Published on | Prototyping: From UX to Front End — Medium Padraig Mannion credit: Dave McElfatrick http://explosm.net/comics/4548/

Siri is a perfect example of how the promises of consumer AI can be massively overblown and the delivered product can be extremely disappointing. There are two ways that I think about Siri – it is a way of controlling your phone using natural language and an annoying product that pretends to be able to do anything but is really very limited. When trying to understand the roles, goals & actions of Siri I think it’s fair to say it has a bit of an identity crisis.

I ran a small study last year to get an understanding of how people interact with Siri. There was a common trend amongst the users

initially they were confused about what they could do
then they’d figure out something small they could do & get excited
then try something fancier
fail miserably & give up on Siri

this is not using real data — but you get the gist

This pattern might be familiar to iOS users. Can you remember the first time you played with Siri? How about the last time you used it to solve a real world problem? Reports also show that Siri usage has dropped year on year so even people who gave it a good shot are tapping out.

However this week I’ve had the pleasure of experiencing what future generations might embrace, a phone that is purely driven by voice commands. This is not a new device… my phone’s touchscreen just stopped working. After furiously tapping it for two days I realised that I could access some features through Siri so for the last week I’ve exclusively interacted with my phone via my own intelligent personal agent 😎 What have I learned from this? Siri is not intelligent, but we haven’t really given it much of a chance to succeed.

The basic expectation for intelligent agents is that we can give them complex challenges and they will be able to extrapolate a number of steps to perform and complete them all accurately. Siri usually makes a complete mess of this.

Let’s not write off Siri just yet though because it can:

open apps
send messages
read messages
view call/missed call list
play/pause podcasts
& call people

but that’s kind of it… These features are useful but it’s infinitely easier to do this with a touch interface (😪). However what if there is no touch interface?

Apple will be releasing Apple Home soon and the word on the rumour mill is that they intend to release an AR headset before 2020 that will heavily leverage Siri. So how can we make it better?

Right now Siri starts in the context of “I’ve no idea what you want, so ask me anything”. It can do anything (not much) at any stage but once you start down a specific path and define the context it does not improve. The flexibility of the GUI allows us to clearly define context and to update the GUI for each context but Siri completely lacks this. Conversational UI has taught us to remove ambiguity by suggesting a limited set of options but that hasn’t really allowed us to progress or change context, it just asks for further specificity.

What this looks like in Siri is yet to be defined. I believe that it will largely be influenced by the Home & Apple AR but it should equally make it’s way back onto the phone/tv/etc…

While I am massively sceptical about voice as a UI, I am given hope in apple’s recent hardware investments but also in what I see in the younger generations- primarily my kids who try to use Siri as a way to get past my lock screen to watch cartoons. If we don’t design a better Siri for us, then lets at least do it for them…

How I broke my phone and learned to love Siri was originally published in Prototypr on Medium, where people are continuing the conversation by highlighting and responding to this story.

How I broke my phone and learned to love Siri

Cool stuff, every week

More Articles