eSky is a leading online travel agency offering flights, hotels, and Flight+Hotel packages (Holidays and City Break), as well as a wide range of ancillary products such as travel insurance, car rentals, transfers, and others. Operating across multiple brands and markets, eSky combines advanced technology with Tourist Guarantee Fund protection to deliver comprehensive, reliable, end-to-end travel booking.

For a travel enterprise operating across more than 50 markets, customer support rarely happens in calm moments. It happens when travelers are under pressure, at the airport, or trying to salvage a trip they have planned for months. And during irregular operations (IROPs)—flight cancellations, diversions, missed connections—contact volumes can spike dramatically within minutes, making it nearly impossible for human agents alone to keep up.

To better support their customers, eSky needed more than automation that simply deflects tickets instead of resolving them. They also needed a better operating model for customer experience: one that could improve service quality, lower cost-to-serve, and scale across the channels customers already use.

Transitioning from flow-based chat to AI customer service

Lukáš Maršálek joined eSky Group as the Digital Customer Support Manager to explore opportunities to that end. “I’ve built a flow-based chat before, and when I joined eSky, I built a flow-based chat again,” Lukáš reflected. “But the experience with flow-based ‘chatbots’ is the same across the board: customers would look for the quickest way to get to a human rather than believing that the chatbot can actually solve their problem.”

At the time, it had become common practice to stand up a flow-based chatbot to manage inquiry volumes by deflecting tickets. “You’re basically trying to manage volume spike and hoping that helps,” he explained. But neither Lukáš nor eSky was satisfied with the status quo. “As an organization, we had to start believing that things can be different. So we decided to go for a proof of concept with Ada.”

Less than a year later, eSky has AI agents delivering personalized customer experiences across three distinct brands, on multiple channels, in multiple languages, markets, and time zones. They have also meaningfully operationalized AI within the organization such that it is connecting siloed efforts, surfacing workflow improvements, and delivering against business goals.

“The moment where we started to believe that we can actually resolve inquiries automatically was a game changer,” Lukáš noted, beaming. “Ada has been a huge contributor to my mission so far. Less than a year ago we were running a proof of concept. Today we're live across three brands, three channels, and dozens of markets. I'm having the best time and we're just getting started."

AI customer service: from point solution to operating model

eSky’s success wouldn’t have been as transformative without Lukáš’s game-changing perspective on AI.

When Lukáš first said his goal was to improve customer experience while reducing costs, people saw a contradiction. Now, they hear “AI” and assume that’s the answer.

“Sometimes it seems like businesses have gone from ‘this is impossible’ to ‘AI will solve everything,’” he said. “And in doing that, they’re overlooking other opportunities for leverage.”

For eSky, the shift wasn't about simply adopting AI. It was a company-wide commitment—backed at the highest level—to rethinking how customer experience is delivered. That meant being deliberate about the approach: AI only delivers real impact when it's applied across the full customer experience, alongside the right systems, processes, and integrations, and when the whole organisation is aligned behind it.

This mindset also put their support team at ease. Lukáš built that confidence by showing—not telling—what AI could take off their plate. By demoing how the AI agent handles repetitive, manual tasks like updating passport details, he made it clear that automation wasn’t about replacement, it was about focus.

As a result, human support agents can spend less time on routine work and more time on complex, meaningful requests. It not only improves the quality of support, but helps prevent burnout in a high-pressure environment.

“AI is not here to take our jobs,” Lukáš said. “It’s here to enhance what we can do, and enhance the customer experience.”

Growing confidently with operational excellence

With that guiding principle in mind, the eSky team made decisions that considered the entirety of the customer experience, looping in cross-functional collaborators as needed.

Three brands, one team, one platform

The eSky Group operates across multiple brands—eSky, eDestinos, and Thomas Cook—each serving distinct customers with different expectations. Behind it all is a lean team managing the entire operation. So how do they deliver personalized experiences at that scale? They start by simplifying what can be simplified.

eSky centralizes its knowledge base in a single language, English. This keeps content easy to maintain, while Ada’s multilingual capabilities translate each experience in real time based on customer preference. As the team puts it, “we just flip the language, and Ada does the job.”

Where personalization matters most, they add control. Instead of forcing a single AI agent to serve every audience, eSky runs a dedicated AI agent for each brand. That decision allows each experience to reflect the nuances of its market without compromise. And it doesn’t add operational overhead. With Ada, all three AI agents are managed from a single platform, with a unified view of performance across brands.

WhatsApp: Meeting customers on their preferred channel

After seeing strong results on web chat, the next step was clear: expand to the channels customers already use. “When we were thinking about a customer that's in distress—they're at the airport, they need urgent support—they’re usually most comfortable with picking up the tool that they normally use for communication,” Lukáš explained. For eSky, that’s WhatsApp.

And it makes sense. WhatsApp is a familiar channel for eSky’s European and Latin American customer base, offering a persistent thread that anxious travelers can return to, revisit, and rely on. It was also a natural extension of how eSky already operates. The team has always focused on chat as a format, not just a channel, and WhatsApp mirrors a similar conversational dynamic.

So the approach was simple: meet customers there, launch the channel, and see how it performs. The team’s bet was a winning one. Customers adopted the channel quickly.

Messenger: Using AI to build a bridge between CX and Marketing

Another chat-based channel that the team was looking to expand to was Messenger, but it presented an unexpected challenge: Marketing.

As part of the Meta ecosystem, Messenger typically falls under the mandate of the Marketing team who use it to engage with their audience. But it’s also a channel that customers often use to reach out to businesses with their issues. “Messenger is an interesting one that I see businesses struggle with because it’s kind of a mixed bag,” Lukáš reflected. “It's a channel where you get engagement from your social posts and also one where customers go to contact you.”

The team knew this was exactly the kind of operational advantage that AI brings to the table. Lukáš approached the Marketing team with a solution: using an AI agent to automatically recognize and resolve customer inquiries, or triage them right to the support team if necessary, and Marketing can still retain full ownership of the channel.

“I showed them what we do on web chat and WhatsApp, and said we could set something similar up on Messenger. They were pretty excited about it,” he said. “It’s perfect because we can handle all kinds of queries under one roof, and it’s all orchestrated automatically.”

Using AI to guide decision making

Lukáš and his team also use AI to expose operational gaps across the business. “AI magnifies any kind of disconnect,” he explained. Some human-only CX teams face a structural challenge: they may not be aware of knowledge gaps because human agents tend to fill them in conversations with customers. On the other hand, AI will stick to knowledge resources fully and reveal any gaps instantaneously.

“We can see everything. We can see where the points of failure are because they always are manifested in the communication channel,” Lukas said.

By plugging Ada into their enterprise AI infrastructure via Ada's MCP Server, the team can quickly surface the areas that need their attention most. By combining AI, analytics, and manual efforts, they identify improvement opportunities and pinpoint root causes for issues.