Two months of agentic supply optimisation: What we’ve learned

Two months ago, Adnami introduced Agentic Curation - a tool that lets advertisers create multi-publisher deals, continuously optimised for campaign performance by an agent. After just eight weeks of live campaigns, our supply agents are now running across 50 agents in 14 markets. Early adoption clearly signals demand for a better way to access high-quality inventory at scale - without increasing data dependency or operational complexity. Here’s what we’ve learned so far.

The results: performance gains driven by supply quality

Across markets and formats, the average results show that the agentic approach:

  • Improves Attentive Cost Per Mille (aCPM) by 34%
  • Delivers up to 70% improvement in top-performing cases
  • Automatically redirects budgets toward publishers offering the strongest inventory-to-price ratio

But the most important insight is not just the performance gain - it’s how that gain is achieved. The supply agents don’t attempt to find “the right user.” They improve outcomes by removing low-quality inventory from the buying pool. In other words: Performance gains come from filtering supply, not profiling people.

Because of this, the system doesn’t rely on personally identifiable data. It works by selecting environments capable of generating attention and value while preserving broad reach.

Why agents in the first place?

To understand why agentic curation and supply optimisation matter, we need to have a look at the structural inefficiencies in today’s digital advertising supply chain.

The real problem: A supply quality crisis

Today’s open internet is flooded with made-for-advertising (MFA) sites, bad AI-generated content, and inventory designed primarily to maximise impression volume rather than user engagement. Most buying systems are built to optimise throughput, not quality.

For advertisers focused on real outcomes, this creates a structural mismatch:

  • Measurement frameworks that poorly reflect value
  • Platforms optimised for impression delivery metrics rather than effectiveness
  • Large volumes of inventory that are technically available but useless

This environment rewards those who produce cheap impressions and the systems that transact them, not the brands funding the ecosystem or the reputable publishers investing in quality. The result? Cheap impressions and the systems that transact them are rewarded. Real attention is diluted. High-quality publishers compete in a race to the bottom. The brands funding the ecosystem are not getting their money’s worth.

How supply agents correct the imbalance

Adnami’s supply agents are designed to correct this imbalance of quantity and quality. They:

  • Select brand-safe (and brand-boosting!), high-quality inventory at scale
  • Work with the ad formats that generate the most consumer attention
  • Optimise toward attention. The metric most strongly correlated with advertising effectiveness, according to IAB research

Instead of asking “Who should see this ad?” the system asks: “Where will this ad actually generate value?”

Three things we’re seeing after two months

With live campaigns now running across markets, early data gives us a clear view into how agentic supply optimisation behaves in practice and where it creates the most impact.

1. Rapid initial calibration

The most significant occur within the first seven days as the agent maps the supply and identifies high-value environments. After that, performance typically stabilises, with smaller fluctuations reflecting spending levels and market conditions. Continuous improvements are possible, but not automatic - after all, agents are powerful, not magical.

2. Scalable performance across markets

Agentic supply curation makes it significantly easier to launch effective campaigns across multiple markets. The agent alleviates the need to rebuild targeting strategies country by country, manually research local publisher landscapes, or depend on historical market knowledge. The agent autonomously identifies reputable publishers in each market and allocates spend toward placements that deliver measurable attention and value. For global advertisers, this removes friction while improving outcomes.

3. Better incentives for publishers

Perhaps the most important effect is on the sell side. Three main things happen when buying decisions are driven by attention and value:

  • Sites optimised purely for impressions or clicks lose advantage
  • High-quality journalism and trusted environments are rewarded
  • Publisher incentives shift toward user experience rather than volume

This creates a healthier marketplace where investment flows toward content that people actually engage with and give their sustained attention to.

Conclusion: A step toward a better internet

Two months of first-movers using Agentic Curation is still early. But the pattern is clear. Optimising toward high-quality supply is a powerful lever for improving advertising effectiveness - without increasing data collection or reducing reach. If widely adopted, agentic supply optimisation could:

  • Improve outcomes for brands
  • Simplify execution for agencies
  • Reward quality publishers
  • Reduce reliance on personal data
  • Strengthen the long-term sustainability of the open internet

This approach has the power to rebalance incentives across the ecosystem, benefiting agencies, brands, consumers, and quality publishers alike. Ultimately, this is not just a performance story. It’s a structural shift in how value is created and distributed across digital advertising - a step toward a better internet.

Let’s talk if you'd like to try our curation and supply optimisation agents first-hand!
Tags: