Ad Formats Designed for Chat

This reveals that monetizing chat is really a product design problem, not just an ad targeting problem. In a conversational app, a standard banner breaks the flow and looks out of place, so Koah uses units that behave like parts of the exchange itself, short text before or after an answer, sponsored sources inside citation panels, suggested next questions, and static placements in unused input screen space. That makes ads feel native to the chat workflow while preserving the app’s core answer experience.

• The format choices line up with where attention naturally sits in chat. Prefix and suffix units attach to the answer itself, sponsored citations sit where users already inspect evidence, and follow up prompts turn an ad from a dead end into the next turn of the conversation.
  1 sacra
• This is closer to sponsored search than display advertising. Google now places ads around AI Overviews on relevant commercial queries, and lets users continue with follow up questions from the overview, showing that ad inventory in generative products is moving toward intent matched, conversation adjacent placements instead of separate banner slots.
  2 blog 4 blog
• The closest emerging peer is Nexad, which pitches itself as AI native advertising for chat and search apps. The broader pattern is that new ad networks for AI apps are being built around query level intent, native text units, and monetization that can plug into answer surfaces without forcing publishers to redesign the whole product around ads.
  5 nex 6 a16z

Going forward, the winning ad formats in AI products will look less like web page inventory and more like small, useful extensions of the conversation. That favors networks that can match commercial intent at the query level and serve placements that fit naturally into answer, citation, and follow up flows, which is where chat products are training users to look and act.