Amazon Is Matching Products to Buyers, Not Keywords. Two Accelerate Updates Show How

TL;DR

What to expect from this article: What Amazon announced about Alexa for Shopping and Amazon Business at Accelerate 2026, how brands can take advantage of this by being clear about who their product is for, and the specific listing and pricing changes in light of this.

Key takeaways:

• Alexa for Shopping reads your entire detail page and searches the web, including review sites and publications, to decide whether your product fits a shopper's need. A listing now has to answer a question, not just match a keyword.

• The fastest fix is on your top SKUs: state plainly who the product is for, the problem it solves, and the concrete details (size, materials, age range, certifications).

• What's said about your brand off Amazon now feeds the match. Third-party reviews and press are part of your listing, whether you manage them or not.

• Amazon Business buyers order bigger, convert better, and return less than consumers, and many are already buying your products at consumer prices.

• The risk with Amazon Business Deals and quantity discounts is discounting volume you'd have sold anyway. Set tiers that start above the order size business buyers already place.

What did Amazon announce about Alexa for Shopping and Amazon Business at Accelerate 2026?

Amazon used Day 2 of Accelerate 2026 to make two related points: shoppers are increasingly finding products through a conversational assistant rather than a search box, and business buyers are a large, under-served segment already buying on Amazon. Both are live now, and neither requires sellers to enroll in anything new to be affected.

The common thread is that Amazon is getting better at matching a product to a specific kind of buyer. For consumers, that's the shopper's situation and needs. For organizations, it's the buyer type, order size, and the certifications procurement teams screen for. In both cases, brands that state clearly who the product is for get matched more often.

How does Alexa for Shopping decide which products to recommend?

Alexa for Shopping matches a shopper's stated need to products it understands, using the full detail page plus what it finds about the product on the web. It isn't just matching search terms. Amazon replaced Rufus with Alexa for Shopping in May 2026, and on Amazon's Q2 2026 earnings call (July 30, 2026), Andy Jassy said more than 350 million customers had used it in the past 12 months, and that US customers who use it spend 40% more per order on average.

Three things about how it works matter for sellers:

• It remembers the shopper. Preferences, past purchases, and budget carry across devices. A shopper who has told Alexa they have a toddler, a small apartment, or a $50 ceiling doesn't repeat it, and products get filtered against that context.

• It reads the whole page. Title, bullets, description, A+ Content, and product details are all inputs. Information buried in an image or left out entirely is information the assistant can't use to recommend you.

• It looks beyond Amazon. Amazon said on stage that Alexa for Shopping also searches the web, including publications and review sites, to match products. Your off-Amazon reputation is now an input to on-Amazon discovery.

Amazon's own framing was direct: discovery is shifting from keyword optimization to clearly explaining what problem a product solves and who it's for. Keywords still matter for traditional search, which hasn't gone away. But a listing that ranks well on keywords and never explains its use case is increasingly invisible inside the conversation.

How should you optimize a listing for Alexa for Shopping?

Make sure each top SKU's page states, in plain sentences, who the product is for, the problem it solves, and the concrete details a shopper would ask about. Most listings already contain some of this, but scattered across keyword-stuffed bullets in a way an assistant can't reliably interpret.

Start with your top 10 SKUs by revenue and check each page for four things:

1. Who it's for. One sentence naming the buyer and situation: "for renters who can't drill into walls," "for kids ages 3 to 6," "for dogs over 50 pounds."

2. The problem it solves. Stated as the problem, not the feature. "Keeps cords off the floor so a robot vacuum doesn't get stuck" answers a question; "premium cable management system" doesn't.

3. The concrete details. Size, materials, weight limits, age range, compatibility, and certifications, written as text. Specs that only appear in an infographic are invisible to a text-reading assistant.

4. Who it's not for. If the product isn't right for a common situation, saying so helps the assistant match it correctly and reduces returns from mismatched buyers.

A quick test: ask Alexa for Shopping the question a real buyer would ask, such as "what's a good leash for a dog that pulls," and see whether your product comes up and how it's described. If the assistant's summary of your product is vague or wrong, the page isn't giving it enough to work with.

In fact an interesting example is a certain pet brand selling dog chews had Alexa suggesting questions about the safety of their product. Since this was never raised as a concern, and its food safety was assumed, the listing’s details did not explicitly address this question, so when buyers clicked Alexa’s suggested question, it responded uncertainly like “There is no direct mention of…”. We thus optimized their listing, including information that explicitly answers Alexa.

Does what's said about your brand off Amazon affect Alexa for Shopping?

Yes. Because Alexa for Shopping searches publications and review sites to match products, third-party coverage now feeds on-Amazon discovery. A product that appears in reputable reviews, comparison articles, or category roundups gives the assistant more confidence about what it is and who it suits. A product with no outside footprint, or a negative one, gives it less.

This is the same shift that's happening in Google and ChatGPT search: AI assistants recommend what they can verify from multiple sources. In practice, that means:

• Search your brand and top products outside Amazon. Note which review sites and publications mention you, what they say, and whether the details match your listing.

• Fix inconsistencies. If a review site lists outdated dimensions or a discontinued variant, that's now a discovery problem, not just a PR one.

• Treat earned coverage as a listing input. Product seeding, review outreach, and creator content were already brand-building tactics. They now also influence whether Amazon's own assistant recommends you.

Should your brand sell to Amazon Business buyers?

If you sell in office, IT, maintenance, janitorial, education, or consumables, almost certainly yes, and you may already be doing it without a strategy. According to Amazon's July 21, 2026 announcement, Amazon Business reached $60 billion in annualized gross sales in Q2, serving more than 11 million organizations worldwide. On stage at Accelerate, Amazon added that business buyers place orders nearly 2.5 times larger than consumers on average, convert at three times the rate, and return 50% fewer items.

Those three numbers together make business buyers unusually profitable customers: bigger baskets, less acquisition waste, and lower return costs. They also shop differently. Business buyers look for:

• Certified products. Many procurement teams have diversity or sustainability purchasing targets and filter for small, women-owned, veteran-owned, minority-owned, or sustainable suppliers.

• Bulk quantities and large packs. A 12-pack that makes no sense for a household can be exactly what an office manager or school district wants.

• Services like custom kitting. Bundled supplies assembled for a specific use, such as onboarding kits or classroom sets.

The catch is that many brands in these categories are already selling to business buyers at consumer prices, without knowing it. That's the starting point, and it's also where the main risk sits.

How do you set Amazon Business pricing without giving away margin?

Set quantity discount tiers that start above the order size your business buyers already place, so the discount buys incremental volume instead of subsidizing orders you'd have won anyway. Discounting volume you would have sold at full price is the most common way B2B pricing quietly erodes margin.

The math is unforgiving. Take a SKU that sells for $40 with a $14 contribution margin per unit (35%). A 10% business discount costs $4 per unit and cuts the margin to $10. To earn the same total profit, you now need 40% more units. If your business buyers were already ordering two or three units at a time and your first discount tier starts at two, you've given away $4 a unit on demand that needed no incentive.

So a safer sequence:

1. Check your current business share. Seller Central's Business Reports can be filtered to show B2B sales by ASIN. Find out what share of revenue already comes from business buyers and their typical order size.

2. Set tiers above existing behavior. If business buyers typically order three units, start the first discount at five or six. The discount should change behavior, not reward it.

3. Use large packs instead of deep discounts. A dedicated multi-pack lowers your per-unit fulfillment cost, which funds the discount without coming out of margin.

4. List every certification you hold. Certifications cost nothing to display and put you in front of procurement filters you'd otherwise never appear in.

5. Test Business Deals on a few SKUs first. Compare the incremental volume against the margin given up before rolling exclusive pricing across the catalog.

Fee discounts on bulk orders, which Amazon also highlighted, can soften the math. But treat them as a bonus to the margin calculation, not as the reason a discount works.

What should you do this quarter?

Both updates are live, and neither needs a big project to act on. The highest-return steps for Q4:

• Rewrite the top of each top-10 SKU's listing so it states who the product is for, the problem it solves, and the concrete specs in text.

• Ask Alexa for Shopping the questions your buyers ask, and check how it describes your products.

• Audit where your brand appears in third-party reviews and fix any outdated or wrong details.

• If you're in a B2B-relevant category, pull your current business-buyer share before setting any discount tiers.

• Add every certification you hold to your listings and Amazon Business profile.

Related reading

For how Amazon's other AI update affects your account, see Amazon's Seller Assistant Can Now Watch Your Account. It Still Can't See Your Business. If you want help with either update, see Kyzenn's Amazon listing optimization and Amazon deals and promotions services.