Five inboxes. One router. Every enquiry self-qualifies.

The traffic was already there.

A boutique villa resort in Malaysia. A multi-villa property serving leisure stays and private events. Five active inbound channels — WhatsApp Business, Instagram Direct, Facebook Messenger, TikTok Business, and a website chat widget. Each channel was its own inbox with its own thread of conversations.

The enquiries were real and they were varied. Couples asking about romantic stays. Families needing villa recommendations. Group bookings, corporate retreats, weddings, private celebrations — every guest type with its own questions. But every one of those conversations had the same five-question core. Someone had to read it, sort it, ask the missing details, write the answer, and pass the lead to the right person.

The job was not to bring in more enquiries. The enquiries were already arriving. The job was to make sure the team could handle them without becoming the ceiling.

What we built.

A twelve-workflow respond.io system, built on a single rule — every enquiry walks itself to a structured lead before a human reads the chat.

It starts with one front door. Every channel — WhatsApp, Instagram, Messenger, TikTok, webchat — funnels into the same router. The guest sees a short welcome, picks a path, and the system branches into one of six segment flows: couple, family, group, corporate event, wedding, or general enquiry.

Each path recommends the right villa type for that guest, asks one short question about the occasion, and ends with four free-text questions to capture the booking details — dates, pax count, villa preference, contact information. By the time the conversation reaches a person, eight custom fields are already filled in on the contact record. Thirteen tags are applied along the way for sorting and reporting later. The team opens the chat already knowing who the guest is, what they want, and when.

When the question is open-ended — what’s the difference between a Lake View and a Pool villa?, do you cater to dietary needs?, can I bring kids? — an AI concierge takes over. It uses the actual rate sheet and villa info as its source of truth, replies in two or three short sentences, and only passes to a human when the question is really outside its lane. When it does pass, round-robin assignment picks the team member with the lightest current load. Nobody watches five inboxes anymore.

Where the value lands.

The ROI of this build is not a single number. It is the difference between two ways of working.

Before, every extra enquiry cost roughly the same human time as the last one — read, sort, ask for missing info, reply, repeat across whichever channel the guest chose. The more enquiries, the more hours. The only ways to scale were to hire more people, reply slower, or stop answering FAQ-style questions altogether.

After, each enquiry walks most of the way on its own. The four-question funnel collects what reservations actually needs. The AI tier handles the long tail of questions that don’t need human judgement. People only spend time on conversations that turn into stays. The cost per enquiry stops growing with volume — and the more traffic the property gets, the more the system pays for itself.

Three places that show up directly:

Time spent sorting enquiries. What used to be a few minutes of hand-sorting per message is now zero. The system has already classified the guest and filled in the booking fields the team needs.

Coverage outside business hours. First reply is 24/7 across every channel. A guest who messages at midnight gets a useful reply at midnight, not the next morning. That window is where bookings are won or lost.

Room to grow. New channels plug into the same router without rebuilding the booking flow. New guest types are a workflow addition, not a redesign. More enquiries don’t mean more staff.

What it took.

About three weeks from kickoff to fully live. Twelve workflows, five channels, eight custom fields, thirteen tags, an AI concierge that only answers from the property’s actual rate sheet and villa list, and a human handoff that does not re-trigger once a guest is already assigned. Round-robin uses respond.io’s least-contacts logic, so adding or removing team members in the UI rebalances the load — no workflow changes needed.

The retainer covers what comes next. Workflows get tuned as the team learns which paths convert and which get abandoned. The AI concierge gets re-pointed at the latest rate sheet and villa info as those change. The system gets cheaper to run every month, not more expensive.

If your business runs on bookings.

This is a hospitality story, but the shape works almost anywhere. If your business is appointment-driven or enquiry-driven and lives on more than one channel — dental and aesthetic clinics, spas, tour operators, event venues, private dining, professional services — the same pattern fits. One router. Branching paths by customer type. Four questions to capture the booking. AI for the long tail. People only when people actually add value.

The math is the same wherever you sit. Every enquiry the system handles is one your team does not have to. The bigger the volume, the bigger the difference.