AI-POWERED TRAVEL ENGINE
Travel planning made effortless with AI-curated packages
Context
With the rise of conversational AI and personalized services, travelers increasingly expect faster, smarter ways to plan trips. Avero is an AI-powered mobile app designed to simplify travel planning by curating optimized flight and hotel packages through a conversational interface.
My role
The challenge
While AI-powered apps have gained popularity, many still fail to deliver on their promise of intuitive and reliable experiences. The challenge was to design an AI assistant that could not only understand user preferences but also deliver high-quality, personalized travel packages that feel reliable, accurate, and human-like — all while avoiding common pitfalls like irrelevant recommendations and awkward interactions.
Design process
Our client, a leading technology company, aimed to revolutionize scheduling processes worldwide by introducing the world's first AI-powered scheduling app.
Phase I:
Empathyze
Our first step was to outline the research question that would guide our research plan:
How might we create an AI-powered travel assistant that provides personalized, reliable travel packages in a seamless, human-like conversation?
To uncover underlying issues and formulate our hypothesis, I conducted five remote user tests on the existing website. These tests were aimed at identifying the pain points users faced when interacting with current travel booking apps and understanding how they would react to a more personalized, conversational approach.
SAMPLE:
Age: between 30-40 years old
Location: Netherlands, Germany and Spain.
KEY INSIGHTS:
AI interactions felt impersonal: Users were frustrated with generic suggestions and lack of contextual understanding.
Lack of trust in AI: People preferred interacting with apps that felt human, not robotic.
Overwhelming choices: Users didn’t want to sift through endless options, especially with AI-generated results that didn’t seem to match their needs.
Phase II:
Definition
Based on these insights, I crafted a hypothesis to address the AI experience gap:
HYPOTHESIS:
“If we design an AI assistant that mimics human conversation, suggests tailored travel packages based on a deeper understanding of preferences, and allows users to easily tweak recommendations, travelers will feel more in control and confident in their choices.”
THE ROADMAP:
Build a human-like conversational interface that feels natural.
Ensure AI can learn and adapt based on user preferences, reducing irrelevant recommendations.
Design for transparency so users understand how AI arrives at recommendations.
Phase III:
Ideation
To solve the AI shortcomings, I explored various chat flow ideas and interactions, including:
Using natural language processing to interpret not just the words but the context behind them (e.g., "I want to relax" = beach destinations).
Implementing real-time preference updates, allowing users to fine-tune their options and see how the AI adapts.
Designing personalized suggestions that felt intuitive, offering users choices they could easily trust without being overloaded.
Early wireframes were tested for clarity — ensuring users could engage with the assistant without getting lost in technical jargon or endless options.
Phase IV:
Design & Prototype
Focusing on clarity and simplicity, I designed high-fidelity screens for the app:
A chat-based interface that felt warm and conversational, with quick responses and tailored suggestions.
Clear visual indicators showing users how their preferences influenced the AI's package choices.
Interactive elements that allowed users to adjust their travel preferences on the fly, like “change my hotel,” and see updated package options instantly.
Key takeaways
Through this process, I learned that human-like AI interactions are key to increasing user trust and engagement. Users respond much better to an AI that feels empathetic and adaptive, rather than robotic or rigid.
Transparency also played a critical role in building trust; users need to understand why the AI is suggesting certain options, which empowers them to make more informed decisions.
Lastly, I found that real-time adaptability was essential for user satisfaction. When users could tweak their preferences and immediately see the results, they felt more in control and confident in their choices, which led to a better overall experience.