Flytrip: WhatsApp AI Agent for a Panama Airline

The Problem

With more than two decades of experience, Flytrip expanded into scheduled commercial airline operations in 2023. Alongside its scheduled routes, it offers private and executive flights, tourist flights, pet travel, medical evacuation, and cargo services.

Like many travel and transportation companies in Latin America, Flytrip's customer journey runs largely on WhatsApp. And as Flytrip grew, so did the volume of enquiries landing in the WhatsApp Business channel. Questions on schedules, seat availability, baggage, fares, and reservation changes; the same questions were coming every day, phrased in a hundred different ways.
‍

The challenge: managing WhatsApp conversations at scale

Initially, a small team handled WhatsApp conversations using a generic chatbot solution. The team spent a lot of time replying to incoming messages.

In travel, static workflows struggle with how travellers actually write and have conversations. Flytrip was using a rule-based WhatsApp chatbot, where every new customer path needed its own workflow to be built and maintained, and travellers could still end up frustrated.

Rule-based, basic chatbots struggle in travel because travel language is complex and every traveller talks and types differently. One person asks about baggage one way, the next asks the same thing in a completely different way, in another language, or with shorthand and typos, and a rule-based bot can't keep up with that variety. So the common questions still reached the same two people, and most answers needed constant human supervision.

The result was an operational bottleneck. Repetitive enquiries were eating the time Flytrip's commercial staff could have spent on ready-to-buy customers, calls, emails, events, partnerships, or complex inquiries that need special attention. And with fares and schedules changing seasonally, there was always the risk of an outdated answer.
‍

What Flytrip wanted to achieve with Maya

Answer repetitive WhatsApp questions naturally, without a rigid script
Free up the team's time for commercial and purchase-ready conversations
Keep a clear human handover for complex requests and customers who ask for an advisor
Motivate users to move from booking via WhatsApp to booking directly on the website once their questions are answered
Preserve the Flytrip voice: having an agent that actually feels like part of the team, speaking as "we," not a disconnected third-party bot

Why Maya?

Flytrip deployed Maya on WhatsApp, using its FAQs, website content, and historical WhatsApp conversations as the knowledge foundation.

Instead of forcing travellers down a decision tree, travellers ask questions in their own words, and Maya understands traveller context and intent while answering using only Flytrip's knowledge, in the traveller's language, around the clock.

When a traveller is ready to purchase, asks for a person, or raises something Maya can't resolve, the conversation is handed to a Flytrip advisor with full context already attached.

‍

Crucially, the team no longer needs to supervise every chat from the first message. They step in when a conversation is routed to them, usually when it matters most commercially.
‍

An intelligent conversational layer, not a chatbot

Beyond answering questions, Maya gives Flytrip a clearer view of their own customers and their own content:

Voice and image understanding means travellers can send voice notes or screenshots (like a flight confirmation or change) and still get a useful answer.
Direct knowledge updates let frontline team members check a response and correct Maya's information themselves, instead of routing every fix through one person.
Maya Insights surfaces conversation topics, customer intentions, recurring patterns, and knowledge gaps.

This surfaced things Flytrip hadn't realised were missing. Knowledge Gaps flagged absent baggage-dimension details and helped catch an incorrect Contadora schedule buried in existing content. In other words, customer conversations quietly became an audit of the company's own information.

Results

Flytrip set out to give a lean team room to breathe, and that's what happened after implementing Maya.

"Questions that a chatbot or prior solutions can't answer are answered accurately by Maya. Our agent, powered by Maya, has evolved at an impressive rate and knows how to interact with our customer base from around the world.

For Flytrip, it's a fantastic software, and the attention we've received from Maya's team during and after implementation has been 10/10."

Amanda Hernández, Planner & Reservations Manager, Flytrip

With Maya handling hundreds of questions each week, the team's interaction with travellers now happens only when a traveller has direct booking intent. That freed-up attention went to high-impact work for the Flytrip commercial team: reservations management, events, partnerships, calls, emails, and more complex enquiries that need the human touch.

Getting the Maya-powered agent to speak Flytrip's brand tone of voice and know its routes took a few rounds of agent calibration. The Flytrip team describes the current agent as completely different from the one from 12 months ago, more versatile and capable of understanding complex, counterintuitive questions in over 50 languages.

Get Answers from Travel AI Experts

Talk to the team building conversational AI agents for over 100 travel brands. Thirty minutes, no pitch, no cost.

Talk to us