Skip to main content

​
Direct answer

Qwairy can show where a selected ChatGPT model recommends a DIRECT competitor and omits your brand. It cannot prove why the model made that choice. Diagnose the observed prompt, answer wording, eligible metrics, and exposed sources before forming a hypothesis.

​
Data required

  • Completed answers from the selected ChatGPT model
  • The exact prompts and active country, language, topic, tag, funnel, and period filters
  • Correct SELF and DIRECT competitor classifications
  • Mention occurrences, recorded positions, sentiment, and exposed citations
  • Comparable answers where your brand is included

​
Workflow

  1. In Cockpit > Compare, filter to the selected ChatGPT model and a fixed scope.
  2. Identify prompts where a DIRECT competitor appears and your brand does not.
  3. Open those answers in Monitor > Response Analysis and read the full recommendation criteria and caveats.
  4. In Monitor > Source Explorer, inspect sources exposed with the same answers.
  5. Compare with prompts where both brands appear to find differences in intent, evidence, market, or answer format.
  6. Check whether the pattern persists across repeated completed answers before acting.

​
Interpretation

A competitor mention or earlier recorded position is an observed output, not evidence about private model reasoning. Product Share of Voice counts SELF and DIRECT mention occurrences. Product Mention Rate is response-level and uses answers with at least one SELF or DIRECT mention as its denominator. Keep recommendation context separate from raw presence. A competitor can be mentioned as an unsuitable option, and a cited competitor domain may support a general fact.

​
Possible next actions

  • Test a page that answers the recurring selection criterion with verifiable evidence.
  • Correct inaccurate or outdated public information found in the observed answer or cited source.
  • Investigate whether the pattern is limited to one prompt, model, topic, or market.
  • Track the same prompt cohort after a content or positioning change.
These are hypotheses. No content change guarantees future inclusion or recommendation.

​
Limitations

  • Qwairy does not expose ChatGPT’s private training data, weights, or decision process.
  • Answers and exposed citations can vary between runs and models.
  • Recommendation detection requires context; mention alone is insufficient.
  • Monitoring results do not represent all ChatGPT interactions.