Reduce Front Desk Calls: The Real Cost of Routine Questions

The last time I checked into a resort with two tired kids, the person ahead of me at the desk was asking what time breakfast started. The person behind me wanted to know where to park. The agent answered both with real warmth, and I remember thinking she had probably said those exact sentences thirty times since lunch. Routine questions cost a resort real money because each one takes handling time plus an interruption, and the fix is moving that volume to channels guests already use, not adding headcount. If you are weighing tools, start with our guide to AI concierge platforms for hotels. If you want the math first, keep reading.

Why does the front desk feel busier than the occupancy says?

Because the same questions are landing on fewer people. In the American Hotel & Lodging Association (AHLA) and Hireology survey published in early 2025, 65% of hotels reported staffing shortages, 71% had openings they could not fill, and front desk roles were the second most mentioned shortage, at 26% of properties, after housekeeping ^[1]. Hotel employment was still almost 10% below its pre-pandemic level in 2025 ^[1].

The money side is no kinder. CBRE's 2025 review of about 2,600 US hotels found salaries, wages and benefits rose 4.8% in 2024 while total revenue rose 2.3% ^[2]. Since 2019, hours worked fell 7.4% while compensation dollars rose 22.1% ^[2]. Fewer hours, costing more, carrying the same question load. That is why a 70% occupancy Tuesday can feel like a sold-out Saturday.

I think the honest framing is that the desk is a queue, and the queue got longer without anyone deciding it should. It helps to understand why the staffing gap is not closing on its own before you budget around it. Treating the desk as a queue is also what lets you put a number on the problem, which is the part owners respond to.

What does one repetitive question actually cost?

A routine question costs two things: the minutes to answer it, and a recovery penalty, because the agent was in the middle of a check-in, a folio correction or a conversation with a guest who needed real help. Anyone who has tried to finish a sentence while a small child asks the same question four times knows the penalty is real. The question takes ten seconds. Getting back to where you were takes longer.

A front desk agent at the end of a double feels this penalty more than she can explain. The questions were all small. The day was still exhausting, because she never got to finish anything. That is the cost that does not show up on a schedule, and it is the one that drives turnover.

So the model needs both numbers. For the wage, I used the US median hourly wage for hotel, motel and resort desk clerks, $16.86 in the Bureau of Labor Statistics (BLS) May 2025 data ^[3]. For the loaded cost I assumed 1.3 times wage, about $21.92 an hour or roughly 37 cents a minute. That multiplier is an assumption, and yours may be higher once you include taxes, benefits and uniforms. Change it and the table below changes with it.

Everything else in the model is also an assumption. There is no public benchmark for front desk questions per room per day. I looked. So the volumes below are a plausible day at a hypothetical 150 room resort running 75% occupancy, about 112 occupied rooms. They exist to show the shape of the cost, not to tell you what yours is. Your tally sheet will. ResortConcierge AI Research hit the same wall: its report on where front desk hours go found no independent time study in the published evidence, and lays out a week-long desk tally to run instead.

What are the 40 questions, and which ones are routine?

Every desk has a repertoire. Pool hours, breakfast hours, Wi-Fi password, where to park, how to get to the beach, can we get more towels, is there a pharmacy nearby, what time is checkout, can we check out late. Write them down for one week on a tally sheet by the phone and you will find most desks hear forty or so distinct questions on repeat. Then sort them. Routine means the answer is the same for every guest. Judgment means the answer depends on the guest, the situation or a feeling.

Illustrative: what a day of questions costs a 150 room resort at 75% occupancy

Read the bottom line slowly. About 530 desk minutes a day is nearly nine hours, or around 265 hours a month, which is more than one and a half full-time people doing nothing but answering questions. The first two rows alone are 220 minutes a day. That is the volume you are paying a human to repeat.

If you want to try your own figures, you can run your own numbers in the ROI calculator rather than rebuilding the spreadsheet.

How much of that volume can move to self-service?

More than most GMs expect, because guests are asking for it. In eviivo's survey of 430 travelers in late 2025, 58% said they prefer SMS, WhatsApp or email links for quick exchanges with a property, and only 4% wanted to download a hotel's app or scan a QR code to communicate ^[5]. Read that as a preference for channels they already have on the phone in their hand.

The appetite for self-service is not new. Oracle Hospitality and Skift surveyed 5,266 consumers in 2022 and found 73% were more likely to stay at a hotel offering self-service technology, and 38% wanted a fully self-service model with staff available on request ^[4]. That study is four years old now, which is why I pair it with the eviivo data, but the direction has held.

Map that onto the table. Hours, directions, Wi-Fi, housekeeping requests, local recommendations and simple bookings are all one-right-answer questions, and together they are 362 of the 530 minutes in the model, about 68%. Late checkout, billing and complaints are the judgment rows. They stay with people, and some of them happen at night, which is its own problem: covering the desk after 10pm deserves a plan of its own.

The caveat I would put in front of any owner is that these shares are a model, not a promise. Nobody can tell you what percentage of your calls will deflect until you have run it. What the data supports is the direction: routine volume wants to be written, and guests prefer it that way.

What can a hotel automate without losing the personal touch?

The personal touch is the judgment queue. It is the agent who notices a guest's voice tighten when the room is not ready, the one who has the time to walk a family to the elevator. None of that happens while she is explaining the parking garage for the twenty-third time. Automation buys the minutes the personal touch needs.

The general practice is a written front door. Guests reach the property by web chat from a QR code or a stay link, by SMS or by WhatsApp, and the routine questions are answered from one property knowledge base. ResortConcierge AI does this from a stay link or an in-room QR code (early access), with nothing to download, which is where most of the routine volume lives, and it is the core of moving routine questions off the desk.

Anything the knowledge base cannot answer, and any message with a negative tone, needs a person, and it needs one fast. The useful version of that is not a notification that pings everyone. It is an escalation queue your team works from, so the desk gets exceptions instead of interruptions. ResortConcierge routes anything the knowledge base cannot answer into that queue with the guest, room, reason and sentiment attached, and staff can add any other conversation.

Two things to say plainly. Voice can be added through early access; without it, plan on the ringing phone staying with your team. And ResortConcierge can work alongside property management systems (PMS) such as Opera, Mews and Cloudbeds, with reservation data set up during onboarding. The written channels and the knowledge base work standalone.

How do you measure whether front desk load actually went down?

Four weekly numbers, tracked against a 90 day baseline you take before changing anything. First, routine questions that still reached a person, from the same tally sheet you started with. Second, average wait at the desk during your worst hour, timed by hand if you have to. Third, the share of escalations handled within your target, whatever you set it to. Fourth, after-hours contacts that woke someone up.

Pick a target for each before you start, and make it modest. Cutting routine questions reaching a person by a third in 90 days is a result an owner can see. Promising to eliminate them is a result nobody believes. The baseline week will also tell you something uncomfortable: how many of those questions were already answered in the confirmation email that nobody read.

The baseline matters more than the tool. I used to tell people in counseling that you cannot know whether something is working if you never wrote down where you started, and the same is true of a desk. A GM who walks into an owner meeting with "it feels calmer" loses to one who says "routine questions reaching a person fell from 65 a day to 22, and here is the sheet."

ResortConcierge's analytics shows question volume by topic and hour, which is the tally sheet done for you once the written channels are live. Until then, paper works. Either way, the numbers belong in the same budget conversation as a realistic staffing plan for 2027, because the point was never fewer people. It was putting the people you have in front of the guests who need them.

Start with the tally sheet this week. One sheet, seven days, every question that reaches the desk. When you can see the shape of your own queue, the next decision, whether that is a hire, a tool or a founding partner seat, gets a lot easier to make and a lot easier to defend.