KNOWLEDGE BASE · GUIDE
Your LLM Visibility Baseline: 8 Questions Before You Fix Anything
Your LLM visibility baseline is eight buyer questions asked across four AI engines, scored on whether your firm appears, and captured as screenshots on day zero. Before you fix anything, you need to know where you stand. Most small and mid-size sites are effectively invisible to AI crawlers: no llms.txt, thin entity markup, and crawler rules written for search, not for answers (skillpilotadvisory.ai/geo, live 15 Oct 2026).
The baseline is the "before" in your before-and-after. Without it, you cannot prove that Scale's 40-point audit and fix pack moved the score. The day-30 recheck compares against this baseline. If you skip it, you are fixing blind.
The eight questions to ask four engines
| Question | What it tests | What a good answer looks like |
|---|---|---|
| Who is the best [your category] in [your city]? | Whether the engine names you in the shortlist | Your firm name appears in the top three answers |
| What does [your firm] do? | Whether the engine can describe your business accurately | A correct one-sentence description, not a guess |
| Where is [your firm] located? | Whether the engine knows your geography | Correct city and country |
| How much does [your product] cost? | Whether the engine has your pricing | Correct price or a correct "check the site" |
| Who founded [your firm]? | Whether the engine knows your founder | Correct name |
| What are the alternatives to [your firm]? | Whether competitors crowd you out | Your name appears alongside competitors, not replaced by them |
| Is [your firm] legit? | Whether the engine has trust signals | A positive or neutral answer, not "I cannot find information" |
| What do customers say about [your firm]? | Whether the engine has corroboration | Real reviews or mentions, not silence |
Ask all eight on ChatGPT, Gemini, Perplexity, and Claude. Screenshot every answer. Date the screenshots. That is your baseline.
How Scale uses the baseline
Scale's 40-point GEO audit scores your site on seven points: crawler access, plain HTML, answer-first structure, schema, llms.txt, entity clarity, and off-site corroboration (skillpilotadvisory.ai/geo, verified 15 Oct 2026). The audit does not measure whether engines cite you. It measures whether they can read, verify, and trust your site well enough to cite it.
The fix pack includes your llms.txt written for your site, schema snippets, and content restructure patterns. The 30-day sprint plan tells you what to fix first. The day-30 recheck scores the fixed items again. Founder-reviewed within 48 hours of intake (skillpilotadvisory.ai/geo, fetched 15 Oct 2026).
What this does not cover
The baseline does not prove the engines will cite you after the fixes. Scale removes the reasons engines skip you. It cannot force a citation. Any firm that guarantees citations is lying (skillpilotadvisory.ai/geo, FAQ, live 15 Oct 2026). The baseline tells you where you start. The audit tells you what to fix. The recheck tells you whether the score moved.
Sources and context
- Scale: 40-point GEO audit, ranked fix list, copy-paste fix pack, 30-day sprint, day-30 recheck. US$199 or RM850, one-time (skillpilotadvisory.ai/geo, verified 15 Oct 2026).
- Most small and mid-size sites are effectively invisible to AI crawlers: no llms.txt, thin entity markup, and crawler rules written for search, not for answers (skillpilotadvisory.ai/geo, live 15 Oct 2026).
- The audit covers GPTBot, ClaudeBot, PerplexityBot, and Google-Extended crawler access, plus content structure, schema, and entity clarity (skillpilotadvisory.ai/geo, FAQ, verified 15 Oct 2026).
- Generative-engine optimization spend is estimated at about US$1.2 billion in 2026, up from about US$848 million in 2025 (skillpilotadvisory.ai/geo, live 15 Oct 2026, modeled industry estimates).
FAQ
How long does the baseline take? About 30 minutes. Eight questions across four engines, screenshot each, date them.
Do I need to repeat the baseline after fixes? Yes, on day 30. Scale's recheck covers the fixed items. Ask the same eight questions on the same four engines and compare.
What if the engines already name me? Then your baseline is strong. Focus on Scale's content structure and schema checks to make the citation more accurate, not just more frequent.
Can I run the baseline without buying Scale? Yes. The eight questions are free to ask. Scale tells you what to fix based on what the audit finds.