AI
Chatbot ads are arriving, but the industry remains skeptical about what actually works
As OpenAI and Google expand ads in their AI assistants, startups and advertisers are discovering which tactics resonate and where the technology still falls short.
WIRED — AI
While Anthropic has pledged to keep Claude free from advertising, the broader AI chatbot industry is moving in the opposite direction. OpenAI has amassed over 100,000 advertisers on ChatGPT since launching ads in February, and Google recently expanded its AI Mode search ads to 10 additional countries. Yet despite this momentum, questions persist about whether advertising in conversational AI will ever feel natural—or profitable.
For smaller platforms that lack massive venture funding or built-in user bases, ad revenue may become essential to sustainability. Koah Labs, a San Francisco ad-tech startup now serving ads across 200-plus apps including DeepAI and Liner, is helping to map what works in this emerging space. The company has drawn just a fraction of OpenAI's advertiser base, but its CEO offers early lessons about the shape of chatbot advertising to come.
Why AI-Written Ads Win More Clicks#
One of Koah's most surprising findings: ads drafted partly by AI—rather than entirely by humans—generate twice as many clicks. The reason likely stems from a simple match between form and context: copy generated by machine learning feels native when it appears next to an AI model's own output.
Yet this efficiency comes with friction. Pharmaceutical firms and other heavily regulated industries require strict human oversight of their messaging to comply with legal mandates. That caution, Baird notes, results in duller, less effective ads. The tension between creative optimization and regulatory safety is shaping how ad platforms must design their workflows.
Context Unlocks Better Targeting#
Traditional search offers limited data: a query like "help with stress" yields ads for therapists. Chatbots reveal much more. A user exploring stress-reduction strategies in a conversation might see an ad for a spa or wellness retreat—a more oblique but ultimately more relevant match, based on the full dialogue thread.
Koah can examine not just the user's prompt and the AI's response, but the entire conversation history to date. Since chatbot queries often run twice as long as Google searches, there's simply more information to work with. The company reports ad click-through rates around 3 percent—competitive with mature search platforms.
The Privacy Bargain#
Showing ads based on full conversation history raises uncomfortable privacy questions. Baird says Koah anonymizes queries and stores them in de-identified form, but declined further details on data practices. He frames personalization as a win-win: advertisers learn what users truly need, and users receive offers that feel less random. An advertiser seeing that a user criticized a competitor's product, for instance, could tailor a landing page to address that concern.

