AI Agents Just Became Your New Media Buyer
- Aseem Singh
- 14 hours ago
- 2 min read
I stopped tracking every model drop this week. The useful news is simpler: AI stopped being a tool you prompt and started becoming a customer you have to sell to.
Three moves that actually matter
1. Gemini hit 1 billion monthly users. Google said the standalone Gemini app crossed that mark. ChatGPT was already there. Distribution is no longer the debate. If your brand is invisible inside answer engines, you are invisible to the next buying layer.
2. Perplexity crossed $750M annualized revenue. That is up from under $250M at the start of 2026. Answer engines are not a side quest. They are a media channel with a P&L.
3. Tencent open-sourced Hy4 — 770B parameters, 1M-token context. Chinese labs keep closing the gap with cheaper, longer-context open weights. Agencies that only know one closed stack are about to look slow.
The marketing shift: agents now buy media
This is the part that actually changes our week at Dzine Prodigy.
The IAB is building a measurement framework for ads served to AI agents, aiming for a November 12, 2026 release. Time already started serving ads to agents. OpenAI made Automatic Advanced Matching the default on ChatGPT Ads pixels. Auxia shipped Agent Studio so marketers can describe a campaign in plain English and let agents plan, build, and launch it across existing tools.
BCG's 2026 CMO survey is the uncomfortable punchline: 96% of CMOs say AI is transforming marketing. About a third have actually rebuilt the work. Only 8% run multi-agent campaigns.
So the gap is not tools. The gap is operators.
What we are doing this week
LAYER 1 — Make the site readable to agents. Clean product facts. No fluff landing pages. Structured answers.
LAYER 2 — Treat ChatGPT Ads, Gemini, and Perplexity like media, not experiments.
LAYER 3 — Put a human on governance. Agents can buy. They cannot own the brand.
Finish one thing before Friday: open your homepage the way an agent would. If it cannot extract what you sell, who it is for, and why it should recommend you, you are already losing the next customer.


Comments