The Siri hierarchy of needs

I was reading a Medium post by John Ganotis, entitled, Siri commands that should work on watchOS, but don’t:

Siri on watchOS is useful for handsfree interactions, especially when one command can skip multiple physical steps. But I’m regularly finding commands that should be feasible with Siri and don’t work. I decided to start documenting them here.

First off, this is an interesting read. I’d think it’d have value for the Siri team, though I suspect they already have a list like this, and that their list is much more comprehensive.

I am a huge fan of Siri. I do think Siri, like most useful technologies, suffers from its own Maslow’s hierarchy of needs. We love it when we first start to use it but, over time, we start to become aware of how incredibly useful it could be. As we move up this Siri needs hierarchy, we become more and more focused on the missing detail, ignoring all the working pieces.

I use Siri constantly. There are plenty of fails, but I keep a model in my head of what works well, what elements of Siri I can count on.

Perhaps my biggest Siri dependence: I constantly ask Siri to schedule meetings, calls, and reminders. I use many variants on time, including things like day after tomorrow, a week from Thursday, in an hour, like that. I also use “when I get home” a lot. Though there are occasional glitches, almost always with Siri understanding my words, I find this technology very dependable.

Scheduling is at the core of the Siri hierarchy. There are many more examples at that level. General queries that are keyword driven and domain specific. Asking about the weather, the result of last night’s baseball game, requests to do well identified things like set a timer or to give directions to a specific address.

The higher you go in the hierarchy, the less dependable Siri will be. The more context comes into play, the larger the domain of possibilities, the higher the odds that Siri will either be unfamiliar with a concept, or that Siri will misinterpret your intentions.

The problem is one of expectations. As visions of what could be, should be possible come into view, the more we expect from Siri. At the same time, as our Apple gear grows ever more sophisticated, the set of possible Siri interfaces grows, and grows exponentially.

I can only imagine how hard the Siri team is working on both keeping up with the growing list of demands, and working on more future-proof, general purpose, AI-driven approaches that can move Siri into the future.

Personally, I find it helpful to think about how hard this problem is, to appreciate what we’ve got, to not get sucked into that vortex of expectation.