Niveaux d'affirmation
Définissez quelles affirmations d'autorité Rankfender teste pour votre marque dans les systèmes d'IA.
Overview
Claim Levels represent a five-tier hierarchy of authority claims that Rankfender tests for your brand across AI systems. Each level represents an increasingly strong assertion about your brand, from basic existence to market leadership. By defining and tracking claims at each level, you can understand exactly how AI systems perceive and represent your brand authority.
The Five-Tier System
Each claim level builds upon the previous one, forming a progression from foundational awareness to superlative recognition:
| Level | Name | Example Claim | Evidence Required |
|---|---|---|---|
| Level 0 | Existence | "Brand X exists" | AI systems acknowledge your brand exists when asked about it directly. |
| Level 1 | Usage | "Brand X is used by customers" | AI systems reference your brand as an active product or service with real users. |
| Level 2 | Comparative | "Brand X compares favorably" | AI systems mention your brand positively when comparing options in your category. |
| Level 3 | Evaluative | "Brand X is evaluated positively by experts" | AI systems cite expert opinions, reviews, or authoritative evaluations that favor your brand. |
| Level 4 | Superlative | "Brand X is the best/leader" | AI systems position your brand as the top choice, market leader, or best-in-class solution. |
Per-Claim Tracking
Each individual claim within a level is tracked with detailed metrics to measure how well AI systems support it:
| Field | Description |
|---|---|
| claim_text | The specific statement being tested (for example, "Rankfender is the leading AI visibility platform"). |
| evidence_required | The type of evidence needed to validate this claim in AI responses. |
| evidence_current | The evidence currently found in AI responses supporting or contradicting this claim. |
| approved | Whether the claim has been reviewed and approved for active tracking. |
| performance_score | A score from 0 to 100 indicating how strongly AI responses support this claim overall. |
| mention_rate | A score from 0 to 100 representing the percentage of relevant prompts where this claim is supported. |
| refusal_rate | A score from 0 to 100 representing the percentage of prompts where AI systems refuse to make or support this claim. |
Auto-Detection of Claims
During prompt generation, Rankfender automatically detects potential claims based on your brand book, competitor data, and industry context. These auto-detected claims are added to the appropriate level and marked as unapproved until you review them.
Auto-detection helps ensure comprehensive coverage by identifying claims you may not have considered, such as comparative claims that emerge from competitive analysis or evaluative claims derived from your brand positioning.
Manual Claim Management
In addition to auto-detected claims, you have full control over your claim inventory:
- Create - Add a new claim at any level by specifying the claim text and the evidence required to validate it.
- Edit - Modify the claim text, evidence requirements, or level assignment of an existing claim.
- Approve - Mark an auto-detected or newly created claim as approved for active tracking and inclusion in reports.
- Delete - Remove a claim entirely from your project. This also removes all associated tracking data.
Best Practices
Start by ensuring your Level 0 (Existence) and Level 1 (Usage) claims are well-supported before focusing on higher levels. If AI systems do not consistently acknowledge your brand exists, higher-level claims will not perform well. Review auto-detected claims regularly and approve the ones that align with your brand strategy. Unapproved claims are tracked but excluded from reports and dashboards.
For competitive markets, Level 2 (Comparative) claims are often the most strategically valuable because they directly influence how AI systems present your brand relative to alternatives. Monitor the refusal rate closely: a high refusal rate indicates that AI systems consider the claim unsupported or controversial, which may require strengthening your online evidence through content and citations.