Our method / YuSeen
How do we assess your products on AI platforms?
YuSeen designs questions around shopper needs in your target market, runs repeated tests on relevant AI platforms, and checks the responses against product information.
We define the test conditions, retain evidence that can be reviewed, and make results comparable where possible.
You receive findings, specific issues to address, and a clear order of priority.
Our assessment principles
Three principles01Findings need supporting evidence
Each issue in a report should be linked to a specific AI response, test record, or source of product information.
Where possible, we retain original responses, screenshots, and source references so clients can review the findings independently.
02One response does not establish overall performance
Generative AI responses vary. A product appearing—or being absent—in one answer does not establish consistent performance on that platform.
We repeat tests under agreed conditions to observe product mentions and descriptions across the sample.
03Improvements must reflect the facts
Specifications, certifications, reviews, and purchase terms must be supported by real information.
We do not invent product claims, create fake reviews, or use content intended to mislead AI systems.
What does a full assessment involve?
Assessment processFrom designing shopper questions to checking product information and retesting.
01Design the test questions
We establish the product category, sales market, key competing brands, and typical purchase needs.
We then build a set of questions suited to that market.
| Test type | Example question | What we examine |
|---|---|---|
| Product recommendations | What's the best carry-on backpack for a 14-inch laptop? | Whether the product appears among relevant options |
| Suitability for a use case | Which backpack can fit under an airline seat as a personal item? | Whether the available information supports that use |
| Competitor comparisons | Brand A vs. Brand B: which is better for a weekend trip? | The comparison criteria and differences AI describes |
| Brand and service information | Is Brand A reliable, and what is its return policy? | Accuracy of brand information and purchase terms |
Once a baseline question set is established, we keep questions, language, and other controllable conditions as consistent as practical during retesting to improve comparability.
02Test on relevant AI platforms
We select applicable AI search or shopping assistants according to the target market and agreed service scope.
Testing may include ChatGPT, Google AI Mode, Microsoft Copilot, Perplexity, and Amazon's AI shopping assistant (Alexa for Shopping in the US, formerly Rufus).
We examine whether products are mentioned, how they are described, which brands they are compared with, and what verifiable sources appear in the response.
Platform availability and the specific scope are confirmed before the project begins.
03Check the product information
We compare test responses with public product information, merchant policies, and relevant data fields, focusing on five areas.
01Product identity and consistency
Whether the brand, model, product name, and descriptions are clear and consistent across pages.
02Specification completeness
Whether dimensions, weight, materials, features, and intended uses are adequately described using units familiar to the target market.
03Purchase and after-sales information
Whether price, stock, shipping costs, delivery times, return policies, and warranty terms are clear and consistent.
04Verifiable brand information
Whether the brand background, applicable certifications, genuine reviews, and merchant information needed to verify a transaction are accurate and easy to check.
05Information accessibility
Whether access restrictions or technical settings prevent external services from retrieving public product information and relevant data.
04Retest after updates
After product updates are published, we generally recommend allowing a 2–4 week observation period before retesting under agreed conditions.
We look for changes in product mentions, information accuracy, and competitive performance.
Website updates may not appear immediately in AI responses, and update cycles can differ between platforms. The observation period is a planning recommendation, not a commitment about a platform's refresh schedule.
What will you receive?
Deliverables01Product mention records
Records of mentions across the test question set, with comparisons to key competitors.
02Information accuracy checks
Confirmed errors, omissions, or unsupported statements in AI descriptions, with supporting evidence.
03Competitive analysis
Observable product and information differences, the language used in recommendations, and factors that require further investigation.
04Prioritized improvements
Specific recommendations ordered by significance and practicality, so your team can assign the work.
How should you interpret the results?
AI models, data sources, and response policies can change over time.
An assessment therefore records observations at a particular time, on particular platforms, under defined conditions. It does not predict every future answer.
YuSeen focuses on longer-term patterns and evidence that can be reviewed. We do not treat a single result as a ranking commitment or guarantee that an improvement will deliver a specific increase in sales.
See how your products currently appear on AI platforms
Send one target market and links to 3–5 priority products to request a free initial assessment.