Ad tech does not need more tollbooths

A modern media buy can pass through systems for planning, identity, data, buying, selling, exchange, verification, measurement, billing, and reporting. Many of those capabilities matter. The problem is that they were assembled one platform at a time, each with its own taxonomy, interface, incentives, and version of the truth.

Advertisers experience that fragmentation as operational drag. Teams translate the same intent across briefs, spreadsheets, dashboards, deal IDs, and insertion orders. Publishers experience it as margin pressure and lost control. Their audience and environment create the value, but the path to demand is often governed by intermediaries they cannot see or influence.

The industry has learned to manage this architecture remarkably well. But managing complexity is not the same thing as eliminating it. Every handoff creates another place for context to degrade, fees to accumulate, and accountability to blur.

Debates about ad-tech take rates often focus on whether a particular fee is too high. That is useful, but incomplete. The deeper issue is the number of times a transaction must be interpreted, packaged, matched, cleared, checked, and reconciled before an ad can run.

Some services add unique value. Others exist because neighboring systems cannot communicate directly. When information is trapped inside closed platforms, translation becomes a recurring cost center. The market pays not only for technology, but for the absence of shared infrastructure.

That is why opacity and economics reinforce one another. Buyers cannot easily distinguish working media from duplicated machinery. Sellers cannot see which layers are essential and which simply control access. Complexity becomes defensible precisely because no participant has a complete view of the path.

AI agents make a different market design possible. A buying agent can carry an advertiser's objective from discovery through negotiation and order preparation without a person re-entering the same intent in a succession of tools. A selling agent can make inventory, pricing rules, and approval requirements legible without forcing a publisher into someone else's closed workflow.

This is more than a faster interface. Dashboards made complicated systems easier for people to operate; agents can reduce how many operations are necessary in the first place. They can search broadly, compare consistently, preserve context, and produce a durable record of what was requested, offered, and agreed.

The result should be a shorter path between budget and inventory. Fewer handoffs mean less reconciliation, less duplicated infrastructure, and fewer opportunities for the commercial intent to disappear. Human judgment remains central—especially in strategy, creative, and relationships—but it can move up the value chain instead of being consumed by workflow.

Agents only create a healthier market if they can transact across organizational boundaries. Without a shared language, every buyer and seller still needs a bespoke integration, and the new interface inherits the old fragmentation.

The Ad Context Protocol provides that common transaction language. It gives buying and selling agents a consistent way to describe objectives, inventory, terms, creative requirements, and outcomes. Publishers keep control of what they offer. Buyers gain a comparable view of the market. Both sides can audit the path from intent to delivery.

Open protocols do not remove differentiation. They move it to the places where it belongs: quality of supply, quality of demand, intelligence, service, execution, and outcomes. A shared rail expands the addressable market because participants can connect once and interact with many counterparties.

Basin is building the marketplace and infrastructure for that agent-native model. We normalize publisher inventory, expose it to buying agents over AdCP, preserve publisher controls, and create an auditable workflow from discovery through execution.

Our thesis is deliberately simple: when agents can understand the market directly, value moves toward the participants who create demand, create media, and deliver measurable outcomes. Software should earn its place by making the transaction more useful—not by making itself unavoidable.

The transition will not happen all at once. Existing systems will coexist with agentic workflows, and many valuable services will adapt rather than disappear. But the direction is clear. Markets tend to reward lower coordination costs, greater transparency, and broader access once the infrastructure makes them possible.

Ad tech is ripe for disruption because its complexity is no longer a fixed constraint. A market built for agents can be more direct without being less capable, more open without surrendering control, and more efficient without sacrificing trust. That is the market Basin intends to build.