I've said this on the podcast more than once, because it cuts through a lot of confused strategy: TikTok knows attention, and Amazon knows intent. They're not really competitors. They're two stages of the same journey — one creates demand, the other captures it. Understanding that division of labour explains most of how modern product discovery works. And understanding what AI is now doing to it explains where the next few years are heading.
The two-stage machine
Start with the stable picture. TikTok — and social generally — is a demand-creation engine. Nobody opens TikTok intending to buy a specific thing; they're there to be entertained, and somewhere in that scroll a product catches them and plants a want. That's attention converted into desire. TikTok is extraordinary at it.
Amazon is a demand-capture engine. By the time someone's searching on Amazon, the want already exists — they're not discovering that they'd like the thing, they're deciding which version of it to buy. That's intent. And intent is the more commercially valuable of the two, because it's closest to the money. Attention is upstream and abundant; intent is downstream and scarce, and whoever owns the intent owns the moment the wallet opens.
For years this has been a comfortable sequence: attention created on social, intent captured on Amazon. The two platforms function in order, not in opposition, and the smart operator uses both — social to create the demand, Amazon to capture it. No conflict.
What AI just changed
Here's the disruption. AI-mediated discovery — both Amazon's own assistant and the open-web assistants like ChatGPT, Perplexity, Gemini and Claude — is moving in on the intent stage. The valuable one. The one Amazon has owned.
Think about the behaviour shift. Increasingly, when someone has intent — a real need, a purchase to make — their first move isn't necessarily to type keywords into Amazon. It might be to ask an AI a full question: "what's the best [thing] for [my situation]?" That question is pure intent, and it's now being asked somewhere other than the Amazon search box. Whoever answers it well is standing exactly where Amazon has always stood: at the point of intent, guiding the decision.
This is precisely why Amazon reacted the way it did. Putting an AI assistant into its own search bar as the default — the whole Rufus-into-Alexa-for-Shopping move of 2026 — wasn't Amazon chasing a trend. It was Amazon defending its half of the machine. If intent starts getting expressed to AI assistants, Amazon has to be the AI assistant people express intent to, or it slowly loses its grip on the most valuable stage in the funnel. The gatekeeper heard footsteps.
Two fronts, and why you need both
So for sellers, the intent stage has split into two fronts, and they're additive, not interchangeable.
Inside Amazon's own AI. Amazon's assistant reads Amazon's product graph — your structured listing fields, attributes, A+ content, reviews, Q&A. Winning here is about the listing discipline these posts have covered: complete, specific, question-answering source-document listings. This is Amazon defending intent on its own turf, and your job is to be the product it confidently recommends.
On the open web. The open-web assistants — ChatGPT, Perplexity, Gemini, Claude — answer intent expressed outside Amazon, and being recommendable there leans on a different set of signals: your own website, your structured product data, your reviews and mentions across the sources these models read. So a brand that sells both on and off Amazon needs a readiness posture for both. But here's where I want to correct a comfortable assumption, because I nearly wrote it myself: don't assume Amazon's catalogue is somehow harder for these open-web models to reference than your own website. If anything, the opposite is likely. Amazon is not going to sit back and let its product data be invisible to the biggest shift in discovery in a generation — it will be working hard to make its content exceptionally easy for LLMs to reference, probably more so than most brands manage with their own sites. Betting that your scrappy website out-references Amazon's catalogue to the open-web AIs is a bet I wouldn't take.
And the advantage most people forget. There's a reason Amazon became the number-one product search engine ahead of Google in many markets, and it isn't search technology. It's that Amazon sits inside the sale and the delivery. It knows the product is in stock, what it costs today, that it'll arrive tomorrow, and it removes almost all the friction between wanting the thing and having it. That end-to-end control — over the transaction and the fulfilment — is an advantage a pure search engine never had, and it's precisely what made Amazon win intent in the first place. There's every reason to expect that same advantage to carry into AI. An assistant that can not only recommend a product but confirm it's available, priced, and arriving tomorrow — and complete the purchase in the same breath — is answering intent more completely than one that can only point you somewhere. Amazon's moat was never just the search box; it was the friction it removed on either side of it. That moat doesn't disappear in the AI era. It may well deepen.
You don't get to pick one. The intent that used to flow cleanly into the Amazon search box is now fragmenting across Amazon's AI and the open-web AIs, and the demand created upstream on TikTok can now get captured at any of those points. Ceding any of them cedes intent — the valuable half.
The strategic read
Here's the through-line. The old model — create attention on social, capture intent on Amazon — still works, but the "capture intent on Amazon" half is no longer a single, guaranteed destination. Intent now gets expressed across multiple AI surfaces, and the operator's job is to be present and recommendable wherever it lands, not just in the one search box that used to hold a monopoly. But don't mistake "contested" for "Amazon's losing." Amazon is defending its half aggressively, making its data AI-friendly, and it still holds the delivery-and-friction advantage that won it intent in the first place. The likeliest outcome isn't Amazon being displaced — it's Amazon remaining a dominant place intent gets answered, now alongside others, rather than alone.
TikTok still knows attention. Amazon still knows intent — and is better placed than most to keep knowing it. The shift is that intent is no longer Amazon's alone to answer, so the operator prepares to be recommendable across all the surfaces where intent now shows up, while respecting that the one wired into delivery still starts with a structural edge.
FAQ
What’s the difference between TikTok and Amazon in the buying journey?
TikTok creates demand by turning attention into desire, while Amazon captures it at the intent stage, when the shopper already wants the product and is deciding which version to buy. Intent is the more valuable half because it is closest to the money.
How is AI changing product discovery?
AI assistants, both Amazon’s own and open-web ones like ChatGPT, Perplexity, Gemini and Claude, are moving into the intent stage. Shoppers increasingly ask an AI a full question about what to buy instead of typing keywords into the Amazon search box.
Is Amazon losing intent to AI?
Not necessarily. Amazon is defending its half by building AI into its own search and making its data easy for AI to reference, and it still controls the sale and the delivery. The likeliest outcome is Amazon remaining a dominant place where intent gets answered, now alongside other AI surfaces.
About the author
Zamir Cajee is co-founder of This Way Up, a UK business specialising in Amazon marketplace strategy, and co-host of The Upside Podcast, where he and the team break down how Amazon actually works and how it lies to you. Zamir has built multi-million dollar businesses from scratch and has been selling into the EU since 2016.
Not sure how recommendable your brand is across Amazon and the AI assistants? Reach out to us for tailored guidance.
Subscribe to the This Way Up newsletter for impactful strategies and insights, or tune in to The Upside podcast for expert advice on successfully managing your e-commerce efforts.