Halifax is a Canadian city nestled on the Atlantic Coast with a population of 500,000 and approximately 7,000 hotel rooms. As Nova Scotia’s capital, it is a thriving tourism hub.

Our legacy visitor information tools—guides, maps, website, kiosks—have long served as anchors to connect visitors with local experiences. Like most destination organizations, we’re continuously evolving as technology advances.

One area of high demand is the Port of Halifax, which welcomed over 200 cruise ships in 2024, carrying more than 350,000 passengers. As traffic increases, so does the need for greeters and printed materials.

To enhance the experience for cruise passengers as they step off the ship, Discover Halifax created a custom, AI-powered itinerary planner and an updated information centre. The new space bridges the transition between the cruise environment and the vibrant city beyond. Developed with our partners at National, the tool was designed for visitors who tend to be older, less tech-savvy, and often first-time explorers.

The user interface features familiar elements—event listings, maps, and insights into our Mi’kmaw and African Nova Scotian communities. The innovative element is an AI-driven recommendation engine, which generates tailored itineraries based on real-time factors.

The system “knows” its location, time of day, day of the week, season, and current weather conditions. Visitors answer a few questions—how many people are traveling, their interests, the time available, and their preferred pace—and instantly receive a customized plan. The tool queries our website and CRM content to deliver options aligned with visitor preferences and logistical realities, such as business hours and weather conditions. 

Plans can be reviewed on-screen or emailed to guests via complimentary Wi-Fi. Future updates will include options to print or text directions.

As a proof of concept, the project has been both rewarding and instructive. As with many AI tools, occasional “hallucinations” happen. For instance, an out-of-the-way convenience store offering coffee appeared as a café alternative. We’re working to refine accuracy to eliminate such anomalies by mid-year.

Despite these quirks, early results are promising. With a small team and budget, we demonstrated feasibility and gathered invaluable insights. We’re now scoping how this solution can extend to hotels, meeting venues, and potentially a mobile version for independent travelers.

Importantly, this AI tool is distinct from our website chatbot, which was developed by Guide Geek. Because it is purpose-built for a specific location and duration, it has a highly controlled prompt structure. Strict geo-fencing and focused content keep recommendations safe and relevant. 

Simplicity has been central to our approach. Our goal is to surprise and delight visitors with personalized recommendations that enrich their time in Halifax.

Four lessons we’d share:

• Experiment boldly: Out-of-the-box tech for every scenario isn’t here yet, but the building blocks are.
• Control variables early: Start simple. For cruises, we focused on attractions over dining and omitted hotels to keep things manageable.
• Embrace iteration: No one has fully cracked this yet. Be comfortable with “good enough” and plan for improvements.
• Observe real use: We monitor interactions onsite and analyze data to uncover and address gaps.

This project solved a challenge that was nearly impossible before AI. We’re proud to have delivered a helpful, innovative tool that supports our visitors and showcases Halifax’s spirit of hospitality.