Should Tourism Boards Build Their Own AI?

Tourism Boards face a watershed decision: build their own artificial intelligence to serve hotels, agencies and tour operators, or leave it to third parties? According to a Skift analysis, three models are already emerging, and the strategy each destination chooses will define its relevance for years to come.

• Model 1: The generic conversational assistant. A chatbot with public destination data. It works for basic info but falls short when a hotelier asks about real-time occupancy or an agent needs dynamic package margins.
• Model 2: The open data hub with AI. The DMO releases APIs and structured data (events, prices, availability) so operators build their own tools. More power, but it requires stakeholders to have developers or budget.
• Model 3: Tailored AI for each profile. The tourism office develops specialized assistants: one for hoteliers (revenue analysis, demand alerts), another for agencies (live product recommendations), another for tour operators (route and cost optimization). The most ambitious and resource-heavy.

My take: Tourism Boards that wait for a perfect product will fall behind. Model 2 (open data hub) is the sweet spot for most: it doesn't block private innovation and puts the destination at the center of the data ecosystem. The worst a DMO can do is build a generic chatbot nobody uses. If they go AI, make it purposeful and profile-specific.