The phone rings at reception during check in. There are four people in the queue, one of them is asking about a late checkout, and the call goes to voicemail.
Nobody logged that. It does not appear in any report. But the person calling wanted to book three nights, and thirty seconds later they booked them on an online travel agent instead, where you will pay a commission on a guest who was trying to give you the money directly.
Missed calls are the only revenue leak in hospitality that leaves no trace in the system that measures revenue. That is precisely why it persists.
0
Rows a missed call generates in your PMS. It is invisible to every report you run.
15%
Typical OTA commission band on a booking the caller was trying to make directly.
24/7
When guests call. Reception is staffed for a subset of it, at best.
The economics are unusually clean
Most automation business cases require assumptions. This one mostly does not, because the direct versus OTA gap is a number you already know.
A booking taken on the phone carries no distribution commission. The same booking made through an online travel agent carries a commission typically in the mid teens as a percentage, and industry figures for the major platforms sit in that band. So every phone booking you fail to answer and that reappears through a channel manager costs you that spread on a guest who had already chosen you.
The uncomfortable part is that the guest did the harder thing first. They looked up your number and rang it. The switch to an OTA was a consequence of nobody picking up, not a preference.
Where the calls go missing
It is rarely the middle of the night, which is the case most people expect. The losses cluster in four windows that any front office manager will recognise.
Check in and check out peaks, when the desk is physically staffed but every person is in front of a guest. Evenings after the desk drops to one person. Weekends, when enquiry volume rises and rota cover falls. And the long tail of calls that arrive while the one available person is already on another call, which no amount of staffing fixes because it is a concurrency problem rather than a headcount one.
That last category is the one worth dwelling on. Two simultaneous calls cannot be answered by one receptionist however good they are. Adding a second receptionist to cover a peak that lasts forty minutes a day is not a defensible cost. This is a shape of problem that suits automation almost perfectly, because the requirement is elastic capacity rather than more hours.
CHECK IN PEAK
Desk is busy
Staffed, but every person is with a guest. Calls ring out in front of a full lobby.
EVENING
Single cover
One person covering desk, phone, and floor. The second concurrent call has nowhere to go.
WEEKEND
Volume up, cover down
Enquiry calls peak exactly when the rota is thinnest. Most direct booking attempts happen here.
What the caller wants is narrower than you think
Hospitality teams often resist automation here because they imagine it needs to handle everything a receptionist handles. It does not.
The calls that carry revenue are a small, repetitive set: availability for a set of dates, rate for a room type, whether the booking can be made now, parking, late checkout, whether the restaurant is open, and directions. These are answerable from your property management system and a page of policies. They are also the calls most likely to be abandoned, because the caller assumes they are simple and gets impatient.
The genuinely hard calls, a complaint, a group enquiry, a wedding, an accessibility need that requires judgement, are a smaller share and they should go to a person. The design goal is not replacing the front desk. It is making sure the simple revenue calls are never the ones that ring out while your team handles the complicated ones.
Nobody chose the OTA. They chose you, rang you, and nobody answered.
the Kaigen team
The integration decides whether it works
An agent that can talk about rooms but cannot see availability is a switchboard with better manners. The value sits entirely in the connection to the property management system and the booking engine.
Three questions separate a deployment that pays from one that annoys guests:
- Can it read live availability and rates, or is it reciting a stale price?
- Can it complete a booking, or does it collect a callback request that lands back on the desk?
- Does it know when to stop and pass a caller to a human, cleanly, with context?
The third is the one that protects the guest experience. A confident agent that will not admit it is out of depth does more damage in hospitality than in almost any other sector, because the caller is often already a guest and the brand promise is service.
After hours is a different job
Overnight calls skew away from new bookings and towards in house guests: a lock that will not read, an early taxi, a question about breakfast. Independent properties and small groups often have nobody on the desk at all past a certain hour.
Handling those well is a guest satisfaction play rather than a revenue one, and it shows up in reviews rather than in the ledger. It is worth separating from the direct booking case when you are building the argument internally, because the two get measured differently and conflating them makes both look weaker.
Q1
How many calls went unanswered last month?
Your phone system knows. Almost nobody pulls the report, which is why the problem stays invisible.
Q2
When do they cluster?
Overlay missed calls on the rota. The peaks are staffed hours far more often than teams expect.
Q3
What is your direct to OTA mix?
The commission spread on shifting a few points of that mix is the whole business case, and you already know both numbers.
We took this apart at a group scale in the direct booking deep dive, and 1st Step Hotel is a live deployment of the same idea. The hospitality page covers how the front desk agent is put together, and how we work explains what stays on your plate.
KEY TAKEAWAYS
- Missed calls are the one revenue leak that leaves no record in the systems you use to measure revenue.
- The business case is unusually clean: it is the OTA commission spread on bookings the caller was already trying to make directly.
- Most losses happen during staffed hours, because two concurrent calls beat one receptionist regardless of skill.
- The revenue calls are a narrow, repetitive set. Complaints and group enquiries should still reach a person.
- Value comes from the property management system integration, not the voice. Live availability or nothing.
COUNT THE LEAK
Want to know what your missed calls are worth?
Twenty minutes on your unanswered call report, when the peaks fall, and what shifting a few points of OTA mix would be worth.
Book a 20 minute audit →FAQ
Will guests know they are talking to an AI?
Often yes, and the design should not hide it. In hospitality the honest framing works better than the alternative: the caller reached someone immediately rather than a voicemail, and gets handed to a person the moment the request needs judgement. Trying to pass a synthetic agent off as staff is also a regulatory risk in several markets, which our compliance guide covers.
Can it take a booking, or only pass a message?
That depends entirely on the property management system integration, and it is the question worth asking first. An agent that collects a callback request has moved work back to the desk rather than removing it. One that reads live availability and completes the booking captures the revenue the call was carrying.
What about complaints and complex requests?
Those should reach a person, and the handover matters more than the handling. A clean transfer with the context already gathered is a better guest experience than an agent attempting to resolve something requiring judgement. Set the escalation boundaries deliberately at build time rather than discovering them from reviews.
Is this worth it for a small independent property?
The case is often stronger for small properties, because the concurrency problem is sharper when there is one person on the desk and no overnight cover at all. The relevant comparison is not against a large front office team, it is against the calls currently ringing out.
