Autos Valls: from the office WhatsApp to an agent that sells in every customer's language
A tourist car rental company in Menorca, with an airport desk and demand that explodes in summer, now answers calls and WhatsApps with AI: 81% of sales with no team involved, twelve languages and a team that grows and shrinks with the season.
Summary
- WhatsApp was answered from the office phone, between the counter and the calls, with no view of the channel.
- Today calls and WhatsApps land in one inbox, sales, questions and support alike, and the AI resolves 81% of sales conversations.
- The same agent has served in twelve languages and almost half of all conversations arrive outside office hours.
- The lesson: in a seasonal business, the messaging channel is the first thing to take off the shared phone.
Menorca lives on tourism, and a car rental company on the island lives on summer. Autos Valls has an airport desk, several branches and a demand curve that goes from almost nothing in winter to full capacity in July and August. That profile turns every service channel into a peaks problem: everyone has to be answered, in their own language, exactly when the counter is full.
This case matters because of how common it is. Many small and mid-sized car rental companies handle WhatsApp the way Autos Valls did: from a shared phone, whenever someone can. Here we look at what changed when that channel moved to an AI agent connected to the reservation system, and what can be copied without being a large company.
1. A tourist car rental company with extreme seasonality
Autos Valls is a family-run car rental company in Menorca. Its customers are almost all tourists, arrive through the airport and write in whatever language they speak: Spanish and English, but also German, French, Italian, Polish, Dutch or Hungarian. The company has to grow every spring and shrink every autumn, and so does the team answering customers.
Before the change, the company's WhatsApp lived on the office phone. It was answered between customers at the counter, whenever the calls allowed, and nobody had a full picture of the channel: how many people wrote, how many got no reply, in which language, at what time.
2. The challenge: languages, hours and season
Three factors added up. First, language: serving a Polish customer well on a Saturday in August needs someone who speaks it or a translation nobody notices. Second, hours: a very high share of messages arrived outside the working day, and those enquiries went cold until the next morning. Third, season: any solution had to hold up through July peaks without forcing hires in January.
The company was not looking for a chatbot to answer FAQs. It wanted the conversation to end in a booking written in its system, and its team, when it had to step in, to do so with the whole conversation in front of it.
3. What was set up: an AI agent on top of the reservation system
Autos Valls switched on a car rental AI agent on WhatsApp, phone and web chat, connected to its reservation system. The same agent handles all three channels with the same rates and the same rules, and everything lands in a shared inbox for the team, including questions and support requests, not only sales.
- Quotes with the reservation system's live availability and rates and confirms the booking in it; last-minute bookings go through the team under a rule set by the company.
- Replies in the customer's language, whichever it is, and when a person steps in the conversation arrives translated in real time both ways.
- When someone is needed, the alert includes the summary, the full conversation and a reply already drafted.
- The back office lets the team create groups, assign and escalate conversations and measure response times and handover reasons, conversation by conversation.
4. Results
- 81%
- of sales conversations handled by the AI with no team involved
- 12
- languages served by the same agent
- 47%
- of conversations arrive outside office hours
At seasonal peaks the agent has handled up to twelve sales conversations started within the same hour and close to thirty in a day, each in the customer's language. And for the first time the company has channel metrics: what gets asked, how long replies take and which conversations needed a person. It uses that data to size the team to the season.
“WhatsApp used to be answered from the office phone, whenever we could. Now everything lands in one inbox, we see every conversation and in high season the team keeps up with all of it.”
5. What you can copy from this case
- Take WhatsApp off the shared phone first: without a common inbox there are no metrics, and without metrics nothing gets sized.
- Demand that the agent books in your system rather than capturing a lead: the difference shows in the conversations that arrive at night.
- Set the auto-confirmation rule yourself (for example, only with enough lead time) and leave the rest to the team.
- Measure the language and hour of every conversation: they are the two data points that justify, or not, seasonal hiring.
Frequently asked questions
Do you need to change reservation system to do something like this?
No. The agent connects to the system the car rental company already uses and works with its availability and rates. Autos Valls did not change system.
What happens with last-minute bookings?
The company sets the rule. In this case, bookings with enough lead time are confirmed instantly and last-minute ones are reviewed by the team before confirming.
How does the team serve a customer who writes in a language it doesn't speak?
The conversation arrives translated in real time and what the agent writes is translated into the customer's language, with a reply already drafted to send.
Written by
Johan Smith
Car Rental Editorial Team
Sources
- Autos Valls case study on Bota — figures and implementation details published by the customer.
- Autos Valls — the car rental company's website.
Related articles
Related tools
Conversational sales
Bota
AI agent for WhatsApp and website bookings.