New data reveals how agentic media buying is changing programmatic ad performance. Analysts compare CPMs, fill rates, and auction strategies to traditional methods. Publishers see shifts in user engagement and revenue metrics.
Media and publishing professionals tracking programmatic revenue models are watching a new shift: agentic media buying is now being measured against traditional non-agentic approaches, with fresh data from DataBeat, a MediaMint company, highlighting key differences in CPMs and fill rates. The analysis shows that agentic buying, which uses coordinated AI agents to autonomously plan, execute, and optimize campaigns in real time, is far more selective in targeting inventory and audiences. This selectivity is driven by a focus on aligning spend with specific campaign objectives, rather than casting a wide net across all available inventory.
According to DataBeat’s report, agentic buyers participated in 86% fewer auctions than non-agentic buyers, yet their CPMs were only slightly lower. This suggests that AI-driven strategies are able to maintain competitive pricing while being highly selective. Despite engaging in fewer auctions, agentic buyers achieved higher fill rates, indicating that their targeted approach does not come at the expense of monetization efficiency. The data also shows that agentic buyers operate on a much smaller share of auction volume but still deliver comparable CPM and fill-rate performance to traditional methods.
Publisher-side metrics reveal additional shifts. In June, users initiated 5.9% fewer sessions and viewed 7.0% fewer pages per visit compared to the previous period, with the average session count holding steady at 1.54. Engagement rates dropped by 3.2%, reflecting a rise in lower-intent traffic. While the average session duration increased by 19%, this was not enough to offset the overall decline in visits and content consumption, suggesting that longer sessions alone do not compensate for reduced engagement-a trend attributed in part to the influence of agentic and AI-driven engines.
The “Programmatic Trends Report” from DataBeat, based on anonymous June 2026 data from its network partners, benchmarks May 2026 results against April 2026 and May 2025. The findings show display CPMs rose 5.9% month-over-month, while video CPMs increased 10.5%, resulting in a 6.2% overall CPM gain. Year-over-year, display CPMs climbed 10.6% and video CPMs 4.2%, for a net overall CPM increase of 24.2%. In June 2026, DataBeat tracked over $55 million in monthly revenue, 35 billion impressions, and more than 200 monthly bidders. ChatGPT’s share as a referral source also grew, with its contribution to total sessions rising from 0.017% in April to 0.037% in May, and its share of referral traffic increasing from 2.1% to 3.4%.
For those interested in the broader implications of automation in digital ad sales, a related analysis explores whether agentic advertising risks recreating the same intermediary challenges that programmatic was meant to solve. More on this perspective can be found in this report examining agentic advertising’s potential pitfalls.
DataBeat, a subsidiary of MediaMint, specializes in programmatic advertising analytics and works with a network of publishers and advertisers across the United States. The company’s monthly reports track billions of impressions and hundreds of bidders, providing benchmarks for CPMs, fill rates, and revenue trends in the digital advertising market.
