On the reg AI SEO - how does ChatGPT talk about you behind your back?
Jan 22, 2026
They dig into how large language models discover and cite web content, including when ChatGPT shows up as a referrer. Conversation covers retrieval-augmented generation, why structured web text helps visibility, and practical steps like metadata, ORCID and plain-language summaries. They also discuss risks like prompt injection, differences between small and large LLMs, and durable audience-building through email and newsletters.
AI Snips
Chapters
Books
Transcript
Episode notes
Use Custom Instructions To Save Prompting
- Use system-level prompts or custom instructions in ChatGPT to avoid repeating formatting and tone requests each time.
- Set preferences like 'ask clarifying questions' or 'use Australian English' to streamline outputs.
LLMs Are Becoming Search Engines
- ChatGPT and other LLMs increasingly act as search engines by retrieving and summarising web content rather than relying on Google alone.
- This retrieval-augmented generation means AI visibility depends on what documents the model can find and parse.
ChatGPT Became A Major Referrer
- Inger noticed ChatGPT becoming her blog's second biggest referrer after Google, prompting investigation into AI visibility.
- That discovery motivated her to research how LLMs retrieve and cite web content for academic audiences.
