Customizable AI Auto Reply Brand Voice / Personality

Description: Hosts currently have no way to configure the tone, personality, writing style, or brand voice used by AI Auto Reply. Tone is learned implicitly from past host replies and Knowledge Hub phrasing, which: Is slow to converge (needs lots of manual replies in the target voice). Can't be defined once and reused across a portfolio. Produces noticeably different tone between Auto Reply (more direct / enforcement-flavored, due to the safety + judgement layer) and Suggest/Improve with AI (warmer, because it inherits the host's own wording). Requested capability: A dedicated Brand Voice / Personality setting for AI Auto Reply, configurable at: Account / global level (default voice for the whole portfolio) Property level (override per listing — e.g. luxury vs. budget vs. family-friendly) Optionally per Auto Reply rule (e.g. different voice for upsells vs. restrictions) Example use case (verbatim from a host): "Our Auto Reply correctly answered a parking question, but the response emphasized 'street parking is not permitted' and several additional restrictions. We would prefer something closer to: 'Absolutely! You'll have dedicated parking on the right side of the driveway, with space for up to two vehicles parked tandem. We ask guests to use the designated driveway parking rather than the street.' Ideally, we would like to define this communication style once rather than embedding tone instructions into every Knowledge Hub article across our portfolio." Desired voice attributes (host example): warm, welcoming, concise, conversational, hospitality-focused, leads with the solution / what the guest can do, then gently explains any restriction. Workarounds hosts use today: Rewriting Knowledge Hub entries in the target voice. Using Escalation Policy to send verbatim host-written messages on tone-critical topics. Keeping sensitive topics on Suggest with AI instead of Auto Reply. Source: Intercom conversation 215476305926529 (Ariel, Oct 2026)