CrawlQStudio

Methodology · DefinedTerm · CrawlQ Studio

The BRAND Score.

Five dimensions. Zero to 100. A published rubric for AI content, applied in CrawlQ Studio by a model-assisted evaluation against your own documents. Free for academic and journalistic citation.

What the BRAND Score measures

The BRAND Score is a compliance score for AI-generated content — a 0–100 rating on each of five dimensions, plus a weighted aggregate. Published by CrawlQ Studio in 2026 as a reference rubric for the AI content category.

How the score is produced matters for how you use it. CrawlQ Studio scores each output with a model-assisted evaluation against the workspace’s own documents. That makes it a guide for the reviewer, not a deterministic measurement: the same text can score differently on a second run, and a person still makes the publish decision.

The five-dimension structure reflects how experienced editors actually evaluate AI output: is it on brand? (Fidelity) is it well-reasoned? (Reasoning) does it speak to the right reader? (Audience) is it saying anything new? (Novelty) and will it ship? (Deliverability). Each answer is scored independently; the aggregate catches outputs that are strong on four dimensions and disqualifyingly weak on one.

The score is named BRAND because the letters of the dimensions spell it. This is deliberate — it turns the methodology into a memorable artefact editors and buyers can cite.

The five dimensions

  • B

    Fidelity

    How closely does the output match the brand's documented voice rules, vocabulary, tone, and prohibited phrasings?

    Reads from

    • • Voice & tone guide
    • • Vocabulary lists (approved + banned)
    • • Prior published content the brand endorses

    Scoring

    0-100. A score above 75 passes; below 60 returns to workflow.

  • R

    Reasoning depth

    How well-grounded is the claim? Does it cite the brand's own knowledge graph, or is it generic internet synthesis?

    Reads from

    • • Foundation documents (brand strategy, positioning)
    • • Customer research transcripts
    • • Competitive intelligence in the knowledge graph

    Scoring

    0-100. Grounded claims with traceable sources score highest.

  • A

    Audience alignment

    Does the output speak to the intended persona — their vocabulary, their pain points, their decision context?

    Reads from

    • • Persona documents
    • • ICP definitions
    • • Jobs-to-be-done frameworks

    Scoring

    0-100. Persona-locked outputs score above 80; generic outputs rarely exceed 60.

  • N

    Novelty & differentiation

    Is this output saying something the brand's competitors are not already saying? Novelty is a moat.

    Reads from

    • • Competitive intelligence in the knowledge graph
    • • Category-level content analysis
    • • Differentiator definitions

    Scoring

    0-100. Novelty scoring penalises clichés, category fluff, and repeated phrasings.

  • D

    Deliverability

    Is the output publish-ready for the specific channel and format? Channel-fit, length-fit, layout-fit.

    Reads from

    • • Channel style guides (LinkedIn, blog, email, ads)
    • • Format constraints (length, links, hashtags)
    • • Accessibility guidelines

    Scoring

    0-100. Deliverability scoring replaces the old 'editorial review' catch-all with a measurable gate.

How to cite the BRAND Score

The BRAND Score methodology is free for academic, journalistic, and editorial citation. Suggested citation:

Quantamix Solutions. (2026). The BRAND Score: A five-dimension compliance score for AI-generated content. CrawlQ Studio. https://www.crawlq.ai/brand-score-methodology

Implementing the methodology in your own tooling requires no licence. The reference implementation — where scoring is grounded in a per-customer knowledge graph and logged with a full audit trail — is shipped as CrawlQ Studio.

Reference implementation

Score every AI output against your own documents.

EU-hosted. Upload your foundation documents; every generation gets scored across all five dimensions before it ships.

See pricingBook a demo

Related