All research
Soft-focus macro photograph abstract color composition — pale golden cream fading into a deep, saturated amber and burnt orange

World FAQ Insights

NewsJune 16, 20261 min read

World FAQ insights, built from real guest questions

Instead of guessing what guests might ask, HOWI now looks at what they actually asked — and turns the pattern into a canonical FAQ a host can review at a glance.

AirNLP Research

A host filling out a world for the first time faces a blank-page problem: what will guests actually want to know? The honest answer was always sitting in the conversation history — it just wasn't organized.

What shipped

World FAQ insights derive a canonicalized, deduplicated list of real guest questions for a world, cached and ready for a host to review:

  • Grounded in real conversations. Questions are derived from actual guest messages, not guessed at or generated from nothing.
  • Canonicalized, not raw. Only aggregated, cleaned-up questions are stored — never a raw transcript.
  • Cached per world, with a source-message count and a computed-at timestamp, so refreshes are throttled instead of recomputing on every page load.

Why it matters

The gap between "what a host wrote down" and "what guests actually ask" is exactly where knowledge goes stale. Surfacing the real pattern of questions turns that gap into a to-do list — and over time, a world's coverage gets better because it's driven by evidence, not guesswork.