Marcio Cyrillo is the director of marketing at CI&T.
Over a year and a half ago I was beyond excited with Siri and the possibilities Apple could explore with it. I wrote this article and argued that Siri took a more personal approach to digital personal assistants and thus could engage and learn more about the user than ever before.
I even risked saying that the search game could be changed with a service like Siri, as Siri would learn about you in a unique way and then start to anticipate results that were more relevant to you.
I clearly remember one night when Siri surprised me in an unprecedented way. I was used to asking her to wake me up in the morning. I believe it just felt better, psychologically, to ask “wake me up at 7am” than setting up the alarm app manually.
One night I was going to bed unusually late and asked her the same question. To my surprise she added to the default message “don’t worry, I won’t forget it”. Isn’t this the reassuring message that you want to hear when you fear that the alarm might not work and then you might sleep-in due to the short night of sleep? I thought that was an amazing touch, a sign that some layer of intelligence was indeed being added to Siri.
Unfortunately though, that was the only moment of bliss in my relationship with Siri.
I still use the service on occasion, but for the most part, I get annoyed with the Siri’s delay in simply understanding what I am saying. A recording of my voice needs to hit the cloud in order to be translated to text back to me before Siri replies. Also, my accent still sometimes gets in the way, and she didn’t learn even a single thing about me or my personality. Generally, it appears that her abilities only grow in incremental bits whenever Apple decides to do so.
One night in New York I was packing for a trip to Brazil and asked Siri what the weather was like in São Paulo. To my dismay she replied, “Here’s the weather for Brasilia, Brazil, through Wednesday next week.” Well, Brasilia, the capital of Brazil, is 500 miles away from São Paulo, and its weather so distinctive, we could say it’s a different country.
Then I decided to try the new Google conversational search with the same question.
“OK Google, what’s the weather like in São Paulo?” — and I see the words being interpreted by Google as I say them, which is great. An instant later, I hear back:
“Here’s the weather forecast for São Paulo, Brazil, for Sunday night.” Oh, wait, do you know that I am traveling on the morning flight and getting there at night? Also impressive, Google used a web service to correct what it had understood initially from my question to something that made complete sense. Siri usually apologizes for not getting what I said.
“OK Google, how do I get to Avenida Paulista?” And I was given driving directions from the GRU airport to Avenida Paulista.
Google’s best response to Siri comes in the form of the mobile service called “Google Now,” which has an iOS version as well as the obviously more powerful Android one. Both versions can leverage the meaningful conversational search I reference above.
While Apple still does a better job positioning Siri as a personal assistant, Google Now appears to consider context far better than Siri does. For example, a few days ago I was in my living room with the TV on when I checked Google Now on my new Android phone. I was presented a card with a question “Are you watching live TV?” And I was able to let Google listen to a few seconds of the program to present me more information about it.
Funny thing though, when I tried to show this feature to a friend who visited me during the day, it didn’t work. But it continues to work at night. I know that Google can detect if my TV is connected to the same Wifi network as my phone, but is Google Now also considering the fact that I don’t usually watch TV during the day? I think it’s plausible.
Another example of context and learning is that in the morning of my latest flight to Brazil. After several trips, Google Now “knows” that I usually take the train to the airport, so it sent me a notification for when I should leave for JFK, a New York area airport, considering the timetables of NYC’s transit trains close to me. It also updated me about gate changes even before American Airlines. I think this is simply awesome.
Google Now is definitely a great response to Siri. Even that impersonal approach found in previous versions of Google Voice Search has changed with the “OK Google,” “OK Google Now,” or “OK Glass” commands. Google, Glass and other “entities” will become our assistants the way we wanted Siri to be.
I knew Google could catch up and maybe this massive improvement was under way at the time of my first article. But it’s a fact that this example exposes one of the major competitive advantages of Google: their cloud services are evolving much faster than those of its competitors. Also, Google is seriously adding layers of intelligence to most of its services so that it leverages patterns found in people’s behaviors.
Whatever the future of digital assistants may be, it’s clear that the service must be fully context-aware, super responsive, and most importantly, learn about you. If Apple doesn’t empower Siri with a true digital brain, the service will soon become a joke when compared to the significant improvements Google Now is achieving.
I can’t wait to see what comes next when Apple seriously addresses Siri’s shortcomings and Google keeps improving their machine learning algorithms. The layer of intelligence that really made a difference in my experience was in fact an “emotional intelligence.” Like all human-centric relationships, Apple or Google’s machine learning algorithms will continue to make us fall in love with them when they find new ways to connect on a personal and emotional level.
Emotional connections become stronger the more people learn about and get to know one another over time. Perhaps the love affair burns most deeply when they continue to find new ways to surprise us. OK Google Now, you get me.