Knowledge Hub fallback for %answers% short code in event messages with configurable timeout

Problem Statement

  • When a New Inquiry (or request-to-book/reservation) rule containing the %answers% short code fires and no Question Rule template matches, the block silently resolves empty
  • Inbox AI cannot reach this surface: whether by design or by collision with the instant rule send, Auto Reply does not respond to messages already handled by an event messaging rule — in months of running Auto Reply with the Inquiries stage enabled, it has never once done so
  • The result: the event message — the guest's first touch, the one that stops the OTA response clock — is structurally a dead zone for AI answers, and %answers% is the only mechanism that could carry one

Current Workaround

  • Maintain the legacy Question Rule template library in parallel with the Knowledge Hub, purely to get any inline answers — two answer sources for one job
  • Question detection misfires often enough that many hosts have disabled most Questions, so in practice the first message says "we'll answer soon" instead of containing the answer, and the substantive reply waits for a human

Impact

  • The most conversion-critical message in the funnel is the one place the product's AI can never answer
  • Duplicate threads and delayed responses when the ack doesn't address the guest's actual question
  • Hosts who have migrated their knowledge to the Knowledge Hub get no inline-answer benefit from it

Proposed Solution

  • When %answers% fires and no Question Rule template matches, have Inbox AI generate the answer from the Knowledge Hub and insert it inline
  • AI-resolved answers pass through the same guardrails and safety pipeline as Auto Reply
  • Booking-stage relevance comes free: Inbox AI is already stage-aware, which was the legacy Questions system's key scoping feature — nothing is lost in the migration
  • Add a per-rule setting to hold the message briefly for AI resolution — mirroring the existing Auto Reply "Delay sending" control — with configurable timeout behavior: send with the block empty, or fall back to a matched Question Rule template
  • Priority order: Question Rule templates where defined → AI as fallback → empty as last resort

Unlock

  • Single-message first-touch resolution: the same message stops the OTA response clock and answers the guest's question
  • Hosts maintain one answer source (the Knowledge Hub) instead of two
  • Re-platforms %answers% — the product's original inline-answer mechanism — onto the engine that has already replaced Questions everywhere else; the Question Replies help article already names Knowledge Hub-driven answering as a future direction

Please authenticate to join the conversation.

Upvoters
Status

Backlog

Board

💡 Feature requests

Tags

AI

Date

1 day ago

Subscribe to post

Get notified by email when there are changes.