adMarketplace expands into AI chat with CPC-based ad model in beta rollout

adMarketplace has launched a beta program for AI Discover, a performance advertising product designed to integrate sponsored placements directly within AI chat environments, as companies test monetization strategies for conversational interfaces.

The rollout follows a six-month alpha phase conducted with Opera AI, during which more than 2 million sponsored placements were delivered across approximately 700,000 AI queries. The results provided early indications that conversational interactions frequently contain commercial intent that can be translated into advertising opportunities.

AI Discover operates by analyzing user conversations in real time to detect product-related intent. Its underlying system includes a proprietary signal extraction layer, known as Commercial Intent Vector, and an ad-serving engine that matches those signals against a catalog of more than 200 million product listings. The system then ranks and inserts relevant sponsored options into the chat interface.

Ads appear in formats such as text links, hover-based placements, or sponsored results embedded alongside or below AI-generated responses. According to the company, a typical interaction can generate multiple potential ad placements, reflecting a shift from traditional search environments where visibility is concentrated in a limited number of positions.

The beta program extends beyond Opera’s ecosystem to include partners such as Dupe and Sezzle, both of which are developing AI-driven commerce experiences. The test phase will run from June through December 2026 across the United States, United Kingdom, Germany, and France, involving a limited group of advertisers and AI platform partners.

Unlike impression-based pricing models, AI Discover is built on a cost-per-click structure, where advertisers are charged only when users engage with a sponsored placement. This approach is intended to align ad spend with user intent in a context where a single conversation may surface multiple relevant options.

The company stated that its ad systems operate independently of the AI model’s reasoning process, emphasizing that sponsored placements do not influence the content of generated responses. This separation is positioned as necessary to maintain user trust in AI-driven recommendations.

Early findings from the alpha phase suggest that AI-driven interactions can compress the traditional purchase funnel, enabling users to move from research to transaction within a single session. The beta program is expected to further evaluate how scalable and sustainable such behavior is across different platforms and markets.