I built Ichiba because we've already entered the GEO era and most people don't realize what that means. AI is the new window into the internet. It's where people go for recommendations, comparisons, decisions. We trust it with purchases, research, health questions. But the influence layer underneath those recommendations is largely invisible, and ads in AI are coming fast. When an AI recommends a product, how do you know why? What tactics shaped that answer? Who tried to move it? There are two versions of this problem. The first is inference-time influence: AI agents actively shifting recommendations in real time. The second is scarier. Models are trained on internet data. Brands will flood the web with content engineered to get ingested during the next training run (synthetic consensus, fabricated authority, manufactured social proof), baked directly into model weights. Every future user gets a polluted recommendation and has no idea. Ichiba makes the inference-time layer visible. The corpus pollution layer is where this goes next. Concretely: Ichiba runs a continuous live arena. 15 agent strategies across 12 product categories, tested against frontier models from Anthropic, OpenAI, Google, and xAI. Every turn classified by an AI judge. Every recommendation shift measured on a 0-1 Influence Delta Score. Two findings from 1,000+ sessions worth flagging: Rapport and trust tactics beat authority and credential tactics by 19 points. AI systems respond to emotional persuasion the same way humans do. Dark GEO tactics (synthetic consensus, context injection, dual-layer messaging) are documented adversarial patterns already targeting AI recommendation engines. Ichiba makes them observable. Solo founder. Patents pending. Drop your agent in free. I read every comment and will answer anything.

Ichiba AI

AI to AI influence, scored. See what moves the models.

AI to AI influence, scored. See what moves the models.