An AI Chat Agent Connecting Users to Personalized Travel Activities

Civitatis offers thousands of travel activities across multiple destinations and languages. Users often arrive at the website with open-ended questions like “What can I do in Rome this weekend?”, “I want a family-friendly activity in Paris.”, or “Are there Spanish-language tours in Lisbon?”. To provide faster, more accurate answers, Civitatis partnered with us to build an AI chat agent. The agent interacts directly with users, asking clarifying questions and retrieving answers from Timbal knowledge bases. The result: a conversational interface that guides users to the most relevant activities in real time.

• 73% faster activity discovery

• 11,000+ AI conversations

• 62% fewer support tickets

Key Results

• Up to 73% faster activity discovery

• Real-time personalized recommendations by destination and language

• Reduced friction in customer support interactions

• One conversational agent connected to Timbal knowledge bases

• Semantic search powered by natural language understanding

Overview

Civitatis offers thousands of travel activities across multiple destinations and languages. Users often arrive at the website with open-ended questions like:

• “What can I do in Rome this weekend?”

• “I want a family-friendly activity in Paris.”

• “Are there Spanish-language tours in Lisbon?”

To provide faster, more accurate answers, Civitatis partnered with us to build an AI chat agent. The agent interacts directly with users, asking clarifying questions and retrieving answers from Timbal knowledge bases.

The result: a conversational interface that guides users to the most relevant activities in real time.

The Challenge: Helping Users Find the Right Activity Quickly

Before the agent:

• Users navigated large catalogs with generic filters

• Keyword search returned imprecise results

• Language mismatches caused frustration

• Customer support handled repetitive questions

• Users struggled to find activities that matched their intent

Civitatis needed a solution that could understand user intent and provide personalized recommendations immediately.

The Solution: A Conversational AI Agent

The AI agent combines chat-based interaction with semantic search over Timbal knowledge bases.

1. Conversational Interaction

The agent:

• Greets users and asks clarifying questions

• Detects the destination

• Determines or confirms the user’s language

• Guides the conversation to understand user intent

2. Semantic Search Over Timbal Knowledge Bases

Using natural language understanding, the agent matches user queries to activities stored in Timbal’s KB:

• “I want something cultural and short”

• “Boat tour at sunset”

• “Activities for kids near the city center”

The AI finds the best matches even when the user’s wording doesn’t exactly match catalog entries.

3. Personalized Recommendations

By combining the destination, language, and semantic search results, the agent delivers:

• Relevant activities in the correct city

• Options in the user’s language

• Suggestions that fit the user’s intent

All within the chat interface, no browsing or filtering required.

4. Conversational Customer Service on the Website

Users can:

• Ask questions in natural language

• Receive tailored recommendations instantly

• Clarify details about schedules, duration, or types of experiences

• Proceed to book the selected activity

The entire experience is interactive, guided, and conversational.

Results

After deployment, Civitatis achieved:

• Faster, more intuitive activity discovery

• Higher personalization of user interactions

• Reduced load on customer support

• Better engagement with activity recommendations

• More effective use of Timbal knowledge bases through conversational AI

Why It Matters

In travel platforms, users often know what they feel like doing, but not which activity fits. By combining:

• Conversational AI

• Semantic search

• Timbal knowledge bases

• Destination and language personalization

Civitatis created an agent that transforms questions into curated travel experiences, improving user satisfaction and support efficiency.

Next Phase

Future improvements will include:

• Adding proactive suggestions based on behavior

• Integrating bookings directly into the chat

• Enhancing understanding of complex user intents

The long-term vision: a personal travel concierge for every website visitor, powered by Timbal and AI.

“Instead of browsing endlessly, users now chat with an intelligent assistant that understands their needs and instantly finds the right activities.”, Product & Support Teams, Civitatis