Stella AI Search Lab|Tracking product and retailer visibility in AI search
A public research dashboard tracking how generative AI recommends and displays products and retailers for retail and e-commerce queries.
夫婦と子ども2人の4人家族が主要家電を買い替える際の推薦・販売元露出を観測する。キーワードプランナーrawdata（2026-08-16取得・15,239語）へ全文突合した現行設計。旧v1は2026-08-16観測分を最後に退役・削除済み。
Highlights for Selected Period
—Retailers That Moved
Only retailers whose ranking changed from the immediately preceding equal-length period are shown.
Leader by Category for Selected Period
The retailer with the strongest visibility in each appliance category during the selected period.
Observation Status
Run the same study for your products
This appliance dataset is one example. We can apply the same method to your products, competitors and target customers.
Daily Trend for Leading Retailers
Coverage = queries where the retailer appeared ÷ observed queries.For example, 24 of 25 equals 96%. First-position rate shows ranking strength, while Top 3 rate shows the depth of high visibility.
—
Ranking for Selected Period
Compared by first-position rateFirst-position rate: the share of queries where the retailer was recommended first.
Retailer Visibility Ranking
Aggregates query-by-day observations in the selected period and ranks retailers by breadth and position. Select a retailer to open its report.
| Rank | Retailers | Type | Queries? | Coverage? | First-position rate? | Second-position rate? | Third-position rate? | Top 3 rate? | Product links? | Appearances? |
|---|
Product Visibility Ranking
Shows which products AI recommended for the selected persona. Rates use the number of observed queries in the product's category as the denominator.
| Rank | Product | Categories | Format | Queries | Coverage | First recommendation rate | Top 3 rate | Average recommendation rank | Retailers |
|---|
—
Query → Product → Retailer
Showing queries from all categories
Queries were submitted in Japanese and are retained below in their original wording as source evidence.
The Stella AI Search Lab is a retail-focused public research project that continuously observes how products and retailers are discovered and presented by AI as generative AI becomes a gateway to shopping. Rather than reducing the results to winners and losers, we aim to create a shared, transparent way for anyone to understand how the AI shopping landscape is changing.
Why We Publish This Research
Recommendations generated by AI are harder to inspect than conventional search rankings, and the same query can produce different answers. That is why we make our method and results as open as possible: so retailers, e-commerce businesses, manufacturers, site operators and shoppers can examine the same data and think about the changes together.
This dashboard does not claim to provide a single correct optimization strategy. By observing how AI understands and recommends products and points people to places to buy, it provides evidence for considering what information is getting through and what is still missing.
What You Can Learn Here
- Which retailers appeared across the widest range of queries
- Which retailers tended to appear near the top of answers
- How visibility differed by category and price range
- Which products, sources and retailers were connected in AI answers
What We Observe
We prepare fixed shopper personas and queries, then submit them through the ChatGPT web interface. We record the answer, citations and sources, product cards, and the retailer list for each product in the order shown.
How Rankings Are Calculated
Within each query, retailers are ranked using the following sequence.
- More products linked to that retailer in the answer
- If tied, the higher position in the retailer list
- If still tied, the greater number of retailer appearances
All rates use the number of successfully observed queries as their denominator.
- Coverage: share of queries where the retailer appeared at least once
- First-position rate: share of queries where the retailer ranked first
- Top 3 rate: share of queries where the retailer ranked in the top three
How to Read the Data
- Generative AI answers can change each time, even for the same query. A daily observation is a point-in-time snapshot, and a ranking change does not necessarily indicate the effect of an optimization measure.
- The current dataset covers ChatGPT only and does not represent other generative AI systems.
- Price, stock and delivery text is recorded as displayed in the AI answer and may differ from the retailer's actual selling conditions.
- Product and retailer naming variations are normalized, with normalization decisions retained in the source research records.
- This dashboard does not judge retailer quality or superiority; it records only how retailers appear in AI answers.
- When considering a purchase or contract, please verify prices, stock, specifications and other details on each company’s official website.
Independent Research
This site is independently planned and operated by Stellagent, Inc. Unless otherwise stated, the featured retailers and service providers did not commission the research, provide the data or supervise the content. Our priority is not to promote a particular company or judge businesses by rank alone, but to accumulate observed facts using a consistent standard.
Trademarks and Logos
Company names and logos help readers accurately identify which retailers appeared in an AI answer. We therefore use the company names, service names, trademarks and logos shown on this site only to the extent needed to identify observed retailers and explain the research results.
All rights in the respective trademarks and logos belong to their owners. Their appearance on this site does not indicate an affiliation with, endorsement by or approval from those rights holders. If you notice an issue with a name or published item, we will review it and respond in good faith.
Updates
Data is added after each observation and accumulated as a daily trend. Last updated is the time of the latest aggregation.
Contact
For questions about the research method, corrections to published names or rights-related inquiries, please contact us through the operator’s contact form.
Stellagent, Inc. develops agentic commerce infrastructure for shopping and booking experiences in the AI agent era.
- Company name
- Stellagent, Inc.
- Representative
- Akihiro Suzuki, CEO
- Address
- WeWork, Ocean Gate Minatomirai 8F, 3-7-1 Minatomirai, Nishi-ku, Yokohama, Kanagawa, Japan
- Founded
- February 22, 2022
Contact
Contact us using the form below with questions about this research.
Stella AI検索対策
Stella AI Search measures how generative AI reads and recommends your products and services, then turns those findings into improvements. We can apply the same observation method used here to your products, competitors and target customers.