Found during discovery, absent from practical advice.
Two models named the client's product during discovery, but subsequent usage and comparison advice did not connect to it. NiubiGEO assessed discovery and practical understanding separately to identify content priorities.
01 / FINDINGS
Findings
Some models understood the product's purpose
Two discovery answers named the client's product and explained its main purpose, establishing a starting point for discovery.
Practical advice did not connect to the product
The other nine answers discussed evaluation methods and comparison conditions without naming the product. NiubiGEO recorded this separately from the discovery result.
Existing documentation was not fully reflected
Public materials documented supported scope and usage conditions, but answers did not fully convey those limits. The diagnosis distinguished omissions in answers from gaps in product capability.
02 / QUESTIONS & OUTCOMES
Question themes and diagnosis outcomes
The following are anonymized intent summaries, not the prompts sent to the models. Results refer to the saved original tests. All four original variants omitted the client's brand, website and product materials; one topic used both a general and a detailed version.
| Question theme / intent | GPT-6 Astra | Claude Fable 5.1 | Gemini 3.8 Flash |
|---|---|---|---|
| Q1 · Product discovery Find developer tools suited to a team's evaluation needs. | Mentioned | Mentioned | Not mentioned |
| Q2-A · Usage review · general Understand how to check the work results a tool presents. | Not mentioned | Not mentioned | Not mentioned |
| Q2-B · Usage review · detailed Explore which process information can support a review of results. | Not mentioned | Not mentioned | Not mentioned |
| Q3 · Solution comparison Identify the evaluation conditions needed to compare tools. | Not mentioned | Not mentioned | Not mentioned |
A mention does not establish a recommendation or an accurate description. Named questions and unprompted discovery are interpreted separately.
03 / RECOMMENDATIONS
Recommended priorities
Add a worked usage example
Use a common task to show inputs, steps, results and limits, linked from the product overview.
Make comparison conditions easier to check
Compare supported scope, requirements and unknowns under the same evaluation scenario so readers can make a practical choice.
Retest discovery and understanding separately
After updating the pages, reuse the saved original questions and conditions to check mentions and the accuracy of purpose and limits.
04 / SCOPE
Test scope
This API round enabled web search and completed all 12 requests, with one sample per question and model. Some responses returned no retrieval record, so actual retrieval is not established for those responses. Brand absence in a methods answer is not a failure. No human interface or software functionality testing was performed. Recommendations have not been implemented or retested and do not establish growth.
NiubiGEO completed this report using model API tests. Human web and app testing can be arranged separately, with human testing supplied by NiubiStar.