AI Customer Service: How Guest Assistance Is Changing
AI customer service allows property managers to respond to guests automatically, accurately, and in multiple languages, 24 hours a day, without delegating every single message to human operators. In the Italian short-term rental sector, where a manager with 15 apartments receives an average of 80–120 messages per week, automating guest assistance means recovering concrete hours every day and reducing communication errors.
Capitolo 01 / 08
01 — Why Manual Customer Service Doesn't Scale
Handling guest messages manually works when you have 2 or 3 apartments. Beyond that threshold, the model breaks down.
A guest arriving at 11 PM who can't find the check-in instructions won't wait until morning: they leave a negative review, or worse, call the emergency number. A German guest writing in German expects a reply in German, not a generic automated message in English. A guest asking whether parking is available wants to know whether their parking space is available, not a standard policy response.
The problem isn't the volume of messages: it's the variety. Every guest has a different situation, a different schedule, a different language. Manual customer service scales poorly because it requires contextual attention every single time.
Dato
According to industry estimates, a property manager with 10 apartments spends an average of 2.5 hours per day managing messages. On a monthly basis, that's roughly 75 hours — nearly two working weeks.
Dato
Tempo giornaliero messaggi ospiti: manuale vs automatico
Peak
150min/giorno
Gestione manuale
Range
30 – 150
min/giorno
The Hidden Cost of Delays
Every delayed response has a measurable cost. Airbnb penalises profiles with low response rates in search results. Booking.com displays average response time on the property listing. A 4-hour delay on a booking enquiry can mean losing the guest to a competitor who responded in 10 minutes.
Automated customer service is not a convenience: it's a competitive requirement.
Capitolo 02 / 08
02 — How AI Customer Service Actually Works
The term "AI customer service" covers very different technologies. It's worth distinguishing them, because the differences directly affect the guest experience.
IF/THEN Rules (This Is Not AI)
Many tools on the market operate on conditional logic: if the guest writes "check-in", reply with message X. If they write "wifi", reply with message Y. These systems are quick to configure but brittle: they don't understand context, they can't handle compound questions, and they don't truly converse with the guest.
A guest who writes "Hi, what time can I arrive tomorrow and is there parking?" is asking two questions in one sentence. A rule-based system answers only the first, or doesn't respond at all.
AI Chatbots with Language Models
AI customer service systems based on large language models (LLMs) understand text contextually. They recognise compound questions, infer the guest's intent, and respond coherently even when phrasing is unusual or the language is mixed.
The practical difference: an AI chatbot understands "there are 4 of us plus a small baby, is there a cot?" and responds with information specific to that apartment, not a generic reply about the children's policy.
Integration with the PMS
AI customer service becomes genuinely useful when connected to real data: booking dates, check-in times, apartment-specific instructions, availability. Without integration with the property management system, the AI responds generically and often incorrectly.
See how Verto AI integrates with Smoobu via live native API: every response uses up-to-date booking data, not static templates.
Percorso
Livelli di automazione del customer service
Manuale
€0 di strumenti, alto costo in oreIl property manager risponde a ogni messaggio personalmente, in qualsiasi orario.
Template manuali
€0Risposte preimpostate copiate e incollate. Risparmio parziale, nessuna automazione reale.
Regole IF/THEN
€5-15/apt/meseMessaggi automatici su trigger predefiniti. Fragile su domande composite o inusuali.
Verto AI
da €21,90/apt/meseAI conversazionale su WhatsApp con integrazione PMS live. Capisce contesto, lingue e domande composite.
Capitolo 03 / 08
03 — WhatsApp as a Channel: Why It Changes Everything
The choice of channel is not a technical detail. It's the variable that determines whether the guest actually reads the response or not.
Email has an average open rate of 28% in the hospitality sector. SMS and push notifications from proprietary apps aren't much better, especially for international guests who don't want to install an app they'll use for three days.
WhatsApp has an open rate above 90% and is read on average within 3 minutes of receipt. It requires no download, no registration, and works in 180 countries. For a guest arriving from abroad, it's already installed on their phone.
Attenzione
Pay attention to the difference between standard WhatsApp Business and the official Meta API. The former has strict automation limits and risks account suspension. The latter — the one used by Verto AI — is the only official channel for integrating AI customer service on WhatsApp in a way that complies with Meta's policies and GDPR.
Zero Setup for the Guest
The concrete advantage of WhatsApp as a channel for automated customer service is that the guest doesn't have to do anything new. They receive a message on an app they already use and reply as they would with any contact. No account to create, nothing to download, no interface to learn.
This drastically reduces friction and increases the effective usage rate of the service.
Capitolo 04 / 08
04 — What an AI System Handles (and What It Doesn't)
One of the most common questions among property managers is: "Does the AI actually understand what guests are asking?"
The honest answer is: it depends on the system and how it's configured. Here is a realistic map.
| Request Type | Manageable with AI | Notes |
|---|---|---|
| Check-in / check-out times | ✅ Yes | Data from the booking |
| Apartment access instructions | ✅ Yes | Configured per unit |
| WiFi password | ✅ Yes | Static per apartment |
| Questions about local services | ✅ Yes | With knowledge base |
| Booking date change request | ⚠️ Partial | Depends on PMS integration |
| Urgent technical issues (breakdown) | ❌ No | Escalation to human operator |
| Complex complaints | ❌ No | Requires human judgement |
| Price negotiation | ❌ No | Strategic decision |
A good AI customer service system also knows when not to respond on its own: it recognises situations that require human intervention and flags them immediately to the property manager, without leaving the guest waiting.
Tip
Configure automatic escalation for keywords such as "breakdown", "emergency", "dissatisfied": the AI responds that you are taking action and notifies the manager in real time.
Multilingual Without Configuration
One of the most practical aspects of modern AI customer service is automatic language handling. A system based on LLMs recognises the guest's language and responds in the same language, without the property manager needing to configure anything.
Verto AI handles more than 25 languages, including less common ones such as Czech or Polish — relevant for managers overseeing apartments in tourist cities with diversified international flows.
Capitolo 05 / 08
05 — Metrics That Improve with Automated Customer Service
Automating guest assistance is not an end in itself. The goal is to improve concrete indicators that impact the business.
Response Rate
Airbnb calculates the response rate based on first replies within 24 hours. With an active AI customer service system, the response rate rises to 100% by definition: every message receives an immediate reply, even at 3 in the morning.
Average Response Time
Average response time drops from hours to seconds. On Booking.com this figure is visible to guests on the property listing and influences their choice. A response time of less than 1 minute is a measurable competitive advantage.
Reviews
The correlation between response speed and final rating is well documented: guests who receive fast, relevant responses during their stay tend to leave higher reviews. Not because they are more satisfied with the apartment itself, but because they felt looked after.
Dato
Property managers who automate customer service report an average increase of 0.2–0.4 points in their average Airbnb rating within the first 3 months of use.
Dato
Metriche operative: prima e dopo l'AI customer service
Hours Saved
Time savings are the most immediate metric. With an automated customer service system handling 80% of requests independently, a property manager with 15 apartments recovers approximately 1.5–2 hours per day. On an annual basis, that's more than 500 hours — time that can be reinvested in acquiring new units or maintenance.
"I hadn't realised how much time I was spending answering the same questions over and over, until I stopped doing it."
— Property manager, Florence, 12 apartments
Capitolo 06 / 08
06 — How to Choose the Right System: Operational Checklist
Not all AI customer service systems are suited to the Italian short-term rental market. Before choosing, verify the following points.
- Uses WhatsApp via official Meta API (no workarounds)
- Integrates with your PMS (Smoobu, Lodgify, Hostaway)
- Understands text contextually, not just keywords
- Handles voice notes and photos (guests send them frequently)
- Supports at least 10 languages
- Has automatic escalation to the manager
- Is GDPR-compliant (guest data processed within the EU)
- Requires the guest to install an app
- Works only via email or SMS
- Does not connect to the PMS in real time
The last three points are negative signals: if a system exhibits them, the guest experience will be worse than well-executed manual assistance.
You can explore a detailed comparison of the main solutions on the market on the comparison page with Hospitable, which analyses the differences between rule-based and conversational AI approaches.
Capitolo 07 / 08
07 — Verto AI: AI Customer Service on WhatsApp for Italian Property Managers
Verto AI is an AI concierge that handles guest assistance entirely on WhatsApp, via the official Meta API. It requires no additional action from the guest: they receive a message on WhatsApp and reply on WhatsApp.
The system is designed specifically for the Italian short-term rental market: it understands typical guest requests (check-in, instructions, local services, issues), responds in more than 25 languages, and integrates with the main PMS platforms via native APIs.
Explore all of Verto AI's features — from multilingual management to automatic escalation, through to the Analytics add-on for monitoring conversations.
Plans start from €21.90 per apartment per month for managers with 10 or more units on an annual contract. For those starting with 1–3 apartments, the cost is €29.90 per apartment per month. The Analytics add-on is available at €4.90 per apartment per month.
If you want to see how it works for your specific situation — how many messages you handle, in which languages, with which PMS — you can request a free demo. No setup required, no commitment.
Capitolo 08 / 08
Frequently Asked Questions
Does AI customer service completely replace the property manager?
No. AI customer service systems independently handle routine requests (check-in, instructions, WiFi, questions about services), but complex situations — breakdowns, complaints, negotiations — still require human judgement. The goal is to eliminate repetitive work, not to replace strategic decisions.
How long does it take to configure an automated customer service system?
It depends on the PMS integration. With Verto AI and Smoobu, basic configuration takes less than 2 hours: units are connected, apartment-specific instructions are uploaded, and escalation is set up. The first automated replies to guests go out the same day.
Do guests realise they are talking to an AI?
It depends on how the system is configured. Some property managers prefer full transparency; others use an assistant name without specifying the AI nature. In either case, what matters to the guest is receiving an accurate and prompt response. The quality of the reply is more relevant than its origin.
Does automated customer service work for longer stays too?
Yes, and it is particularly useful in this context. Guests staying 30 or more nights generate more messages over time, with more varied requests. An AI customer service system maintains consistent responses throughout the entire stay, without the manager needing to remember the details of every previous conversation.
How does AI customer service integrate with Airbnb and Booking.com?
AI customer service systems do not connect directly to OTA platforms to reply within their internal chat — those remain managed by the property manager or the PMS. The AI operates on the WhatsApp channel, which is activated after the booking is made. This allows operational communication (instructions, questions during the stay) to be moved to a more efficient channel, leaving the OTAs for the commercial phase.
Keep going
Related resources
Same topic
01Want to see Verto on a real conversation?
No credit card required. Self-service setup in 30 minutes. Plugs into the channel manager you already use.