Ad Tech Insights: AI on the Sell-Side, Deals API and Addressability in Europe

In December, our recap focuses on three developments shaping publishers' priorities heading into 2026: the growing practical impact of AI on the sell-side, early but concrete progress in programmatic deal automation through IAB Tech Lab's Deals API, and new insights into addressability and measurement readiness across Europe.

In this summary, our experts outline what these shifts mean in practice: how AI is affecting traffic, content and operations on the sell-side, what publishers can realistically expect from the first version of the Deals API, and why strengthening first-party data, contextual strategies and measurement capabilities is becoming critical heading into 2026.

•••

AI on the Sell-Side: What Actually Matters for Publishers

It's everywhere: in every conversation, newsletter, podcast and feed. It's a threat and an opportunity, good and bad. And it's not going anywhere; it's here to stay. The industry is continually launching new AI standards and projects. But what matters today and what matters tomorrow can be hard to understand for all of us who work in digital advertising and publishing. So let's recap the biggest buzz acronyms and see if they make sense.

One way to approach everything going on is to observe the landscape from two perspectives: audience and sales. Let's start with the audience.

AI, discovery and the audience problem publishers face today

The negative side is that LLM crawlers are hitting publishers' servers more than ever, in most cases generating unwanted tech costs and sometimes also ad calls. At the same time, LLMs are grabbing part of publishers' audiences with content they have taken from publishers. Double trouble, and an urgent one indeed. Common sense says that this is the AI issue we should solve first.

From a tech cost perspective, publishers should focus on blocking all unwanted crawlers immediately. In some cases, it might be easier to turn this around and allow only humans and a few well-known friendly crawlers, provided you can tell friendly crawlers apart from impostors pretending to be something they are not.

For blocking or allowing only certain parties, the technology already exists. Infrastructure providers like Cloudflare have launched their own solutions, and companies like Gravito offer monitoring and blocking solutions.

Content access, control and monetization

LLM content crawling can also be good for publishers, or at least tolerable, if it brings an audience or money. Once analytics and control are in place, the next step is monetization. We've already been reading positive news about leading publishers signing agreements with leading LLM providers. That's a great start, but we haven't heard any news of smaller, independent local publishers making these deals. Understandable, but worrying.

How will a local publisher sort this out? One option is standardized monetization mechanisms, which are in the works. IAB Tech Lab's Content Monetization Protocols (CoMP) and other initiatives are building solutions to address this issue, but the challenge is that these projects take time, while audiences and content are being taken right now. So maybe the best approach for publishers unable to sign deals with LLM providers is to block LLM scrapers as tightly as possible and revisit this once working, scalable monetization standards are in place.

AI in advertising sales: automation, agents and realism

The other side of the AI coin is advertising sales, and it's a busy time on this front as well. The biggest launch and buzz of the second half of the year has been around the Ad Context Protocol standard (AdCP).

The standard aims to bring a new wave of automation, with agentic AI, to advertising buying and selling. Right now, AdCP mainly focuses on scaling direct sales-type activities and does not aim to replace OpenRTB. The vision is bold and lucrative. What if buyers could scale their direct buys to hundreds of websites without significant manual work?

But everyone who has experience working with sell-side ad tech knows that it takes quite a lot of coffee, cigarettes and complex development before real-life complexity, with changing inventories, campaign updates, workflows, integrated systems and evolving APIs, is ironed out and simplified for sleek agents.

So, my advice for publishers at this point is this: if you have the resources, jump on the wave, join the community and start testing with your tech partners. For those who don't have resources to burn on uncertain pilots, my advice is to follow this development closely and get to know companies that might help you with implementation if or when the time comes. It's fair to say that there is some skepticism in the market, but who knows, maybe we will all be agentic soon.

Supporting standards: UCP and ARTF explained

And then we have the supporting standards, which might cause me some hair loss if I had any left.

First, UCP, which stands for User Context Protocol. Why would we need this? Well, based on IAB Tech Lab CEO Anthony Katsur's release, existing ways to communicate people's intents, interests and the contexts they spend time in do not scale to the upcoming agentic era. We need something lighter, faster, reliable and privacy-compliant, and UCP might be the solution for that.

If I got this right, we are talking about a new language made for AI systems to exchange complex data combinations quickly as simple strings. Like an old married couple, where one partner understands immediately from the other's look that a mistake was made. No words needed. But yes, words-to-math is what we are talking about. Does this bring more ad revenue? Well, not yet. And this sits more on tech providers' tables than publishers'.

And one more acronym: ARTF, also from IAB Tech Lab, stands for Agentic RTB Framework. ARTF is not an AI solution itself; it's more like a standardized playground where AI applications can play by jointly agreed guidelines and rules, running tasks for orchestrating and hosting platforms. A simple example: the host platform could be the publisher's ad server and the agentic service provider an ID provider. There could then be a secure container with ID decryption and matching logic.

In this scenario, publishers wouldn't need to share their own IDs with ID providers; matching is done privately, locally and in real time within the container. In this example, the container could act as a scalable, agile, local clean room running ID matching in real time. Did I get it right? Of course, the ID example is only one of many; there can be a wide range of applications and host platforms utilizing the standard.

What publishers should focus on next

So, plenty of AI-related conversation topics for our holiday tables to fill awkward silences, and things to follow after New Year's. We'll continue to closely follow everything in the digital advertising domain and remain actively involved in industry-level work through associations. A lot is going on, and some trends are more important right now than others. We are always ready to provide advice and help for our customers. If you need help with integrations, setting priorities or anything else, let us know!

Our own AI priorities focus on making the Relevant Yield AI assistant better and better, teaching it new concepts and tasks, and making it more independent and agentic. We are also building entirely new applications to support publishers' operations and sales. Their launches are not too far away, so stay tuned. Once again, an exciting year ahead.

- Petri Kokkonen - CEO & Partner, Relevant Digital, Member of the IAB Finland Board and Co-Lead of the IAB Europe Programmatic Working Group

•••

Deals API: What Should Publishers Realistically Expect from It?

IAB Tech Lab's new Deals API is currently in its public comment period until the end of January.

Traditionally, when creating a deal after negotiation, the seller has had to set up the deal in their supply-side platform (SSP) and send the details to the buyer, who then has had to do the same in their demand-side platform (DSP). Some SSPs and DSPs have already eased this manual process with their own APIs. However, until now there has been no standard way of doing this. Sometimes there has been an API, sometimes manual work, and what has been transferred through APIs has varied.

The Deals API is meant to standardize this process across systems. It allows deal details to be sent from the SSP to the DSP in a standardized way, which will hopefully lead to wide-scale adoption and remove some of the manual work and need for troubleshooting. It allows very minimal two-way communication, where the SSP can inquire whether the deal has been accepted and whether it is active. At least in the first version, it does not facilitate negotiations or alterations to deals, so everything needs to be confirmed beforehand.

So the Deals API is meant to address some, but not all, of the challenges that everyone buying or selling through deals is all too familiar with. It does not revolutionize setting up deals, nor does it guarantee that all deals set up in the future will be adopted by buyers. In the first version, it does not even guarantee the buyer that the deal will stay exactly the same as when it was sent. However, what it could do is standardize how deals are set up and potentially remove some of the manual work often associated with the process. It also creates potential for increased transparency by detailing the curator or other intermediaries, which has not been done before.

Overall, Deals API v1 would be a good first step in improving programmatic deals. It will most likely leave many hoping for subsequent versions with more two-way communication and the ability to make changes after a deal has been established. Nevertheless, it's a welcome first step.

- Tuukka Aaltonen - Head of Programmatic, Relevant Digital, Vice Chair of IAB Finland's Data & Martech Working Group

•••

Addressability in Europe: Key Takeaways for Publishers

IAB Europe's new Addressability Report highlights persistent gaps in how targeting and measurement are implemented across Europe. Adoption of privacy-centric solutions such as unified IDs, data clean rooms and advanced contextual targeting remains uneven, and many publishers still have limited access to cross-platform data. Measurement without cookies remains one of the biggest challenges, with attribution inconsistencies slowing progress across markets.

For publishers, the most crucial takeaway is clear: first-party data and contextual strategies are becoming the default foundation for maintaining performance. As browser-level consent models reduce the availability of user-level identifiers, publishers need setups that can flexibly support multiple ID solutions and stronger contextual signals. The report also shows how wide the knowledge gaps still are. Many teams cannot assess which identity approaches actually improve results, making transparent reporting and testing critical as publishers move into 2026.

This is where platforms like Relevant Yield provide practical support: publishers can compare setups, track the real performance impact of different ID solutions, measure the effectiveness of contextual strategies and evaluate how each configuration behaves in production. In addition to reporting, Relevant Yield's HB Manager offers a wide range of integrated ID partners that can be enabled and tested effortlessly, as well as built-in integrations for contextual targeting, making it easier to experiment and identify what truly works for their inventory.

While no universal addressability framework has emerged, the direction is unmistakable. Success will depend on strengthening data foundations, improving measurement capabilities and adopting flexible architectures that can adapt as signal loss accelerates and privacy requirements evolve.

- Suvi Leino - Head of Marketing, Relevant Digital