Amazon doesn’t just want to sell you things anymore. Now it wants to sell other retailers the ability to sell you things — using the same AI that drove $12 billion in incremental sales on Amazon.com last year.
This week, AWS announced the Agentic Shopping Assistant (ASA) — a packaged version of the technology behind Alexa for Shopping, available for any retailer to drop into their website in roughly 60 days. Kate Spade is the first customer, with an “AI Gift Concierge” already live on KateSpade.com. You describe the occasion, the recipient, the vibe — and it recommends a gift. No search bar. No filter dropdowns. Just a conversation.
That’s a small detail that changes everything.
What Amazon Actually Did
Alexa for Shopping — the same system that merged Amazon’s Rufus chatbot and Alexa+ earlier this month — was used by over 300 million people last year. Conversational shopping sessions convert at 3.5x the rate of traditional keyword search. Amazon tested this at massive scale, called itself “Customer Zero,” and now it’s packaging the playbook for everyone else.
AWS provides the foundation: Amazon Bedrock, AgentCore for auth and observability, OpenSearch for catalog retrieval. Retailers bring their own product data and brand rules. Kate Spade’s implementation runs on Anthropic’s Haiku 4.5 model. The whole thing spins up in two months instead of two years.
This is infrastructure commoditization happening in real time. The AI shopping assistant is no longer a competitive differentiator for Amazon — it’s a revenue stream. And that tells you exactly how confident they are that the next battlefield is somewhere else entirely.
The Creator Economy Is Downstream of This
Here’s where it gets interesting for platforms like mine.
Most creators treat brand partnerships as a separate lane from their content — a deal gets signed, an affiliate link goes in the bio, and that’s it. But agentic commerce is about to collapse the distance between discovering a product and buying it. When a fan asks Kate Spade’s AI “what should I get my mom for her birthday?” and gets a curated recommendation with a buy button — that’s a creator’s job, done by a chatbot.
Creators who understand this early will use it as leverage. Those who don’t will find their affiliate revenue eaten by AI concierges that don’t need a discount code.
For Tonimus specifically, this reshapes how we think about content-to-commerce linkage. Right now, a creator posts, a fan engages, maybe clicks a link (we hope). That funnel is long and potentially leaky. As more retailers deploy conversational shopping assistants, the content a creator publishes becomes the top of a funnel that can close much faster — if the infrastructure is there to connect them.
The aggregated behavioral data Tonimus builds across creators isn’t just useful for growing audiences anymore. It’s increasingly valuable for understanding what kinds of content actually convert — not just hearts and views, but purchase intent and which hooks become clicks. Which formats best drive discovery. That’s data Amazon is sitting on for their own catalog. Tonimus could be building it for the creator economy.
The Bigger Picture
Amazon licensing this technology is a shot across the bow at everyone building in the commerce-meets-content space. Shopify has its AI tools. Stripe just launched its Agentic Commerce Suite. Google’s been pushing Shopping AI for years. The race isn’t to build the best product anymore — it’s to own the layer where intent becomes transaction.
Creators live at the exact intersection of intent and trust. Their audiences don’t just follow them — they ask them what to buy, where to eat, what to wear. That’s not influence. That’s agentic commerce, with a human face.
Platforms that help creators understand and monetize that relationship — with real data, real automation, and real attribution — aren’t just tools. They’re becoming commerce infrastructure, or what we like to call CreatorIntel.
Amazon knows it. The question is whether creators do.