Legends Tours organizes guided city walks in Bruges, Ghent, Antwerp, and Brussels. Their tours are all about local storytelling, authenticity, and connection.
Every guide has their own story to tell, and every city reveals a different side of Belgium.
Direct bookings for Legends through the website were already strong. The real challenge was not to get more visitors, but to get them to choose private tours instead of free ones.
Free tours attract a wide audience, but private tours generate more value for both the guide and the company.
Legends wanted a simple, human way to help visitors discover which experience fits them best. A guided flow that naturally highlights the value of private tours without feeling like a sales push.
We built a guided selling tool that helps every visitor find their ideal city tour.
It asks a few simple questions, understands what they’re looking for, and instantly recommends the best match - complete with a personalized itinerary and email follow-up.
Behind that smooth experience is an AI-driven workflow that filters over 100 tours, balances free and private options, and learns from every interaction.
For Legends Tours, it means less noise, more relevance, and a booking experience that feels effortless from start to finish.
Most travelers don't realize how different a private tour is from a free one.
Private tours offer flexibility, exclusivity, and deeper storytelling, but those advantages easily get lost in a long list of tour options.
By turning the decision process into a guided conversation, Legends helps visitors reach their ideal match faster.
The new configurator asks a few questions, learns what people value most, and instantly recommends a tour that fits.
Built with n8n, OpenAI, Fillout and Loops, the configurator combines automation, AI reasoning, and structured data to deliver personalized tour suggestions.
Hosted on a self-managed Hetzner stack with Docker and Caddy, it’s fast, multilingual, and fully scalable for future features.
It feels like a friendly local giving advice - not an algorithm trying to sell.
The result is clear: less confusion, more clarity, and a growing share of private tour bookings.
Behind the scenes, the configurator runs on a custom-built AI workflow in n8n.
Each user input is sent through a prompt engine that crafts a context-aware request for the AI agent. The agent then matches preferences with the right tours using a structured output parser and real-time data retrieval.
From there, the flow compiles a personalized itinerary and automatically triggers a transactional email with the proposal.
It’s a seamless mix of prompt engineering, automation, and smart data handling – built for scalability, reliability, and multilingual logic.
Hey, you know us. We're quite the creative nerds. In the 2% chance you'd be interested in the tools we used, here's the breakdown.
And if we're getting real deep?
A configurator that doesn’t just ask questions - it helps people make the right choice for their preferences.
For Legends Tours, this means:
Each interaction with the configurator becomes a small feedback loop, helping Legends understand travelers better and refine their offer season after season.
