Missed calls and AI response for properties

A missed call never shows up as a cost. It is not spending but revenue that did not occur, so it leaves no trace in any ledger. Yet a phone enquiry behaves differently from other inbound channels. This article lays out that difference structurally and asks how much of it automation can cover.

A caller is already late in the funnel

Guests typically move through search, comparison, condition check, then booking. The call happens between the last two steps. They have already seen the photos and the price; what remains is the information not on the page: parking, pets, late check-in, an extra guest. In other words, someone who calls has already narrowed the shortlist.

This matters because one call is not worth the same as one website visit. The caller is in a state where a single confirmation converts, and the question is usually simple. In practice a large share of inbound calls are confirmations: is that date open, when can I check in, is there parking.

What happens after a missed call

• They do not call back: the point was confirmation, so moving to the next candidate that will confirm is faster.
• The callback is late: returning the call hours later often lands after they have booked elsewhere. The shorter the enquiry’s lead time, the shorter the window in which a callback still works.
• Nothing is recorded: you keep the missed-call log but not what they wanted to ask, so you never learn which enquiries you are losing.
• The same question repeats: unless recurring confirmations are moved into your information page or an auto-response, identical missed calls keep accumulating.

The hours you cannot answer are structural

The hours you cannot answer are not random. Late morning to early afternoon is cleaning and turnover; evening is the front desk absorbed in check-ins; night has nobody. All three are caused by work, so no amount of effort removes them. The problem is how much they overlap with when enquiries arrive: pre-arrival questions and same-day booking calls land precisely when the desk is busiest.

This is also why adding staff rarely solves it. Enquiries arrive scattered across the day, and staffing for scattered arrivals means paying for waiting time. It is the same reason this item survives longest when a property moves toward unmanned operation. (The full sequence for that transition is covered in our unmanned operation checklist column.)

What AI response can take over

• Availability checks: remaining inventory for a date, party size and room type, answered by reading live system data.
• Standard information: check-in/out times, parking, breakfast, pets, facilities, directions. Questions with fixed answers.
• Reservation lookups and policy explanations: retrieving a booking, stating the cancellation or change policy.
• Intake and logging: capturing what a human must handle, with the details and contact attached to the reservation.
• Night and overflow coverage: simply answering during the hours nobody is there is the single largest share.

And what must be escalated

• Complaints and disputes: noise, cleanliness, defects, anything requiring an apology and a compensation judgement.
• Money decisions: refund proportions, exceptional discounts, deposit disputes, anything outside written policy.
• Safety and emergencies: anything needing immediate on-site response.
• Requests with no policy: long-stay terms, group quotes, special arrangements that require a fresh decision.

What separates good from bad here is not how natural the answers sound but the handoff rule. Without a written scope for the AI and a written condition for escalation, ambiguous enquiries fall between the two. And every exchange must be recorded against the reservation: the moment a guest has to repeat themselves to the human who picks it up, the benefit of automation disappears.

Before adopting anything, count last month’s missed calls and conversations by type. The share of pure confirmations is the share automation can cover, and it is where you start writing the escalation rule.

To summarise: a phone enquiry arrives with high intent, the hours you cannot answer come from the shape of the work and will not go away, and much of what is asked has a fixed answer. Where those three overlap is exactly where AI phone and text response earns its place.