%20(1).png)
October has been a month on the road for the Relevant Digital team. We started in London at the AOP Publishing Tech Talk, before heading to New York for Prebid Summit and DanAds Summit, and wrapped up the month at IAB Sweden Programmatic Day in Stockholm. Thank you to everyone we met along the way! In this month's recap, we bring together two topics that came up again and again in our conversations, and that could shape how publishers monetise their inventory in the coming years.
First, Ronny Linder explores the proposed changes to Prebid auctions following the ruling in Google's ad tech antitrust case, and what a more unified auction could mean for publisher yield optimisation. Our CEO Petri Kokkonen then shares takeaways from conversations with publishers on agentic advertising: how AI and automation are being used today, and where the real value lies right now.
Both topics point in the same direction: publishers will have more options, more data and more automation available to them. The real advantage will come from having the control and the tools to act on them.
Prebid, Google and the Future of Publisher Yield Optimisation
Following Prebid Dev in late September, our team also attended Prebid Summit in New York in October. Both events offered valuable opportunities to connect with industry experts, exchange ideas and hear fresh perspectives on where programmatic advertising is heading.
One particularly interesting topic at this year's Summit was the potential evolution of Prebid auctions following the recent remedies ruling in Google's ad tech antitrust case. While Google will not be required to divest its AdX exchange, the court has ordered Google to enable Prebid integrations and give publishers greater access to bidding data. The ruling focuses on open web display advertising, but its implications for publisher monetisation could be significant.
How Google's Integration Could Change Prebid Auctions
Today, Prebid typically runs its auction before passing the winning bid to Google Ad Manager (GAM), where it competes with other demand sources, including Google's AdX. This structure has long raised questions about competition, particularly because Google controls the final auction and has visibility into the winning Prebid bid. Under the court's remedies ruling, Google must enable AdX demand to participate in Prebid auctions independently of its ad server, and provide publishers with better access to winning and losing bid data. However, the court has not specified exactly how to implement these integrations.
To help address this, a Prebid working group has proposed a new auction model that would allow ad servers to participate directly in Prebid auctions as "Adserver Bidders". This could allow direct-sold demand to compete alongside programmatic bidders in the same auction. Under the proposed model, Prebid Server would handle auction decision-making, while also supporting publishers using Prebid.js. The framework would be open to any ad server, not just GAM, potentially giving publishers greater freedom to choose their advertising technologies and reducing their dependence on a single ad server. The proposal is still under discussion, and Google has not committed to the proposed architecture, but they are in active development dialogue with Prebid.
What Could This Mean for Publisher Yield Optimisation?
For publishers, one of the biggest opportunities is greater transparency into how demand sources compete for their inventory. Today, separate auction stages make it difficult to get a complete picture of bidding behaviour. A more unified auction environment, combined with access to Google's winning and losing bid data, could provide a clearer view of competition and greater control over auction decisions. This would create opportunities to evaluate pricing strategies, test configurations and optimise yield based on more complete information.
Greater transparency alone, however, will not automatically translate into higher revenue. The real value comes from acting on the data. At Relevant Digital, we've long focused on giving publishers the tools to manage, analyse and optimise their header bidding operations. Relevant Yield already combines Prebid and GAM data in a unified reporting environment, helping publishers compare demand performance, test different configurations and make informed optimisation decisions.
The proposed changes could also improve auction efficiency by reducing the need for separate auction stages, potentially lowering latency. More timely bidding data, combined with automated testing and optimisation, could help publishers respond faster to changes in demand and pricing. Ultimately, the significance of these developments will depend on how they are implemented. For publishers, the opportunity is not simply a different auction architecture, but greater independence, more effective competition between demand sources and better control over advertising revenue.
- Ronny Linder - CPO, Partner and Member of Several IAB Tech Lab Task Forces
Agentic Advertising: From Industry Vision to Publisher Reality
The agentic dream industry organisations are painting is interesting. In the future, where an advertiser briefs a strategic agent on campaign goals, and after that this chief agent starts working with an army of channel and data specialist agents, creating campaigns and constantly optimising everything is fascinating. In this scenario, humans on the advertiser and publisher side focus on overseeing and leading agents; agents do all the operative work. Right now, this is still a dream, but it might be true in the future.
How Publishers Are Using AI and Automation Today
Based on talks with the UK's leading publishers, agentic is on everyone's mind, but real-world use cases are more about targeted solutions where companies introduce automation into parts of sales and ad ops processes. This can mean, for example, connecting 1st-party data platforms, inventory and pricing information through MCPs to LLMs to speed up the "from brief to proposal" timeline and improve proposal quality.
AI is also implemented quite often in the ad ops workflow, aiming to remove all the manual steps of taking the data and creatives from some sources to other systems, for example, to ad servers, invoicing, and BI systems. Of course, automation has existed long before the current hype, but AI tools make it easier, and building automation can be fairly low cost when solutions are built in-house.
Who Will Provide Agentic Solutions for Publishers?
From a sell-side ad-tech company's perspective, it is always interesting to think about who should provide agentic solutions for publishers. This depends on which part of the sales process we are talking about, but when agents need to have access to inventory, ad servers and order management systems are natural places to add an agentic dimension. These systems already sit next to the inventory, either through deep integrations or as the inventory management engine. From a programmatic perspective, a sales agent can integrate with an SSP and can be provided by one. In this scenario, an agent can, for example, offer advertisers a better tool to establish new deals with suitable buyer targeting and support deal troubleshooting.
Familiar ad-tech giants will provide more and more of their agents and agentic tools to publishers. Using those might be the easiest way to start learning, but there is also some healthy scepticism towards out-of-the-box solutions. It is good to be mindful and understand whether those tools are transparent and unbiased, and whether you want to give the partner more access and data. The alternative, of course, is to build agents in-house or ask independent providers for help.
Agentic Direct vs. Agentic Programmatic Advertising
There have been talks about whether agentic will be more direct or programmatic; this has been one of our favourite internal coffee-table debates, too. We can stop arguing about that; it will be both. We'll see two new revenue channels: agentic programmatic and agentic direct.
MCP servers are business as usual, and new standards keep coming out, but one interesting thought came up in the room with publishers: Do we even need all these protocols for every scenario? Maybe bigger publishers could plug their assets directly to the agency's buy-side agent? Relevant question, and might make sense for bigger publishers. But with how many publishers, would agencies do this? Is there a risk that smaller publishers would be left out?
Where AI Delivers Value for Publishers Today
There is a lot of excitement around implementing AI and agents to sell-side workflows and the actual sales. But also healthy realism. Right now, the biggest wins are on the process efficiency side and integrating data sources and systems into one engine. Short-term revenue uplift is most likely still on the machine-learning side of AI, for example, through setup and floor-price optimisation.
- Petri Kokkonen - CEO & Partner, Member of the IAB Finland Board and Co-Lead of the IAB Europe Programmatic Working Group