Programmatic advertising has never had more data or automation. Yet for many advertisers, campaign activation still involves juggling dozens of publisher deals, navigating fragmented supply paths, and spending hours on manual optimisation.
Curation has emerged as one of the industry's most promising solutions, bringing greater control and efficiency to programmatic buying. But while today's curated deals solve many operational challenges, they're often static by nature.
This is where AI agents come in. Instead of simply packaging premium inventory, agentic curation continuously analyses, adapts, and optimises supply before bids are even placed. The result is a smarter, more efficient approach to programmatic buying that improves both campaign performance and operational efficiency.
Curation aims to tackle a well-known dilemma in the market. For years, advertisers have had to choose between two imperfect options.
The open marketplace offers virtually unlimited scale, but often lacks transparency and control, leading to inefficient supply paths. Buyers face higher costs, reduced access to quality audiences, limited control over where ads appear, and increased exposure to fraud, brand safety issues, and low-quality "Made-for-Advertising" (MFA) environments.
On the other hand, private marketplace (PMP) deals provide access to premium publishers and greater control, but they rarely scale easily. Running campaigns across multiple markets often requires researching local publishers, negotiating dozens of individual deals, working across several SSPs, and constantly monitoring and optimising each deal manually.
Rather than buying from the entire open marketplace or managing countless direct deals, curation packages carefully select inventory into a single Deal ID that advertisers can activate through their DSP.
The benefits are clear:
However, there's still one major limitation. Curated deals remain largely untouched after they're created.
Traditional curation represents a significant improvement over open marketplace buying, but it's still largely static. Most optimisations still happen manually, after impressions have already entered the auction.
In other words, advertisers are still reacting to performance rather than proactively improving the supply available for bidding. This is where agentic curation changes the game.
Recent advances in AI have introduced so-called agents, systems that can independently evaluate information, make decisions within predefined guardrails, and take action without requiring constant human intervention.
Applied to curation, these agents continuously analyse supply and optimise curated deals based on the campaign's objectives.
That means it can:
This represents a fundamental shift in where optimisation happens. Rather than relying solely on DSP optimisation after impressions become available, agentic curation improves the quality of available supply upstream. Instead of treating curation as simply a marketplace tool, it becomes a live optimisation engine.
The value of agentic curation isn't that it uses AI. The value lies in solving some of programmatic advertising's biggest challenges.
Campaign teams no longer need to create and manage dozens of individual publisher deals across different markets. Instead, advertisers activate a single intelligently optimised Deal ID while the optimisation happens automatically behind the scenes. This significantly reduces manual work and operational complexity.
Ultimately, successful campaigns aren't about generating more impressions. They're about buying better impressions. Because agentic curation continuously evaluates supply quality and allocates budget where it performs best, advertisers can improve both campaign effectiveness and media efficiency.
Across the first six months of offering our own Agentic Curation solution, results already show:
These improvements demonstrate the value of optimising supply before bidding rather than relying solely on downstream optimisation.
Curation isn’t just another trend. It’s a smarter way forward for advertisers who demand more transparency, quality, and performance from every impression.
It's conquering the industry's biggest challenges, and with agents, curation is no longer just about packaging inventory; it's an intelligent optimisation layer. The closer optimisation happens to the impression, the greater the opportunity to improve performance.
Rather than asking whether AI belongs in programmatic advertising, the more important questions are where intelligence should sit, how transparent optimisation should be, and how shared standards can strengthen the open internet while improving outcomes for advertisers.