HavenstayHospitality

Havenstay Hospitality operates 12 boutique hotels and inns, each staffed by a small front desk and concierge team that fields guest requests over phone, in-room tablet, and text message throughout the stay. A large share of these requests were routine (checkout extensions, wifi troubleshooting, local dining recommendations) but staff still had to personally respond to every single one, often during check-in rushes when response times slipped noticeably.

Guests increasingly expected a fast response over text message specifically, having grown used to that pace of service elsewhere, but Havenstay's small on-property teams simply couldn't staff for instant response around the clock at every property without a significant cost increase.

Havenstay wanted routine requests handled instantly without losing the personal touch that differentiated its brand, and needed any guest showing frustration or a complex issue (a maintenance problem, a billing dispute) routed to a real staff member immediately rather than stuck talking to a bot.

Aivora AI built a guest assistant that connects to Havenstay's existing SMS and in-room chat channels, trained on each property's specific amenities, local recommendations, and house policies so responses felt property-specific rather than generic. A sentiment and complexity classifier runs on every incoming message, routing anything showing frustration, a service failure, or an unusual request directly to the on-duty staff member's device within seconds.

Routine requests (the majority of volume) get an instant, on-brand response drafted in Havenstay's established guest communication voice, developed closely with Havenstay's guest experience team over four weeks of tone workshops. Staff retained full visibility into every conversation and could step in at any point, with a simple takeover control built into the same interface they already used for guest messaging.

The team piloted the assistant at two properties for six weeks, tuning the escalation classifier based on which conversations staff chose to take over, before rolling out to all 12 properties over the following ten weeks.