RS Gonzales
Names 5 answer engines: ChatGPT, Gemini, Claude, Perplexity, Google AI.
Read 12 September 2026 · rsgonzales.com/hvac-ai-search-optimization/
The scores
| Criterion | Score | Weighted of its weight | Contribution |
|---|---|---|---|
| Named engine coverage | 4 / 5 | 20 / 25 | |
| Measurement method shown | 2 / 5 | 8 / 20 | |
| Citation change over time | 1 / 5 | 2 / 10 | |
| Entity and schema competence | 3 / 5 | 18 / 30 | |
| Falsifiability | 0 / 5 | 0 / 15 |
What the page actually says
- Engines
- Five named engines.
- Method
- Claims monitoring of AI appearance and citations with monthly reporting. No method shown.
- Change over time
- Names a client and makes a specific comparative claim about outranking competitors with far more reviews. No method behind it and no time series.
- Entity and schema
- Schema, Google Business Profile signals, and llms.txt.
- Falsifiability
- No stated limit, and the page presents llms.txt as “the file AI reads to recommend you”.
What we would ask them about
2 things on RS Gonzales's own page, worth raising before you sign. Not a judgement about the work.
- AskPresents llms.txt as a mechanism engines use. No engine has confirmed reading it.
- AskMakes a specific performance claim about a named third-party client with no method attached.
The check we would run on this one
RS Gonzales names its engines and stops short of saying how it measures them. Open rsgonzales.com/hvac-ai-search-optimization/ and look for a prompt count, a run cadence, or any reporting artefact you could be shown. If none appears, the gap is disclosure rather than capability, and it closes the moment they write those three things down. Ask for them before signing rather than after.
Disagree with any of this? The corrections log is where it gets settled, in public, with a date.