An online travel agency is building a comprehensive AI assistant. This assistant must handle three different types of user queries: Personalized Recommendations: "Based on my past trips, suggest a quiet, relaxing beach destination for my anniversary." Live Bookings: "Check if there are any available seats on the 8:00 AM flight to San Francisco next Tuesday." General Advice: "What is the best way to get from the airport to downtown in Rome?" To ensure the answers are accurate and relevant, what is the correct grounding strategy for each of these tasks?