Anyone in the media industry can tell you how third-party signal degradation has made traditional ad targeting less reliable. At the same time, AI has made it easier to analyze publisher data to identify high-quality audiences in a fraction of the time it takes manually.

Programmatic advertising demand increasingly flows to flow wherever inventory can be understood quickly, evaluated objectively, and activated without friction. For sellers, visibility in that process is an advantage that’s quickly becoming a necessity.

The automated ad buyer has arrived

Earlier this year, a media agency completed the first publicly-documented, fully-agentic campaign. An AI agent interpreted a client brief, queried an SSP’s inventory, evaluated options, and returned a recommendation for human approval before initiating the campaign. After launch, the agent was set to continue optimizing the campaign’s pace, targets, and performance.

This autonomous advertising campaign delivered:

• 82% reduction in supply chain cost
• 40% lift in impressions
• 30% lower CPM
• 98% average video completion rates

That’s just one example, but the results are promising — and replicable.

That SSP is now running similar programs with more agency partners. Marketing agencies are executing live agentic media buys with direct publisher connections. CTV infrastructure companies are building systems specifically to make streaming inventory accessible to agentic buyers.

What we’ve learned is that agentic buying can already reach and engage with both programmatic and direct deals. Unfortunately, teams that wait for agentic buying to mature might start falling behind.

Publishers need to compete on accessibility and speed

When buyer agents evaluate inventory, the sellers who appear first are the ones whose inventory clearly answers who you reach, how you know that, what is available, at what price, and under what conditions. An agent needs clean, clear signals to evaluate before it can act. If the signals aren’t legible, the agent will move onto inventory that is.

The manual discovery-to-activation cycle wasn’t built for machine speed. Media sellers with a strong agentic foundation will be set up to compress these timelines from weeks to minutes, giving them the ability to close more direct and programmatic deals than slower competitors.

Scale still matters for getting on a buyer’s radar. Accessibility is increasingly what determines whether you get the deal.

What the sell side is building

Buyer agents are becoming a staple of the ecosystem, but I’m keeping a close eye on the layers the sell side is developing.

Audience agents can synthesize first-party data, identify niche targets, and package inventory into structured outputs that buyer agents can evaluate directly. Sales teams can deploy those outputs in RFP responses without rebuilding audience packages manually for each cycle.

Seller agents are the next layer in development. These systems are designed to communicate and transact directly with buyer agents. This will compress the direct deal cycle from negotiation across weeks to structured exchanges in minutes.

Even before the agents start autonomously transacting, sellers who prioritize their identity infrastructure and agentic tooling are positioned to benefit. Sell-side agents can improve direct sales workflows and ensure buyer agents will be able to evaluate their inventory well before the exchange starts.

The standards defining the agentic advertising system

Open protocols, like the Ad Context Protocol (AdCP), which define how agents discover inventory, describe audiences, and transact with each other, are still in development. But they’re rapidly maturing, and major players are hedging their bets on the standards that will best suit their needs.

What I find most interesting about this moment is that publishers still have a say in what comes next. These protocols will impact how your inventory is represented in an agent query, how pricing is structured, and how brand safety and consent signals are exchanged. 

The companies running agentic tests now are shaping how agentic advertising will look. However, those who wait will inherit a framework built around someone else’s inventory.

Where publishers should focus

I’d recommend taking steps to take as soon as you’re able:

1. Audit your data foundation for agent readiness

A self-assessment is the best place to start. Score your organization across the dimensions of agent readiness, including: your data, inventory signals, content and context signals, technical infrastructure, and organizational awareness. An agent readiness self-assessment is a good way to find out where you stand and what you need to prioritize.

2. Deploy audience agents in monitored workflows

From there, the clearest opportunity involves deploying audience agents against your existing direct sales workflow. You can start with one property or deal type before a broader rollout. Measuring against your current baseline is how you’ll know what’s working.

3. Assign a team member to an agentic standards working group

On standards, I’d get someone into the AdCP and MP working groups sooner rather than later. You need a representative who understands your deal structure well enough to read a proposal and identify what works with your business and what could pose some issues.

The demand shift is underway. Buyer agents are already querying inventory, the sell-side is building the stack to meet them, and working groups are writing the standards that will govern how they transact. Publishers who build now, rather than treating it as a problem for later, will be the ones buyer agents find first. They’ll also be in a position of strength, shaping where the industry is headed at this time of rapid change.