25 05 2013
Google’s personal assistant
Has Google surpassed Apple’s Siri?
I predicted in The Software Society that the “personal assistant model” would be the dominant way we interacted with digital systems and services in the future. I posited that it could even become a consistent ubiquitous interface available on devices from smartphones to PCs to TVs. Eventually, you won’t buy a new device unless it supports your assistant.
The key feature of the personal assistant model was that it accepted natural language—we could just talk or type to it without reference to a manual. In addition, the assistant would attempt to give us an answer rather than list of web sites (“direct-to-content” search).
The personal assistant would indeed be “personal.” It would remember what it previously did with a particular user and have access to personal information such as contacts. It would integrate with apps such as calendars or music players so that it could be asked, for example, to remind us of an event later or play a particular song. It would help us with features on the specific devices, such as texting messages.
Announcements at the recent Google developers conference suggest Google web search is moving in this direction. One announced option is to speak to the search engine and have it speak back, at least in the Chrome browser. To invoke search by voice, one wakes up the app by saying, “Okay Google,” so perhaps the name of the personal assistant is simply “Google.”
Voice or text search using natural language is a key feature in Google’s announcement. Direct-to-Content is embodied in Google’s “Knowledge Graph,” displayed in response to a search request in addition to a list of web sites. Through Google Now, Google will even tell you what it thinks you need to know without your asking. Since Google will make search with the new functionality available everywhere through a browser, it is even moving in the direction of being ubiquitous. With natural language interaction, a voice option, and direct answers, is Google search now simply a ubiquitous personal assistant named “Google”? Has it surpassed Apple’s Siri, which won’t even let you type when it isn’t appropriate to speak?
Not quite. The main lack in Google’s offering currently is the “personal” element. Since Google wants to be available on all devices, it doesn’t necessarily know what apps are available on the devices or how to interface with them. It thus lacks this critical function of a full personal assistant. In this dimension, Siri has the advantage of being specific to the Apple device one is using and being able to interface directly with those apps. If Siri becomes ubiquitous, she will most likely be ubiquitous only across Apple products.
But Google’s answer in the long run may be to provide as much functionality as possible with Google apps or services in the network. There is a new Reminder “card” in Google Now, apparently providing a device-independent reminder application. The Google Maps app is popular even on Apple’s iPhone. And if manufacturers of mobile phones that compete with Apple cooperate, Google search may eventually be able to “search” device-specific features on many devices.
Whether Google is providing a version of a “personal assistant” depends on your definition. But whatever the functionality is called, it is migrating toward what I described as a personal assistant model.
And expect Apple to respond. The next generation of Siri will almost certainly have more search (direct-to-content) features. And Apple can exploit a tight integration with the platform and apps so that Siri can help us be more efficient in our use of them.
The ultimate result of this trend toward the personal assistant model will be increasingly powerful natural-language interaction with our devices and a tighter human-computer connection. This is a powerful partnership, as I emphasize in The Software Society.