Amazon Expands Rufus AI As 300 Million Users Show 60% Higher Purchase Rates

Amazon.com Inc. AMZN is pushing deeper into AI-driven commerce through its upgraded shopping assistant, Rufus, though early usage suggests the experience is still evolving. Initial interactions have produced inconsistent recommendations, reflecting the difficulty of translating large-scale consumer data into precise purchase suggestions. Even so, the company is positioning Rufus at the center of its e-commerce ecosystem, aiming to embed AI directly into how users discover and buy products. This approach draws on Amazon's extensive data advantage, though its past efforts with Alexa indicate that execution timelines could extend longer than initially expected.

The latest enhancements introduce scheduled actions, enabling users to automate recurring shopping behaviors such as replenishing everyday items or selecting gifts based on prior activity. This feature, now available to all U.S. customers, represents what some analysts describe as an early version of fully agent-driven commerce. Chief Executive Officer Andy Jassy indicated that around 300 million customers used Rufus in 2025, and those users were about 60% more likely to complete a purchase. That level of engagement could translate into incremental revenue over time if adoption continues to expand, although the system still requires user confirmation before finalizing transactions.

At the same time, competitive pressures are building as rivals explore similar AI-driven shopping models. Efforts linked to OpenAI and internal developments at Walmart Inc. WMT highlight the broader industry push, though early integrations have faced limitations. Amazon has restricted third-party AI agents from operating on its platform, citing customer safety concerns, and has taken legal action against certain firms over data access. Over time, this stance could attract regulatory or consumer scrutiny if interoperability becomes a priority. For now, Amazon appears focused on leveraging its first-party data and scale to refine Rufus, with the broader outcome likely depending on how quickly it can improve reliability and user trust.