Advertising for the agentic internet
Advertising for the agentic internet. OpenAds detects AI agent visitors to publisher sites and serves dynamic sponsored content targeted to the agent's purpose, so search agents surface brand content to end users. A human contextual DSP (generative display ads at 3x CTR benchmarks) serves as a self-funding distribution channel and training environment. Non-human attention is valid. The industry has spent decades blocking bots, but all attention is valuable; agents are the biggest net-new source of attention since smartphones, and attention always gets monetized. Building a DSP from scratch also forced them to learn the undocumented mechanics of targeting context rather than identity and cookies. Working at the intersection of AI and adtech since 2022, including building the first ad network for ads in chatbots before pivoting out. Steven wrote about ads in chat, pay-per-scrape, and ads for agents 6-18 months before the rest of the market. CTO co-founder Michael Bishop was principal engineer at Moat (acquired by Oracle for $850M), advisor Brian O'Kelley invented ad exchanges and founded AppNexus. "We know more about ads than anyone in AI, and more about AI than anyone in ads." Agent traffic grew 7,851% in 2025 and by March 2026 non-human traffic outnumbers human traffic. Agents are new entities coming online faster than humans can be born and will represent a fast-increasing share of economically relevant attention. On the human internet, ads outgrew SEO and paywalls 100:1, and the same dynamic applies: ad-subsidized free access to publisher content will beat pay-per-scrape micropayments.
Agents need better ways to capture attention and convert it into sales in an increasingly crowded advertising space.
OpenAds offers a human contextual DSP that utilizes reinforcement learning for adaptive advertising, enabling local business insight-driven campaigns.