buying could disintermediate infrastructure platforms. We believe what is actually unfolding reinforces the importance of scaled sell-side infrastructure. We embedded an advertising context protocol, or AdCP, seller agent directly into SpringServe in Q4 and executed what we believe was the industry’s first agent-to-agent campaign. Scope3 served as the buyer agent on behalf of MiQ, with media running across LG and Warner Bros. Discovery inventory. While still early, this marks an important milestone.” (February 25) – transcript on Motley Fool

LLMs & CHATBOTS

Scraping video data helps the chatbot

According to Digiday, video is becoming an LLM bot crawler favorite and, consequently, generative engine optimization (GEO) techniques are increasingly taking into account what’s in (transcripts) and around (metadata, chapters) video.

YouTube is the main video source discussed in the story which speaks to Google’s ability to effectively squeeze text and information out of the audio/visual format.

SEO company BrightEdge told Digiday’s Sam Bradley that YouTube citations are showing up more than Reddit in AI chatbots like Google’s Gemini and OpenAI’s ChatGPT.

Read: “Why YouTube has become key for brand GEO strategies.” (February 25) – Digiday (subscription)

From tipsheet: OpenAI crawling YouTube transcripts could become a battleground between the two AI giants or any AI tech firm seeking to unlock the YouTube transcript trove given Google’s T’s & C’s.

In 2024, the battle almost took place for unlicensed scraping. And then, nothing happened — at least publicly.

TECH

IAS Agent ‘beta’ is driving workflow

Lisa Utzschneider, CEO of ad verification company, Integral Ad Science (IAS), joined the latest episode of the AdExchanger Talks podcast.

Now that her company is newly-private thanks to Novacap’s acquisition of IAS in November, she remarked how freely she can speak compared to her public company days.

Utzschneider addressed the status of the company’s agentic strategy which began with an AI assistant known as “IAS Agent.” Launched in December with promised availability in “Q1 2026,” the agent is a workflow assistant addressing reporting, recommendations and campaign setup.

Ms. Utzschneider said:

“We launched a beta. It was oversubscribed. And I have to say the brands are thrilled with our IAS agent solution, which is basically accelerating the ability to get key insights that they need to drive greater efficiency and greater ROI.

So for example, with IAS agent and the automation, we’re able to get the insights directly into the hands of the brands five times faster because of the automation. And the brands are seeing operational efficiencies of 50% or higher. And this is even just in beta phase.

So the brands are completely leaning into our agentic solution. There’s no other agentic solution like this in the market. And we’ll just continue to fuel our agentic solutions and investments again to drive efficiency and ROI for the brands.”

Hear more on AdExchanger Talks on the Apple Podcasts app. (February 24)

PROTOCOLS

‘Upselling the agent’ and WebMCP

Marketecture’s Ari Paparo, a media entrepreneur with deep experience in ad technology product, reviews the potential opportunities unlocked by Google Chrome’s new, open source, WebMCP protocol.

Announced earlier this month, according to Google, WebMCP “aims to provide a standard way for exposing structured tools, ensuring AI agents can perform actions on your site with increased speed, reliability, and precision…” — i.e. the agents’ website, if you will.

On the publisher-side, Mr. Paparo sees a robust list of possible opportunities beginning with an efficient way to facilitate a content marketplace as well as “upselling the agent.”

He explains, “If your user is using an agent to access your content, why not upsell? Tell the agent your website is better if the user installs your app. Or maybe buy some merch. The agent might pass along some of these learnings to the human, right?”

Paparo, ever the product maven, offers over a dozen ideas across the buy and sell-sides — even “weird ideas” (his words) such as: “Could you more easily upsell the AI on buy-now-pay-later offers or other post-conversion add-ons than a human?”

Read: “WebMCP Opportunities” (February 25) – Marketecture

BRANDS

Automating the CMO function

In Adweek, a new survey by market research firm Newtonx found that AI may be shifting the priorities of the marketing department.

NewtonX’s Daniel Sills said that AI has affected marketers’ day-to-day with 63% of respondents saying “it has already had a moderate or significant impact.”

Adweek reported:

“But for CMOs and their teams, ‘it’s putting more demand to generate actual, measurable growth,’ Sills said.

‘Executive leaders believe in AI’s potential to increase productivity and cost efficiency,” he continued. ‘So the expectations for growth are [increasing] simply because of the perception of what AI can do.’”

BRANDS

Agentio scaling opportunity, not headcount

Agentio, which uses custom AI agents to facilitate campaign workflow that results in “fully contracted Creator programs,” is seeing success according to its CEO.

Yesterday on Linkedin, Agentio CEO Arthur Leopold shared an agentic case study about premium dog food brand, Maev, which wanted to scale beyond its creator program on Instagram – but didn’t want to scale its own headcount — and in to YouTube.

Mr. Leopold shared, “One person. 200+ Creator partnerships per month. 2x-3x increase in Earned Media Value ROI with Agentio, vs. Instagram & TikTok organic creator posts.”

Taboola’s chatbot ad network

The automation of ad platforms

Chatbot monetization on the open web

Open Web traffic is growing

LLMs & CHATBOTS

Developments

AGENCIES

Pitching “proprietary” in an AI world

TECH

Apple’s lurking AI revenue strategy

FINANCIALS

Earnings reports

LLMs & CHATBOTS

Scraping video data helps the chatbot

TECH

IAS Agent ‘beta’ is driving workflow

PROTOCOLS

‘Upselling the agent’ and WebMCP

BRANDS

Automating the CMO function

BRANDS

Agentio scaling opportunity, not headcount

MORE