Our Take
Rufus grew 115% monthly active users and 400% engagement year-over-year, but Amazon buried the name—a sign the real play is integration into Alexa's installed base, not chatbot differentiation.
Why it matters
Practitioners building shopping agents need to watch how Amazon is collapsing separate products into its core assistant. The conditional purchase automation (buy if price drops and time threshold passes) signals where consumer AI is headed: less chat, more do.
Do this week
Product teams: audit your agent's task-chaining logic before Q2 planning so you can map which automations require explicit user confirmation versus which can execute conditionally.
Rufus becomes Alexa's shopping backbone
Amazon has merged Rufus, its shopping-focused chatbot, into Alexa for Shopping, a unified interface available across the Amazon Shopping app, Amazon.com, and Echo Show devices. The consolidated assistant handles product questions, side-by-side comparisons, and price tracking across a year of history. Users can ask conversational shopping questions directly in Amazon's main search bar instead of toggling to a separate chatbot window.
Rufus served more than 300 million customers in 2025 and showed strong adoption metrics: monthly active users grew 115% while engagement increased nearly 400% year-over-year (per company disclosure). Despite these gains, Amazon is retiring the Rufus name from its shopping interface, though the underlying technology continues to power parts of the experience.
The automation layer
The new assistant can set scheduled shopping actions tied to conditions. For example, it can add an item to a customer's cart if the price reaches a target threshold and has not been purchased within a set period. Amazon's Buy for Me agentic AI feature can complete eligible purchases using stored payment and address information, with no additional user confirmation required at execution time.
Alexa for Shopping also surfaces products from third-party retailers via Shop Direct, expanding beyond Amazon's first-party catalog. The assistant draws context from a customer's shopping history, browsing, purchases, and past Alexa voice interactions to generate product recommendations and guides for larger purchases.
Full Echo Show shopping access
Amazon is adding full-store shopping to Echo Show, with initial availability on Echo Show 15 and Echo Show 21 for Alexa+ subscribers. Users can browse, search, and shop using voice, touch, or both. The rollout supports cross-device continuity: customers can start a task on their phone and resume on a laptop or Echo device.
The feature is available free to signed-in Amazon customers without requiring Prime membership, an Echo device, or the Alexa app.
Integration, not differentiation
The consolidation of Rufus into Alexa signals a shift from product isolation to assistant-centric architecture. Rather than defend Rufus as a standalone shopping chatbot, Amazon is embedding its capabilities into a broader conversational layer that spans search, voice, and device ecosystems. This is a maturation pattern: early-stage chatbots prove adoption; successful ones dissolve into platform infrastructure.
Conditional purchase automation is the under-discussed feature here. It moves beyond recommendation and comparison into agentic execution. A customer sets a rule once; the assistant monitors and acts without re-confirmation. This reduces friction for repeat purchases and restocking, but it also raises questions about consent granularity and behavioral nudging that Amazon has not publicly addressed.
Amazon also disclosed that free cash flow fell to $1.2 billion over the trailing 12 months, primarily due to a $59.3 billion increase in property and equipment purchases, with the company attributing much of this to AI investments. The scale of infrastructure spend underscores the capital intensity of running large language models at consumer volume.
What to watch
Practitioners building shopping or task-automation agents should monitor the execution permission model Amazon is using. Conditional automations that complete purchases without real-time user confirmation at trigger time are rare in consumer AI. If this pattern scales without user friction or regulatory pushback, it will shift expectations across e-commerce and subscription management.
The cross-device context continuity—phone to web to Echo—is table stakes for agents aiming at household use cases. Test whether your agent can retain conversation state and task status across device boundaries without requiring users to re-authenticate or repeat context.
Track whether third-party retailers gain material volume through Shop Direct. If Amazon's agentic assistant drives measurable purchase lift for non-Amazon sellers, it becomes a distribution channel worth integrating into. If adoption remains marginal, it remains a PR feature.