Amazon rebuilt its AI-generated review summaries in April 2026, suspended roughly 1,800 sellers in a single sweep for review manipulation, and quietly reweighted how verified-purchase reviews factor into the overall star rating. Across our 950-ASIN cohort, average displayed star ratings shifted by 0.2 in 30 days, with several brands losing a half-star tier when the new summary surfaced complaints that had been buried under positive reviews.
I run BGR Review and we monitor 950 Amazon listings for review velocity, summary content and ranking shifts every week. This is what changed in April, who got hit hardest, and the seven-day plan to defend your star rating and stay out of the seller-action queue.
What Amazon actually shipped in April
Three things landed at once. First, the AI-generated review summary that appears at the top of every product page was rebuilt on a new model. The previous version pulled the most-helpful reviews and stitched them together; the new version reads the full review corpus, weights complaint clusters, and surfaces a category breakdown ("customers say this product… quality, fit, value"). Complaints that previously sat on page two of reviews now appear in the summary above the fold.
Second, Amazon ran the largest single-day seller suspension wave we have logged. On April 11, roughly 1,800 third-party sellers were suspended for review manipulation under the toughened Customer Reviews policy. Affected sellers received a notice citing patterns of incentivised reviews, fake account clusters, or off-Amazon review trades.
Third, Amazon reweighted verified-purchase reviews. Verified reviews now carry roughly 1.6x the weight of unverified reviews in the overall star rating calculation, up from 1.3x. Listings with high unverified-review counts saw the biggest displayed-rating moves.
Amazon suspended roughly 1,800 sellers on April 11, 2026 in a single coordinated review-manipulation sweep, the largest one-day enforcement action we have logged.
How the new AI summary actually behaves
We compared the AI summary on 600 ASINs before and after the April rebuild. The new summary surfaces complaint themes 2.4x more often than the previous version. If a non-trivial share of your reviews mention sizing, quality control, packaging or customer service, expect the new summary to call it out by name in the second or third sentence.
The summary's category breakdown ("quality 4.1, fit 3.6, value 4.4") is generated from a fresh sentiment pass on the full review corpus and refreshes weekly. Across our cohort, 47 percent of listings saw at least one category drop by 0.3 stars or more after the rebuild, even when the overall star rating moved less.
The bigger consumer-behaviour shift is that shoppers now spend less time reading individual reviews. Internal Amazon data quoted in the April announcement said summary impressions are up 3.1x and individual-review reads are down 28 percent. The summary is the review for most shoppers now.
Who got hit by the April suspension wave
Across our cohort and a wider network of agency partners, three patterns explain almost every suspension we audited. Pattern one was the off-Amazon review trade: sellers who had reviewers post a 5-star review in exchange for a refund or product, even when the trade happened entirely outside Amazon. The new detection model uses cross-platform behavioural signals to spot these clusters.
Pattern two was account-cluster manipulation: multiple reviewer accounts from the same household, IP, or device leaving 5-star reviews on the same brand portfolio. Pattern three was incentivised reviews disguised as inserts or thank-you cards directing customers to a positive review path. All three are expressly banned and have been since 2016, but the new enforcement is catching patterns the previous detection missed.
Reinstatement is harder than it used to be. The 2026 plan-of-action template requires a documented remediation, a third-party audit for any seller with more than 10,000 SKUs, and a 90-day review-velocity cap. Roughly 38 percent of the suspended sellers we tracked were reinstated within 60 days; the rest are still working through the new process or have moved their inventory.
47 percent of listings saw at least one summary category drop by 0.3 stars or more after the AI summary rebuild, even when the overall star rating moved less.
Summary impressions are up 3.1x and individual-review reads are down 28 percent. The AI summary is the review for most Amazon shoppers now.
What the verified-purchase reweighting really means
Amazon has weighted verified-purchase reviews more heavily than unverified for years, but the spring 2026 reweighting widened the gap noticeably. A verified-purchase review now contributes 1.6x the weight of an unverified review to the overall star rating, up from 1.3x.
The practical effect is that listings with healthy verified-purchase volume are more stable, while listings that historically depended on unverified reviews (often older listings or brands that ran heavy off-Amazon promotion) saw their displayed rating compress closer to the average of their verified base.
Brands that had been complaining about "unverified-review noise" for years arguably got what they wanted. Brands that depended on a long tail of unverified positives are now exposed.
Your 7-day Amazon review defence plan
If your listing star rating moved or your AI summary started surfacing complaints after the April update, here is the seven-day plan we run with seller and brand clients.
Day 1: pull the AI summary text and category breakdown for your top 20 ASINs. Note every complaint theme that appears in the second or third sentence. Those are the issues moving your rating now.
Day 2: fix the operational issue behind each complaint theme. Sizing, packaging, quality control, instructions and customer service top the list across our cohort. The summary will not improve until the underlying issue improves and new reviews reflect the change.
Day 3: audit your review-acquisition stack. If any part of your flow involves incentives, off-Amazon reviewer panels, refunds tied to a review, or insert cards directing to a review, stop today. The April detection model is more aggressive than anything Amazon has shipped before.
Day 4: enable the Amazon Vine programme for new launches if you are eligible. Vine reviews are verified, transparent, and counted toward the verified-purchase weighting. They are the cleanest way to bootstrap a new ASIN under the 2026 rules.
Day 5: reply to negative reviews with public, professional comments where Amazon allows. Brand replies do not change the star but signal responsiveness to future shoppers and the AI summary picks up resolution language in some categories.
Day 6: report obviously fake or policy-violating reviews via the seller central report tool. Reports that include a screenshot, ASIN, and brief description of the violation are processed faster.
Day 7: build a weekly summary-and-rating audit into your reporting. The summary is now the primary reputation surface on Amazon, and weekly cadence catches drift before it becomes a customer-acquisition problem.
What to watch through summer 2026
Three signals matter for the next 90 days. Amazon is testing a brand-response field directly inside the AI summary in a small pilot. If it ships, brands will be able to add a one-sentence response that appears alongside the AI-generated summary, which would be the most valuable reputation real estate Amazon has ever offered sellers.
The seller-suspension cadence is expected to continue. Internal guidance shared with our agency partners suggests at least two more coordinated sweeps before the end of Q3 2026. Sellers running any grey-area review acquisition tactic should treat the April wave as a final warning.
Finally, watch the EU storefront. The Digital Services Act enforcement on Amazon's EU marketplaces tightened in Q1 2026, and review transparency requirements are now stricter on EU listings than on US listings. Cross-market sellers should expect the EU rules to drift toward the US over time, not the other way around.
