Amazon CEO about AI Shopping: Retailers vs. Third-Party Agents

18 Feb 2026

Amazon CEO about AI Shopping: Retailers vs. Third-Party Agents

Author: Oleksandr Kovalov

Role: Founder & CEO @ ANavigator

Amazon’s CEO, Andy Jassy, said something last week that didn’t get nearly enough attention. He told shareholders that when AI-powered shopping becomes mainstream, customers will end up buying through retailers — not through standalone AI platforms like ChatGPT. His reasoning wasn’t complicated: retailers already control everything that actually gets someone to check out. Selection, pricing, delivery speed, and the kind of trust that takes years to build.

This isn’t a debate about which AI model is more advanced. It’s a conversation about structural advantage — and who already has it.

Here’s our take on what this means for brands and sellers.


AI Can Recommend. Retailers Can Convert.

Third-party AI tools are good at one thing: surfacing options. But when it comes to actually closing a sale, they’re missing most of the infrastructure that makes that happen. They don’t have real-time inventory data. Their pricing logic isn’t always verified. They can’t access your order history or behavioral patterns, and they have no fulfillment layer to stand behind.

Retailers have all of that already built in.

Amazon doesn’t need to estimate when your package will arrive — it controls that answer. It doesn’t need to guess whether a product is in stock — it knows. That kind of integration is worth more than a smarter model when the goal is getting someone to actually complete a purchase.


Rufus: The Data Behind the Thesis

Amazon’s own numbers back this up.

In 2025, nearly 300 million customers used Rufus, Amazon’s AI shopping assistant. Shoppers who engaged with Rufus were around 60% more likely to complete a purchase compared to those who didn’t. Over the 2025 holiday season, AI-driven traffic to retailers jumped close to 700%.

That traffic didn’t stay in external chatbots. It ended up inside retailer ecosystems.

The pattern is consistent across the data. People browse widely, they use AI to explore and compare, but when they’re ready to buy, they go back to the platform where their account already lives, where their payment details are saved, and where they’ve bought before without issues.


AI Hype vs. What Actually Drives a Sale

A lot of the conversation in the market right now is about which AI system is the smartest. That’s the wrong question for commerce.

Intelligence alone doesn’t close a sale. What closes a sale is a saved credit card, a reliable returns process, a next-day delivery promise, and a history of the customer trusting that the product will show up as described.

None of that can be copied overnight by an external AI agent. It took Amazon years of infrastructure investment, logistics buildout, and consistent customer experience to get there. An AI interface sitting outside that ecosystem starts from zero on all of those dimensions.

Amazon isn’t winning this race because its AI is better. It’s winning because its AI sits on top of something that already works.


The Real Advantage Is the Relationship, Not the Model

The most durable competitive advantage in AI commerce isn’t technological — it’s relational.

Retailers already hold customer accounts, purchase histories, behavioral data, loyalty programs, and fulfillment networks. When you layer AI on top of a relationship that already exists and a system the customer already trusts, you get leverage. When you layer AI on top of a blank interface with no history, you get friction.

This isn’t a new pattern. Distribution and embedded relationships have consistently outperformed standalone technological breakthroughs throughout commerce history. The best technology doesn’t always win. The most embedded technology usually does.


What This Actually Means for Sellers and Brands

If Jassy’s thesis plays out — and the data suggests it already is — a few things follow from that for anyone selling on Amazon.

AI optimization inside Amazon’s ecosystem is going to intensify. The tools Amazon builds natively will have more influence over discovery and conversion than anything happening outside the platform. Clean listings, structured data, and strong content are going to matter more as AI-assisted search becomes the default way people shop, not the exception.

And brands that have been leaning heavily on external traffic funnels without strengthening their in-platform presence are going to feel that gap. If AI embeds itself deeper into how Amazon surfaces and recommends products, the brands that are already well-optimized inside that ecosystem will pull further ahead.


The Question That Actually Matters

In an AI-driven commerce environment, what wins — the smartest standalone technology, or the strongest existing customer relationship?

Amazon is betting on relationship, infrastructure plus AI integration together. Given that they’re starting with 300 million active customers, years of behavioral data, and a fulfillment network most companies will never replicate, that’s not a bold bet. It’s a logical one.

For sellers, the direction is straightforward: optimize for the platform where the transaction happens — not just for where the conversation starts.

 


If you want more analysis on how AI is shifting Amazon’s performance, advertising systems, and marketplace dynamics, follow the ANavigator blog.

If you need help with PPC, DSP, analytics, or a long-term growth strategy, reach out to the ANavigator team at info@anavigator.co

Visit our website: anavigator.co

Follow&Subscribe to our ANavigator Weekly Digest on LinkedIn.

— The ANavigator Team

Follow us on LinkedIn:

our activity

News and podcasts

LATEST UPDATES

Amazon’s Reference Pricing Crackdown: The Strike-Through Era Is Getting Harder to Game
Blog
April 9, 2026
Amazon’s Reference Pricing Crackdown: The Strike-Through Era Is Getting Harder to Game
Amazon is closing one of the most widely used conversion tactics on the platform. Starting April 23, 2026, the rules around reference pricing are changing in ways that will force many sellers to rethink how they display discounts — and for some, that change will hit hard. This is not a minor policy clarification. It is a structural shift in how Amazon validates the prices that appear crossed out on product pages. If your listings rely on strike-through pricing to communicate value to shoppers, the window to get this right is already open. What Amazon Is Actually Changing There are two separate updates, rolling out at different times. The first takes effect on April 23, 2026, and targets the List Price — what most sellers know as the Manufacturer's Suggested Retail Price. To meet Amazon's updated requirements, a List Price must fulfill one of two conditions: either the product must have been sold at that price by another retailer within a recent timeframe, or it must have been purchased on Amazon as the Featured Offer at the same price. What that rules out is a List Price that exists only on paper. A figure you entered into Seller Central months ago and never actually sold at — or that no other retailer has ever charged — will no longer qualify. If Amazon cannot verify it, it will not display it. The second update lands on May 18, 2026, and changes how Typical Price is calculated. Under current rules, Typical Price reflects the median non-promotional price customers paid for a product over the past 90 days, excluding promotional sales. Starting May 18, this exclusion will no longer apply automatically. If a product's Featured Offer price is below its non-promotional median for more than half of a 90-day period, Amazon will include all sales — promotional and non-promotional — in its Typical Price calculation. In practical terms, sellers running prolonged discounts or price campaigns for more than 45 days within the 90-day window may see their Typical Price recalibrated downward, potentially eliminating the strike-through pricing used to attract customers.     Why Amazon Is Doing This Amazon has linked these reference pricing updates to its Price History Graph, which is visible on product detail pages. The graph records the lowest Featured Offer price each day, offering shoppers a transparent view of a product's pricing history. The message is clear: reference prices need to reflect reality, not just support a discount narrative. Under previous conventions, sellers could submit a List Price that served primarily as a reference point for discount percentages rather than as a price at which actual transactions occurred. That approach is now being closed off directly. This also fits a broader pattern. Amazon ended FBA commingling on March 31, restructuring inventory accountability across its fulfillment network. Each of these changes moves in a consistent direction: greater specificity, more granular accountability, and less operational latitude for sellers who have relied on platform flexibility. What This Means for Conversions and Advertising Strike-through pricing works because it gives shoppers a reference point. When the crossed-out number disappears, the perceived value of the deal weakens with it. For listings that have been built around a prominent discount display, this is a conversion problem — not just a compliance one. From the advertising side, campaigns built around promoted product listings that display meaningful savings percentages will need re-evaluation if the underlying reference prices are invalidated. A Sponsored Products campaign driving traffic to a product that no longer shows a reference price will convert at a lower rate, increasing effective cost per acquisition without any change to bid strategy or keyword targeting. That is a real cost that does not show up in your bid data — but will show up in your ACOS. What Brands Should Review Now There are several things worth auditing before April 23 arrives. Check which of your ASINs currently display a List Price or strike-through, and verify whether that List Price is substantiated — either by real Amazon sales history at that price or by verifiable pricing at another retailer. If neither exists, the display will be removed. Review your promotional calendar against the 90-day pricing window. Amazon will only display a suggested List Price if it follows Amazon pricing policies and meets the necessary savings threshold. If Amazon identifies an inflated List Price, not only will the display be disabled, but your account may be flagged for a violation of Amazon policy. Think carefully about how long your discounts run. Under the new Typical Price rules, a deal that runs for more than half of any 90-day window effectively becomes your new baseline. Running a permanent promotion is no longer a way to show savings — it becomes the price itself. The Bigger Point For years, inflated reference prices were a relatively low-risk way to make a discount look more meaningful than it was. That approach is not going away entirely, but it is getting much harder to sustain without real pricing evidence behind it. The brands that will navigate this best are the ones that build a genuine pricing strategy — not just promotion mechanics. That means understanding your price history by ASIN, being deliberate about when and how long you run discounts, and ensuring your List Price reflects something a customer could actually verify if they checked. Amazon is not just enforcing a policy here. It is changing what a discount means on its platform. For sellers who have relied on strike-through pricing to drive conversions, this is one of the more important operational changes of the year. The deadlines are closing, and the review work needs to start now.  Book a call to get a FREE AUDIT by the link below:     Book a call FREE AUDIT   Follow my Weekly Newsletter on LinkedIn:  / amazon-digest-for-brands-7232361008185372672   Follow me on LinkedIn:  / ookovalov  Follow ANavigator on social media:  / anavigator    /@anavigator_official  / anavigator7    / @anavigators   LinkedIn page to contact us:     Author: Oleksandr Kovalov Founder & CEO @ ANavigator — The ANavigator Team
Learn more
Prime Day 2026 Rewards Brands That Plan Earlier, Not Louder
Blog
April 8, 2026
Prime Day 2026 Rewards Brands That Plan Earlier, Not Louder
Prime Day 2026 is already underway — at least from a planning perspective. Amazon opened the deal submission window on March 24 and has set a May 26 cutoff for deal submissions, with FBA inbound deadlines shortly after. That alone is a signal that brands that treat this as a June or July problem are already behind. There is also more cost pressure this year than many teams are expecting, and the structure has changed in a way that matters.     What the New Fee Structure Actually Means Amazon has revised its Prime Day deal fee structure from a flat fee to a performance-based model. Sellers now pay a $100 upfront fee plus 1.5% of promotional sales, capped at $5,000, instead of the previous $1,000 flat fee for Event Best Deals. On the surface, this looks like a lower barrier to entry. And for smaller brands running a handful of SKUs at moderate volume, it may well be. But for brands with strong sales velocity, the variable component adds up quickly, and it needs to be part of the margin math before a deal is scheduled — not after. There is also an early submission incentive: if you schedule your Lightning Deal or Best Deal before April 30, 2026, you receive $50 off the upfront fee per deal, bringing the cost down from $100 to $50. For sellers running multiple deals, this can reduce promotional costs significantly. That discount is real, but it only makes sense if the underlying deal is already profitable. Submitting early to save $50 on a deal that does not work is not a saving — it is a locked-in loss.     Why This Is More Than a Campaign Task This is where many brands get Prime Day wrong. They treat it like a campaign setup task, when in reality it is a business planning task across pricing, inventory, margin, and ad budget discipline. By the time campaigns go live, those decisions should already be made. The pricing threshold is also unforgiving this year: your deal must be at least 5% lower than the lowest price offered in the trailing 30 days. The old approach of raising a price before discounting it is no longer viable. That means pricing decisions you make now, weeks before the event, will directly affect which deals you can run and at what depth. Sellers who spread thin across 20 deals with shallow discounts consistently underperform sellers who go deep on three to five high-velocity products. Prime Day rewards concentration. The Inventory and Advertising Reality Getting the deals right is only part of the equation. CPC rates during Prime Day spike 40% to 80% above normal levels. Brands that set advertising budgets based on their regular weekly spend will either run out of budget mid-event or dramatically overpay for traffic that converts poorly. The budget planning conversation needs to happen now, not in late June. On inventory, the inbound arrival cutoff for minimal splits is May 27, and for optimized splits it is June 5. If you source internationally, those dates are not six weeks away — they are already inside your manufacturing and shipping lead times. A supplier order placed in late April may not make it.     What Brands That Execute Well Actually Do Differently The real difference between strong and weak Prime Day execution is usually not creativity. It is preparation. Brands that start earlier have time to: Choose the right SKUs rather than defaulting to their full catalog, review profitability before committing to a discount depth, secure enough stock without over-indexing on inventory that may not move, plan advertising budgets for both branded and non-branded traffic, and decide where to push harder and where to protect margin. The brands that perform best are not the ones that react late with bigger spend. They are the ones who go into the event with a clearer plan, better numbers, and more control. What to Watch Between Now and May 26 This is still an evolving picture, with official Prime Day dates not yet confirmed. But there are several things worth tracking closely: whether the event lands in late June or the traditional mid-July window, how the $100 plus 1.5% fee structure plays out in practice for different volume tiers, whether the early submission discount meaningfully changes deal quality across categories, how FBA capacity constraints affect inventory positioning closer to the cutoff dates, and whether Amazon tightens pricing eligibility rules further as the window approaches. Amazon typically does not publish its full promotional calendar far in advance, which leaves many sellers reacting to last-minute announcements. That reactive approach often leads to stockouts during critical periods or overspending on inefficient advertising. On Amazon, stronger Prime Day results usually start well before Prime Day itself. The window to make the decisions that matter is open right now — and it closes faster than most teams realize.    Book a call to get a FREE AUDIT by the link below:     Book a call FREE AUDIT   Follow my Weekly Newsletter on LinkedIn:  / amazon-digest-for-brands-7232361008185372672   Follow me on LinkedIn:  / ookovalov  Follow ANavigator on social media:  / anavigator    /@anavigator_official  / anavigator7    / @anavigators     LinkedIn page to contact us:     Author: Oleksandr Kovalov Founder & CEO @ ANavigator — The ANavigator Team
Learn more