In Defense of Voice Assistants

From time to time I hear a lot of complaints from people about voice assistants. Whether it be that they just don’t find them useful or they never work right, every case tends to follow some common themes. Now, I don’t mean to pull a Steve Jobs with the old “you’re holding it wrong” type thing, but I would be willing to bet a lot of users simply aren’t using their assistants correctly in some way.

I’ve been using Siri since the iPhone 4S, and not just as a gimmick but as a genuinely useful extension of my technology experience. When the 4S came out in 2011, Apple also added a new Reminders app to iOS, which happened to coincide with the newly-added Siri, which could set reminders. Before I upgraded to the 4S, I was loving the Reminders app already. It was a super handy way of ensuring I don’t forget things, or having geo-fences trigger when I enter or leave somewhere. The addition of Siri made this already-useful feature immensely more useful, as suddenly I didn’t need to unlock my phone, open Reminders, type it out, adjust the date/time/location, etc. I could just dictate the entire thing to Siri.

Assistants get a lot of flack, but overall if you just learn the interface even a little, it can go a long way. I certainly would love my assistant to do more, but it still makes my life easier in a lot of ways. The bonus is, the more you use it, the more it improves both in language parsing and understanding you. Perhaps not directly, but this data can at least improve the assistant over time by the company knowing how users are interacting and what results they are getting.

Admittedly, while assistants have been useful to me, I’ve been a nerd about keeping up with each new function they’ve added to Siri. Part of the problem with many of these assistants is that it’s not always obvious what they can and can’t do. Much like the frustrations users experience with poor user interface and experience in traditional software GUIs, the same could be said of assistants. Users expect one thing but may get a different result, or nothing at all. This is where truly intelligent assistants will greatly improve things, but we’re not there yet.