Disclaimer: The purpose of this article is to help Amazon sellers better understand the evolving landscape of product listing optimization—particularly in light of Amazon’s introduction of Gen-AI algorithms and related features. This content is intended solely for educational and informational purposes.
The insights and recommendations shared herein are based on public information, seller feedback, and independent analysis. This article is not affiliated with, endorsed by, or sponsored by Amazon.com, Inc. or any of its subsidiaries. It does not represent Amazon’s official views, policies, or future developments.
Sellers are encouraged to refer to Amazon Seller Central, official documentation, or consult their account representatives before implementing any changes to their operations based on this content. The authors of this blog assume no liability for actions taken based on the information provided.
--------------------
Several Amazon have recently reported a noticeable drop in product impressions on Amazon. Coincidentally, three days ago, Amazon announced the launch of a new AI-driven product explanation feature. These two developments remind us of what happened before last year’s Black Friday, when many listings disappeared from the first page of search rankings.
In this article, we will analyze Amazon's A9 algorithm and the new Gen-AI algorithm, as well as the newly introduced "Retailer Channel Product Promotion" ad feature. The goal is to help sellers understand how to regain lost traffic on Amazon.
Weekly Content Overview:
- Amazon AI Voice Shopping Expert Feature Introduction
- Comparison of Gen-AI Algorithm and A9 Algorithm
- Core Mechanism of Gen-AI and Listing Optimization
Part 1: Amazon AI Voice Shopping Expert Feature
On Friday, May 21, Amazon officially announced its new AI Voice Shopping Expert feature. According to Amazon, the feature:
- Is based on Amazon's Gen-AI shopping model (like Rufus)
- Appears below the main image on mobile product detail pages
- Analyzes product functions, features, customer reviews, and off-Amazon performance
- Allows customers to receive comprehensive, voice-based buying advice without needing to browse product pages themselves
Currently in beta testing, the feature is primarily targeted at:
- U.S. consumers
- Functional and complex products that require recommendations
Part 2: Comparison Between Gen-AI Algorithm and A9 Algorithm
Many sellers still rely on the traditional operational mindset of "impressions - clicks - conversions." However, since 2022, Amazon has been shifting toward a massive Gen-AI-based shopping model. As of 2024, AI shopping has become the core of Amazon’s algorithm. Below is a comparison of the two:
Why It's Time to Let Go of the A9 Algorithm:
- A9 was based on Google's keyword search model, introduced in 1998. It extracted meta titles, descriptions, and links to match with user queries.
- Amazon's A9, built after acquiring a Silicon Valley company in 2009, added e-commerce elements to the search algorithm.
- Ranking factors included:
- Keyword indexing (based on product titles)
- Ranking (based on click-through, add-to-cart, and conversions)
- Price, reviews, discounts, ads, and inventory influenced exposure and conversion, not ranking directly
- From 2018 onward, Amazon clamped down on ranking manipulation
- With increasing regulatory scrutiny (e.g., by the FTC), Amazon needed a more secure and fair system. Thus, the Gen-AI model was born.
The Logic Behind the Gen-AI Algorithm:
- 2020: Amazon started Gen-AI model development
- 2022: Introduced AI tools (image recognition, 3D modeling, virtual try-on)
- 2023: Urged sellers to complete attribute fields for AI training
- Feb 2024: Backend model "Cosmo" launched
- July 2024: Frontend model "Rufus" launched on mobile
- Nov 2024: Rufus expanded to desktop; listing titles shortened from 120 to 80 characters, the rest auto-generated by AI
- Nov 2024 Black Friday: Homepage redesigned to prioritize AI-driven recommendations over keyword search
- Spring 2025: "Interest" feature launched, guiding users via personal profiles
- May 2025: AI Voice Shopping Expert launched
- Amazon is guiding customers to:
Stop manual searching
- Stop reading reviews themselves
- Rely on Gen-AI for fair, authoritative product analysis
Part 3: Core of Gen-AI and Listing Optimization
Amazon's Gen-AI is a large, self-evolving model that affects both frontend and backend. Compared to traditional A9-based exposure, Gen-AI traffic is more massive, accurate, and high-converting. Key ranking and recommendation priorities include:
- Conversion rate: Still the #1 factor
- Attribute completeness: Precise details on color, material, size, usage, target audience
- Price competitiveness: Participation in Deals (BD/LD) increases exposure via Rufus
- AI functionality: Add 3D models, virtual try-ons, interactive demos
- Review summary tags:
- Green check = recommended
- Red dash = not recommended (hurts conversion)
- Return rate: Avoid "Frequently Returned Item" tag by adding detailed videos and manufacturer explanations
- Brand search volume: Higher brand visibility leads to higher recommendation priority
- Repeat purchase rate: Shown in Brand Analytics; signals loyalty
- Off-Amazon influence: Social media, third-party reviews, PR mentions
- Trend alignment: Product relevance to trending categories increases recommendation chances
Additional Recommendations:
11. Keep product titles within 80 characters
12. Use ChatGPT to test if your images/videos/A+ are readable by Amazon's AI
Strongly advised:
- Do not edit existing listings directly
- Instead, use Seller Central A/B Experiments to test AI-generated content
- Compare ad performance between old and new versions to assess effectiveness