A Tale of Two (or more) Assistants

Reading the comments on just about any post about voice assistants these days will quickly reveal a variety of opinions on each. A lot of people are quick to point to the Google Home or Amazon Echo devices and say how far ahead their assistant devices are compared to something like Apple’s Siri. This is a slightly unfair comparison, since Google Home and Amazon’s Echo (Alexa) assistants are using purpose-built hardware with an array of mics and always-plugged-in power. Line-powered mics will immediately give any assistant a major advantage in that it can be listening much better than on any battery-powered device like a phone or watch. Of course, this really only affects how the device hears a user and interprets their voice. That is only a small (but very important) part of the overall assistant experience.

With Apple’s HomePod soon entering the market, we should have at least a more level playing field in terms of mic hardware to compare against. I expect we’ll see a reduction in the misinterpretation of commands with this, but as many are quick to point out, Siri still falls short of other assistants in terms of overall functionality. Essentially, the story goes that Google Home is way further ahead than the rest (which is not entirely false, but the gap may be perceived bigger than it really is), with Alexa and Siri being matched in many cases. This unfortunately is a bit of an over simplification, as there are plenty of things Alexa cannot do given the platform is limited to voice (no screen). People are quick to point out the vast array of third-party support Alexa has for “skills” and claim this is a huge advantage over Siri, but using any of these skills it’s pretty easy to see most of them aren’t very practical and serve mostly as a gimmick to demo Alexa. While it is a leg up over Siri, their API is very limited and won’t be revolutionizing the voice assistant industry any time soon (largely due to the complex nature of triggering these skills, lacking a good natural language parser).

Google Home has the obvious advantage of being connected to Google’s enormous stockpile of user data. If it weren’t the leader in this space, there would be reason for concern. This means it tends to understand commands better and can do a decent job at returning search results and trivia (even if they aren’t necessarily accurate due to the nature of Google’s popularity algorithm).

Amazon’s biggest advantage (and let’s be honest, the main purpose of the Echo) is the ability to sell you stuff. While it can handle minor trivia, it usually falls short. It has many of the other “standard” functions found in all modern assistants like timers, weather, news and calendars, but overall its skill set is pretty limited when you venture too far from the shopping aspect. Its command system is very rigid compared to that of Siri or Google, meaning you can’t use natural language to trigger things (well, you sort of can, but each app has to do its own parsing, which is rather inconsistent, not to mention triggers are still often unnatural). It has home automation, but it requires very precise wording to get even a limited set of functionality.

Siri is the oldest assistant in terms of software, but for a variety of reasons can be limited in some of the spaces Google or Alexa are good at. It can search the web, but may not always return a direct answer, rather a search result. I would argue this is Apple’s typical way of erring on the side of caution, ensuring they don’t just assume the correct answer is the top result like Google does (and has often proven to be wrong, popular doesn’t mean accurate). It can’t shop for things, but that’s something they would need third-parties for (and I doubt Amazon is going to opt-in any time soon).

On the grand scale, Siri has been able to do a lot of things for a while, with new features added all of the time, but many of these functions are now common place amongst assistants. Where I think Apple has the advantage with Siri is hardware. While the others have user data to benefit from and produce software results, Apple has been busy laying down a foundation of hardware that provides access to Siri. Whether Apple Watch, iPhone, iPad, AirPods, Apple TV, Mac or HomePod (did I miss any?), there are a variety of ways and places you can access Siri. Whether you’re at home (HomePod), in the car (CarPlay), or walking around (Apple Watch/AirPods), Siri is always there ready to go. The other assistants are largely tethered to their assistant-in-a-can that stays home.

We’re starting to see Alexa added to portable devices, but most of them are still pretty tethered. Google Assistant is of course available on the phone, Android Wear watches and the Pixel Buds, but the reviews for all but the phone have been pretty bad. This is where Apple has the advantage. They know hardware, they’ve been great at producing this stuff for a while now, and on top of that, can make Siri very accessible on the go. Siri may not be the most capable right now, but what good is any assistant if it’s not available when you need it? All it would take for Apple is to plop in a new “Siri 2.0” and suddenly they have a plethora of hardware that people already have and can use with it. Similar to their release of ARkit, overnight you would have a ton of users with access to this new assistant. The competition can make great AI, but what good is that if you can’t use it?