The situation: It is still early days for buying products through platforms like ChatGPT or Gemini—only 7.2% of US digital shoppers say they regularly use AI shopping tools, per a March Bizrate Insights/EMARKETER survey—but genAI platforms are actively testing which experiences resonate with shoppers.

The latest example: Anthropic’s Claude AI recently expanded connectors—its app-like functionality—beyond business tools to consumer services including Instacart, StubHub, and Uber Eats. At the same time, Anthropic changed how those connectors appear.

Rather than requiring users to reference an app directly in their prompts, Claude automatically pulls in relevant integrations when they can help answer a query. When multiple connectors apply, Claude surfaces several options and lets users choose. For example, it might present the Uber Eats connector for a user seeking nearby healthy lunch options or surface TripAdvisor for a hotel within a set budget.

Anthropic is committed to keeping Claude ad-free, meaning users won’t see paid placements or sponsored results. If multiple connectors are relevant, they are ranked based on utility rather than monetization.

The contrast: This approach differs from other platforms such as OpenAI’s ChatGPT, which has taken a more commercially driven path. It initially rolled out in-app checkout functionality that failed to gain traction, then pivoted to a model where consumers complete purchases via third-party apps that plug into the chatbot.

By experimenting with more commercial integrations and monetization pathways, platforms like OpenAI increase the likelihood that ranking and visibility could be shaped by paid placement rather than pure user value, which could potentially dampen trust and demand over time.

“OpenAI is flat out sprinting in the direction of revenues, while Anthropic seems to understand this is a marathon,” said Nate Elliott, EMARKETER principal analyst.