For most of the past decade, artificial intelligence in advertising has worked quietly in the background. It has helped platforms predict which users are more likely to convert, adjust bids in real time, optimize campaign budgets and select better-performing creatives. The advertiser or media buyer, however, has remained firmly in control of the process: choosing where to buy traffic, setting campaign parameters, comparing sources and deciding where the money goes.
That distinction is beginning to disappear.
Key takeaways
- In September 2026, IAB Tech Lab released AAMP 3.0 with OpenProposal, a standard for buyer and seller AI agents to exchange media proposals.
- 58% of respondents in IAB Europe's 2026 AI report expect agentic ad buying to reach operational use or scale within a year.
- An agent can only evaluate supply it can discover, so structured, searchable traffic marketplaces become more important, not less.
- Price and volume are easy to compare. Quality is not, so transparency becomes a publisher's competitive advantage.
On September 22, 2026, IAB Tech Lab introduced AAMP 3.0, the latest version of its Agentic Advertising Management Protocols framework. One of its most significant additions, OpenProposal, is designed to create a standardized way for buyer and seller AI agents to exchange media proposals. In practical terms, the industry is beginning to build infrastructure that would allow AI not simply to optimize an advertising campaign, but to participate in discovering, evaluating and buying media.
This may sound like another technical development in programmatic advertising. It is considerably bigger than that. If agentic advertising develops in the direction the industry currently expects, one of the most familiar roles in digital advertising could change fundamentally.
Instead of opening several platforms, contacting publishers, reviewing offers and manually comparing prices, an advertiser could eventually give an AI agent a commercial objective: find 5,000 qualified insurance leads in Spain at a maximum cost of €35 per lead. The agent could then search available supply, compare publishers, evaluate commercial conditions and performance information, and identify the opportunities that best fit the advertiser's requirements.
The media buyer is no longer simply using AI. AI is becoming the media buyer.
From automated bidding to automated buying
Programmatic advertising has been automating transactions for years, so at first glance this may not appear revolutionary. Algorithms already decide which impressions to bid on and how much to pay for them. Google, Meta and other advertising platforms already use machine learning extensively to optimize targeting, bidding and campaign performance.
But there is an important difference between optimizing a transaction and deciding where the transaction should happen in the first place.
Much of today's advertising infrastructure still depends on humans making those broader decisions. A buyer chooses the platform, identifies potential traffic sources, communicates with publishers or networks, compares offers and determines which relationships are worth pursuing. Even in highly automated environments, the machine generally operates within boundaries established by a person.
Agentic advertising expands those boundaries.
The emerging model gives AI the ability to interpret an objective and potentially execute more of the workflow required to achieve it. That could include discovering available inventory, requesting information, evaluating proposals and eventually initiating or recommending transactions.
There are signs that the industry expects this transition to happen relatively quickly. In research published by IAB Europe in September 2026, 58% of respondents said they expect agentic ad buying to reach operational use or scale within the next year. AI adoption in marketing more broadly is already widespread, with 86% of respondents reporting that their organizations use AI for marketing activities.
The important question, therefore, is no longer whether AI will become part of media buying. It already is. The more interesting question is what happens when AI begins making decisions that previously depended on human relationships, experience and discovery.
AI can compare traffic. First, it has to find it.
This is where the transformation becomes particularly relevant to the performance marketing and affiliate industries.
Digital traffic remains surprisingly fragmented. A company looking for insurance leads in Spain, dating traffic in Germany, casino players in Brazil or loan applications in the United States may have dozens or hundreds of potential sources available. But those sources are scattered across affiliate networks, agencies, direct publisher relationships, private communities, messaging groups, conferences and personal contacts.
Some of the best supply never appears in a conventional advertising platform at all.
A media buyer may know that a particular publisher generates exceptionally strong finance traffic because they worked together three years ago. Another opportunity may appear through a recommendation from an affiliate manager. A third may be discussed privately at an industry event.
AI does not automatically have access to any of this.
An agent can process thousands of opportunities far faster than a human, but it cannot evaluate an opportunity that it cannot discover. This creates an interesting paradox: the more automated media buying becomes, the more important structured and accessible marketplaces for traffic may become.
For an AI agent to make useful decisions, advertising supply has to become understandable not only to people but increasingly to machines. The agent needs to know:
- what type of traffic is available;
- which GEOs it covers and which verticals it performs in;
- how much volume can be delivered;
- what pricing model is accepted and what conditions apply.
Once that information becomes structured, traffic stops being simply "available." It becomes searchable and comparable.
That is potentially a major change for traffic marketplaces.
Their role may evolve from displaying offers to becoming discovery infrastructure connecting advertiser intent with publisher supply. Instead of a buyer manually searching through hundreds of opportunities, the starting point could simply be the commercial requirement: "I need this audience, in this market, at approximately this price."
The marketplace provides the supply. AI helps navigate it.
When buying becomes easier, quality becomes harder
There is another side to this development that may prove even more important.
AI can dramatically reduce the time required to identify and compare traffic sources. What it cannot automatically guarantee is that the traffic is good.
Anyone working in performance marketing knows that two sources offering the same GEO, vertical and pricing model can produce completely different results. A €25 lead can ultimately be far more expensive than a €50 lead if the first source generates low-intent users, duplicates, fraudulent activity or poor downstream conversion.
This problem becomes more significant, not less, when buying is automated.
If an AI agent can evaluate hundreds or thousands of offers almost instantly, it needs signals that allow it to distinguish cheap traffic from valuable traffic. Price and volume are easy to compare. Quality is much more complicated.
- Where did the traffic originate?
- How was the user acquired?
- Is the traffic direct or resold through several intermediaries?
- How old is the lead?
- Is the source compliant with the requirements of the advertiser and the market?
- What has the publisher delivered historically, and how consistently has that traffic converted for previous buyers?
These are questions experienced media buyers already ask. In an agentic environment, the challenge is turning those answers into information that can be evaluated systematically.
That could make transparency one of the most valuable assets in the next generation of advertising marketplaces.
Consider two publishers offering insurance leads in Spain. One offers leads at €28 but provides very little information beyond GEO and volume. Another charges €38 but provides clear information about acquisition channels, historical performance, delivery capacity and the origin of the traffic.
A human buyer may decide that the second offer represents better value. An AI agent could potentially reach the same conclusion in seconds, but only if it has access to the right information.
This means publishers may eventually compete on much more than price. Their reputation, historical performance, transparency and ability to describe their inventory accurately could become part of how machines rank and select advertising opportunities.
In that environment, a publisher's real product is no longer simply traffic. It is traffic together with the information required to trust it.
The media buyer is not disappearing
The arrival of agentic advertising inevitably raises the question of what happens to human media buyers. The likely change is less about removing people from the process and more about changing where their value lies.
A large amount of media buying work is operational: searching for inventory, collecting information, comparing offers, configuring campaigns, producing reports and continuously adjusting parameters. These are precisely the kinds of tasks AI can perform increasingly well.
Strategy is different.
Deciding which markets to enter, understanding why a particular audience behaves differently, evaluating reputational or regulatory risk, building long-term publisher relationships and recognizing opportunities that are not obvious from performance data all require a broader commercial context.
As automation expands, the media buyer can move away from operating every part of the transaction and toward directing the system that does it. The same shift is already visible in search, where Google is moving campaigns from keywords to AI Max and the advertiser's job moves toward defining the outcome.
Publishers face a corresponding change. Relationships will continue to matter, but discoverability will matter more. A publisher may have exceptional traffic, but that traffic has limited commercial value if the right buyers cannot find it.
The ability to clearly communicate what traffic is available, where it comes from, where it performs and under which commercial conditions it can be purchased becomes increasingly important in a market where discovery itself is becoming automated.
The next generation of traffic marketplaces
The first generation of digital advertising marketplaces helped organize inventory. Programmatic technology then automated much of the transaction. Agentic advertising may now begin automating discovery and decision-making.
That does not eliminate the need for a marketplace. It changes what the marketplace needs to provide.
If an advertiser can tell an AI agent exactly what it wants, there still needs to be somewhere for that agent to discover what the market can offer. There still needs to be supply from publishers across different GEOs, verticals and traffic types. There still needs to be enough information to compare those opportunities meaningfully.
Most importantly, there still needs to be trust.
This is the direction in which Traffberry is building its traffic marketplace. Advertisers can discover media opportunities from publishers instead of relying entirely on fragmented networks and existing relationships, while publishers can make their available traffic visible to buyers actively looking for it. Buyers can also compare ad network profiles, where every fact links to its source.
Today, those buyers are people. Increasingly, they will be people working alongside AI. Eventually, some parts of the discovery process may be handled primarily by agents.
The underlying requirement remains the same: good advertising starts with finding the right traffic.
As media buying becomes faster and more automated, access to transparent, relevant and discoverable supply becomes more valuable, not less.
If you are looking for traffic, explore the media opportunities on Traffberry. If you have quality traffic to sell, publish your opportunity and make it discoverable to advertisers searching for supply.
AI may be changing who, or what, buys the traffic. The market still needs a better place to find it.
Sources
- IAB Tech Lab: IAB Tech Lab introduces AAMP 3.0 with OpenProposal
- IAB Tech Lab: Why we're bringing agentic advertising into media planning with OpenProposal
- IAB Tech Lab: AAMP, Agentic Advertising Management Protocols
- IAB Europe: The Impact of AI on Digital Advertising Report 2026