CLAWSIGNAL RESEARCH / OWNED ACCOUNT

ClawSignal AI visibility baseline

A dated measurement with the zero result left intact. This is ClawSignal's own account, not a customer case study and not a claim of future lift.

Observed July 19, 2026 · Published July 22, 2026

40recorded checks
4AI assistants
0brand appearances
0citations

What the baseline found

ClawSignal ran 10 fixed service prompts across ChatGPT, Claude, Gemini, Grok. The production records contain 40 checks, with 0 ClawSignal appearances and 0 citations. A tracking system is useful only if it can report an unfavorable result without smoothing it into a success metric.

AssistantChecksAppearancesCitations
ChatGPT1000
Claude1000
Gemini1000
Grok1000

Method and provenance

  1. The monitored entity was ClawSignal's owned production project.
  2. The prompt set contained 10 service-intent queries.
  3. Each prompt was checked on ChatGPT, Claude, Gemini, and Grok.
  4. Appearance and citation fields were read from production scan records.
  5. No customer data, estimated lift, or synthetic success result was added.

Source tables: ai_scan_runs, ai_visibility_checks, and ai_share_of_voice. The July 19 aggregate row is stored as partial even though its adapter counters record four attempted, four succeeded, and zero failed. We preserve that discrepancy as a data-quality limitation instead of silently relabeling the run.

What this proves, and what it does not

Supported

  • The four-assistant scan path produced stored observations.
  • The system retained a zero-visibility result.
  • The baseline is specific enough to repeat after changes.

Not yet supported

  • No causal lift can be inferred from a baseline.
  • No customer outcome is represented here.
  • No claim is made that one scan predicts future recommendations.

Start with a result you can repeat

Run your own scan, keep the first result intact, and measure changes against it.