The infrastructure gap

Every human sees ads
on Google. No one sees ads
on Hermes.

Autonomous agents are scouring product graphs and buying on behalf of users. Scour is the ads layer that makes sponsored listings appear in those queries — the same way search ads work for humans, but built for machines.

Agentic Commerce Sell-Side Infrastructure ACP / AdCP
The problem

Today

Humans search Google, see sponsored listings, click, buy. Amazon, Google, OpenAI all have ads inside their AI agents. But these are first-party surfaces — captive, closed ecosystems.

Tomorrow

Users delegate shopping to open-source autonomous agents — Hermes, OpenClaw, custom agents — that scour the open web. Nobody is building ads for that layer. The entire $72B retail media market is built on human browse behavior.

The vision
When your agent searches for the best running shoes, the answer includes both organic results and listings that brands paid to be in. The same way it works today — just with machines as the buyer.
$20.5B
AI-driven retail spend in 2026, up 4x from 2025
50%
Projected share of online shoppers using agents by 2030
$72B
US retail media ad spend in 2026 — all built on human search
How it works
01

Buyer agent searches

An open-source agent — Hermes, OpenClaw, or any ACP-compatible agent — queries Scour for product listings matching a user intent.

02

Scour returns the shortlist

Scour responds with ranked product listings. Sponsored listings are embedded as structured data — visible to the agent's ranking evaluation, not hidden from it.

03

Brand gets in front of the agent

Brands bid to appear in relevant agent queries. Scour's ranking model balances bid value with product quality signals — because agents respond to credibility, not just checks.

The ad layer exists for humans.
The agent layer is wide open.

Scour is building the infrastructure that makes sponsored listings work for autonomous agents — protocol-native, quality-weighted, built for the next decade of commerce.