Methodology

How the SpotAQ score works

SpotAQ turns how AI models talk about your brand into a single score between 0 and 100. This page explains what we measure, how we calculate the score, and where its limitations are.

1. Inputs: what we measure

For each check, we run a controlled set of prompts against AI engines such as ChatGPT, Perplexity, and Google AI. For every response, we extract four main signals:

  • Mention rate – how often your brand is mentioned when AI answers category-level questions.
  • Sentiment – whether the answer describes you in positive, neutral, or negative terms.
  • Source quality – which domains are cited as sources: your official site vs third-party sites.
  • Google grounding – whether Google AI grounding finds your site and at which position.

2. Scoring model: how we get to 0–100

Internally, each test is scored using a simple model:

  • If your brand is not mentioned at all, that test receives a low baseline score.
  • Positive sentiment and official-site citations boost the score.
  • Negative sentiment and "we can't find information about this brand" patterns reduce the score.
  • Strong Google grounding (appearing as a trusted source in grounded answers) provides an additional bonus.

The overall SpotAQ score is the average of these per-test scores across all prompts in your check.

3. Why scores can be "low" even when you rank in Google

Traditional SEO looks at how you rank for keywords in search results. SpotAQ looks at how AI answers questions. It's possible to rank well in Google but still be absent from AI answers because:

  • You have few third-party reviews or directory listings.
  • Your positioning doesn't match the language buyers use in prompts.
  • AI has more training data and citations for larger or older competitors.

4. Data sources and privacy

We currently use a mix of public APIs and browser-based integrations to query AI engines. We do not log your prompts or brand data beyond what's required to generate your report.

  • We do not sell or resell your individual reports.
  • Aggregated, anonymized data may be used to improve the scoring model.

5. Limitations

No AI visibility metric can be perfect. Important limitations of the SpotAQ score include:

  • AI models are non-deterministic; repeated runs may vary slightly over time.
  • We currently test a curated set of prompts, not every possible way a buyer might ask.
  • We do not have access to proprietary training data; we only see model outputs and cited sources.

6. How to use the score

The SpotAQ score is best used as a directional, comparative signal:

  • Tracking your own progress over time.
  • Comparing how often competitors are recommended relative to you.
  • Prioritizing work on content, Schema.org markup, and directory listings.

It should not be treated as a compliance metric or an absolute statement of "AI truth". Instead, think of it as a structured prompt to ask better questions about how AI sees your brand.