Signal & Technical¶
Research domain: RAG mechanics, llms.txt, schema markup, crawl signals, and the technical foundations that determine whether content gets retrieved by AI systems.
When the Orchestrator Routes Here¶
- Technical foundations are unknown or suspected broken
- Content isn't appearing in AI answers and the reason is unclear
- Before content strategy work — a site with blocked crawlers or zero schema is not ready for content gap analysis
What This Specialist Researches¶
| Area | What it covers |
|---|---|
| Crawl permissions | robots.txt — which AI crawlers are allowed |
| Bing indexation | ChatGPT's retrieval pipeline is Bing-powered |
| llms.txt | Perplexity citation prioritization signal |
| Schema markup | Structured data that makes facts machine-readable |
| Chunking quality | How well content breaks into retrievable segments |
| Factual density | Specific, verifiable claims per page |
How It Verifies Claims¶
- Schema claims: validated at
validator.schema.org, never assumed from page inspection - Crawler access:
robots.txtread directly, never assumed from CMS defaults - Bing indexation: live
site:search, never inferred from Google status - llms.txt: checked via direct URL, never assumed absent without checking
Uncertainty flags are mandatory when a finding is inferred rather than directly observed.
Session Output Format¶
Findings are presented as a table with columns: Signal · Status · Priority. Followed by a root cause statement, recommended sequence, and explicit statement of what cannot be determined without additional access.
See 1-signal-technical/examples.md for a complete session example.