AI ads are here — and they don't look like banner ads
Why the next generation of advertising will be native to AI-generated conversations, not bolted onto them.
For twenty-five years, the web has been monetized by a single pattern: pages full of content, with ad slots stamped into the margins. Search results with sponsored links on top. Articles with banners down the side. Social feeds with promoted posts slipped between friends.
That pattern only works when the user is looking at a page. AI-native products aren’t pages. They’re conversations.
The page is disappearing
When you ask ChatGPT for a new laptop recommendation, it doesn’t return ten blue links with ads on top. It returns a single answer. When you ask Claude to plan a trip, it doesn’t show you a SERP — it writes you an itinerary. The interface between a user’s intent and their outcome has collapsed from a page of options to a single generated response.
That’s great for users. It’s existential for the open web’s advertising model. You can’t put a banner next to a chat bubble. You can’t run a display network inside a streaming token response. The surface area that made web advertising work — the page — is being compressed into a paragraph.
What comes next
The question isn’t whether AI products will carry ads. They will — AI inference is expensive, and free tiers need revenue. The question is what shape those ads take.
A few patterns are emerging:
- Sponsored suggestions inside generated answers. When the model recommends three laptops, one of them is paid placement — clearly marked, relevant to the query, and ranked honestly against the unpaid options.
- Product cards attached to responses. Not an interruption to the conversation, but a structured card the user can expand when the model’s answer surfaces a concrete buyable thing.
- Intent-aware placements. Ads that respond to what the user is trying to do, not what they happen to be reading. The conversational context is richer than any URL ever was.
What all of these have in common: the ad is part of the response, not a frame around it. If it isn’t useful in the moment the user asks, it won’t get served.
Why this is good news for advertisers
Contextual advertising has always been a compromise. You bid on “running shoes” and hope the user reading the running-shoe article is actually shopping for shoes. AI flips this. The user literally told the model what they want. The intent signal is the prompt itself.
That’s a higher-quality signal than any cookie ever produced — and it arrives without the tracking infrastructure that regulators are rapidly dismantling. For the first time in a decade, the advertising industry has a path to more targeting precision with less surveillance.
What we’re building at Growl
Growl is the monetization layer for AI-native products. We give AI product teams a way to plug advertising into their generated responses without building an ad stack from scratch — real-time bidding, creative matching, and transparent labeling, all behind a single SDK.
In the posts that follow, we’ll get into the specifics: how RTB works when the “page” is a streaming response, what privacy looks like when the ad decision happens server-side, and why the old IAB taxonomy doesn’t map cleanly onto conversational intent.
Ads aren’t going away. They’re just going to stop looking like ads.