Here's the problem with your example: it's one example. AI assistants don't have general intelligence. All AI assistants currently enable features with specifically developed integration. Google has a flight booking service, and so they've integrated that into their voice assistant. That's really cool, but it doesn't mean Google is fundamentally better than Siri.
There are three fundamental things to look at when evaluating a voice assistant:
- Syntactic understanding: how well does it hear you? Does it accurately transcribe your voice commands into words?
- Semantic understanding: how well does it understand what you're asking for? Is it capable of understanding your intention regardless of how you phrased the command, or does it require rigidly defined commands?
- Capability: what types of tasks can it do? Does it have access to lots of useful data and integrations?
Points 1 and 2 can blur a bit but this is, in my opinion, the fundamental framework that ought to be used for evaluating these assistants. Your example with the flights only goes to point 3 and it is, as I said, a very limited analysis; maybe Siri is missing flight information but has a bunch of other useful integrations that Google lacks.
Now I happen to believe that Google has an advantage in points 1 and 2. 3 is unclear to me. On the whole I would say Google is better. But the flight booking example says little on its own.