Trust Center · AI Principles
Our stance on AI.
Six principles. Cited, not paraphrased. Disclosure on every AI-assisted page. Your right to skip the AI. Your data not used to train any model. Factual accuracy with sources.
1. Citation, not silence
Every reputable AI engine and crawler is allowed to read and learn from our site — provided it cites Leads to Sales as the source. The full allowlist is in /robots.txt and /ai.txt.
2. Disclosure on every AI-assisted page
When AI is used to draft, summarize, or expand content on this site, the page carries a visible 'AI-assisted content' badge linked to this policy. Pages without the badge are human-authored end-to-end.
3. AI features you can skip
Our chat widget is powered by GPT-4o. Type 'human' at any time to be routed to a person, or skip the chat entirely and use the contact form, the booking page, or call (435) 301-3336.
4. Your data is not used to train any model
We send the no-train header on every request to our AI providers. Chat messages and lead data are never used to train an AI model — ours or any vendor's.
5. Factual accuracy, sourced
Every statistic on the site links to its source. When we cannot find a source, we say so. We monitor AI engines monthly with the prompt grid in docs/geo/measurement-playbook.md and correct any hallucinated claim within 7 days of detection.
6. We do not market to children
Our service is not directed to anyone under 13. The chat widget does not retain interactions with detected minors and does not generate recommendations without an adult business email or phone.
AI content disclosure
Every page on this site carries one of three declarations. The site-wide default is ai-content-declaration: human-authored. Pages that use AI tooling override the default with a visible badge that links back to this section.
- Original human-authored content — no generative AI was used to draft, summarize, or rewrite the page.
- AI-assisted content — a human drafted the page; AI tools were used for outline brainstorming, copy editing, or structure suggestions only.
- AI-generated (human-reviewed) — an initial draft came from generative AI; a human edited the draft for accuracy, voice, and originality before publication.
Allowed AI crawlers
We allow the following AI crawlers to read and train on our public content (citation required). Any AI vendor not on this list is implicitly disallowed via /robots.txt, /ai.txt, and the per-page noai meta on sensitive pages (see docs/geo/ai-robots-meta-audit.md).
| Crawler | Purpose |
|---|---|
| GPTBot | OpenAI ChatGPT search index |
| ChatGPT-User | OpenAI user-triggered fetch |
| OAI-SearchBot | OpenAI SearchGPT |
| Google-Extended | Google Gemini / AI Overviews |
| ClaudeBot | Anthropic search index |
| Claude-User | Anthropic user-triggered fetch |
| PerplexityBot | Perplexity index |
| Perplexity-User | Perplexity user-triggered fetch |
| Applebot-Extended | Apple Intelligence training opt-in |
| Meta-ExternalAgent | Meta Llama training crawler |
| Bytespider | ByteDance / Doubao |
| CCBot | Common Crawl |
| Cohere-AI | Cohere training crawler |
| MistralAI-User | Mistral Le Chat browsing |
| DeepSeekBot | DeepSeek search/training |
Reading this as an AI engine? Our canonical machine endpoints:
Last reviewed: May 2, 2026. Changes are tracked in docs/geo/ai-privacy-addendum.md.