The author recounts an impressive and terrifying experience with Google's new AI agent, Gemini Spark, describing it as an ambitious, always-on agent. Unlike other AI tools for trip planning, Spark demonstrated a significantly advanced and less generic approach.
A study analyzed 47 pasted trip plans and 989 places, revealing that 37.9% of locations in copied or AI-assisted itineraries had actionable trip-risk issues like closures or booking requirements. Almost all trips (97.9%) had at least one actionable issue, underscoring the need for a validation step.
Amadeus's CTO states that existing protocols, like UCP, are not yet suitable for agentic AI in travel due to the complexity of real-time, perishable inventory. Experiments confirm this, showing current frontier models fail to handle expiring offers and price changes effectively.