GEO methodology and platform scope
How GEO Lens creates questions, collects AI answers, computes metrics, and frames result limits.
Use this page for questions like “what does GEO Lens measure?”, “what are TEEA stages?”, and “why do results change?”.
What GEO Lens Measures
GEO Lens measures how a brand appears in AI answers, AI search, and AI recommendation scenarios. It is not a traditional SEO rank tracker or an ads platform.
The diagnosis checks whether:
- AI answers mention the target brand.
- The brand appears in recommendation, comparison, or action-oriented answers.
- The brand appears near the top of the answer.
- Competitors are mentioned more often.
- Answers cite official pages, third-party content, cases, reviews, or other sources.
- Specific user-journey stages need more content support.
Results reflect the selected questions, platforms, rounds, collection time, and available sources for that run.
Diagnosis Flow
Brand profile
-> Confirm question set
-> Choose platforms, rounds, and mode
-> Reserve estimated credits
-> Collect AI answers
-> Extract brand, competitor, rank, and source signals
-> Compute GEO metrics
-> Generate the report
-> Share, export, or retestBetter brand profiles produce better questions and more stable brand recognition.
TEEA Stages
| Stage | User state | Good questions |
|---|---|---|
| Trigger | The user has a need but no brand in mind | Pain points, risks, preparation, selection start |
| Exploration | The user is looking for recommendations | Brands, lists, options, how to choose |
| Evaluation | The user compares the target brand with competitors | Target brand reviews and comparisons |
| Action | The user wants pricing, demo, purchase, or support | Contact, purchase path, fit, pricing, service |
Common question counts:
| Type | Questions | Distribution |
|---|---|---|
| Quick diagnosis | About 25 | 5 / 10 / 5 / 5 |
| Presales report | About 45 | 5 / 30 / 5 / 5 |
| Standard diagnosis | About 45 | 5 / 30 / 5 / 5 |
| Deep diagnosis | About 60 | 10 / 30 / 10 / 10 |
Always review the questions before running a diagnosis.
Supported Platforms
Candidate platforms may include Doubao, DeepSeek, Tencent Yuanbao, Qwen, Baidu AI+, Kimi, Quark, Weibo Zhisou, mobile variants, and Antafu.
The actual available list depends on the diagnosis page and backend configuration. Do not treat one platform as the whole market.
Modes
| Option | Meaning | Recommendation |
|---|---|---|
| Search | Search-enabled answer collection | Default for most runs |
| Reasoning Search | Heavier reasoning or deep-answer mode | Use for important questions |
| Rounds | Repeat the same question on the same platform | More stable, more credits |
| Screenshots | Save collection evidence | Useful for troubleshooting when enabled |
Limits
AI answers may change because models, retrieval behavior, content sources, safety filters, timing, devices, and configuration change.
Retests are most comparable when they reuse the same questions, platforms, rounds, and mode.
Assistant Help
You can ask the GEO Lens assistant to explain methodology, check question quality, compare platform choices, or explain why a retest changed. Logged-in actions still require confirmation and follow membership, concurrency, and wallet limits.