Guatemala Rent a Car: from a chatbot that only chatted to bookings confirmed from the first hour
Having an AI chatbot is not the same as selling with it. Guatemala Rent a Car learned that with one that answered well and stopped right when the customer wanted to book. With an agent integrated with Karve, the booking is confirmed on WhatsApp without anyone on the team stepping in.
Summary
- Guatemala's largest national car rental company had an AI chatbot on its website: it chatted, guided and stopped at the moment of booking.
- Every booking waited for an agent to pick it up; replies were delayed and a share was lost.
- An AI agent integrated with Karve, with WhatsApp as the main channel, confirmed bookings from the first hour in operation.
- 81% of sales conversations are resolved without the team, nearly one in three quotes ends in a booking and returning customers are already recognised.
Guatemala Rent a Car's chatbot was not bad. It answered questions, guided the customer and spoke naturally. The problem was structural: it was not connected to the reservation system, so at the decisive moment, when the customer said "I want it", it had to hand over to a person. And there, between the wait and the change of channel, bookings were lost.
This case matters because many companies are at that point today: they already have "AI" on the website and do not see the return. It looks at what changed when a bot that chats was replaced by an agent that transacts in Karve, and why WhatsApp, the most used channel in the region, was the way in.
1. Guatemala's largest national car rental company
Guatemala Rent a Car is the largest national car rental company in the country. It runs on Karve as its reservation system and serves customers who, in the region, prefer WhatsApp to any other channel for asking and booking.
Before the change it had already bet on artificial intelligence: a website chatbot that chatted well and took repeated questions off the team. What it did not do was close.
2. The challenge: a bot that reached the door and did not go in
As soon as the customer wanted to book, the chatbot stopped and an agent had to step in. The booking sat waiting for someone to pick it up, the reply was delayed and a share of those bookings was lost along the way. The web bot did not allow pausing and resuming either: if the customer closed the tab, the conversation was gone.
The company needed the conversation to end in a booking confirmed in Karve, with no waiting, and the customer to be able to come back to the conversation whenever they wanted.
3. What was set up: an agent connected to Karve, with WhatsApp as the main channel
A car rental AI agent was integrated with Karve and WhatsApp was switched on as the main channel. From the first hour in operation it quoted with real availability and rates and confirmed bookings without the team stepping in. WhatsApp also changed the relationship with the customer: they are identified by their phone number, can pause the conversation and pick it up later, and if that number already exists in Karve, the agent recognises them and does not ask for their details again.
- Quotes and confirms in Karve, with live availability and rates, from the WhatsApp conversation itself.
- Recognises the customer by phone number: an automatic sign-in, convenient for the customer and for the company.
- When a person is needed, it hands over with the full context and the agent works from the conversation: details already given come pre-filled, and they quote or close against Karve in seconds.
- Improves with use: questions from real conversations go into its knowledge base and its answers are reviewed against them.
4. Results
- 81%
- of sales conversations handled by the AI with no team involved
- Nearly 1 in 3
- quotes ends in a confirmed booking
- 1 in 3
- conversations arrive outside office hours
Bookings that used to wait for an agent are confirmed on the spot, and the team steps in only where needed, with the full conversation in view. Conversion improves in those cases too, because the human agent quotes and closes from the same conversation without typing anything again.
“Chats get resolved without anyone on the team stepping in and conversion has improved. And when someone on the team does step in, they quote and close the booking from the same conversation in seconds.”
Read the full Guatemala Rent a Car case study on Bota's website →
5. What you can copy from this case
- Ask your current chatbot what happens when the customer says "I want it": if the answer is "I notify an agent", it does not sell, it qualifies.
- Connect the agent to the reservation system before anything else; without real availability and rates, AI can only chat.
- Start with the channel your customers already use; across much of Latin America that is WhatsApp, and it also identifies the customer by number.
- Demand that the agent learns from real conversations: what it escalates today should be answerable next month.
Frequently asked questions
Why didn't the previous chatbot book?
Because it was not connected to the reservation system. It could chat and guide, but at the moment of booking it had to pass the conversation to a person.
What does the customer gain with WhatsApp over the website chat?
They are identified by their phone number, can pause and resume the conversation whenever they want and, if they are already a customer in Karve, do not have to give their details again.
How long did it take to confirm the first booking?
It confirmed bookings automatically from the first hour in operation, with Karve's availability and rates.
Written by
Johan Smith
Car Rental Editorial Team
Sources
- Guatemala Rent a Car case study on Bota — figures and implementation details published by the customer.
- Guatemala Rent a Car — the car rental company's website.
- Karve — reservation system used in this case.