Google has announced a conversation-based tool to control smartphones, smartwatches and other devices.
Google Assistant can be used to find information, play media and carry out tasks – such as booking cinema tickets – via a back-and-forth chat between the user and the software.
The firm also announced a voice-activated device with a built-in speaker called Google Home to provide the tech in living rooms.
It will compete with Amazon’s Echo.
Amazon launched its own dialogue-based smart home device in 2014, and has since included the Alexa voice assistant that powers Echo in some of its other products.
Google’s chief executive Sundar Pichai credited Amazon with having pioneered the idea.
His announcement was made at the start of the firm’s IO developers’ conference in California.
“Google’s new Assistant is its attempt to bring together a set of disparate efforts that have lacked a coherent brand,” commented Jan Dawson from Jackdaw Research.
“Referring to the combined functions of Google Now, OK Google, and other elements has been tough in the past, because there wasn’t a single name for this functionality.
“This should help Google compete more effectively both with Amazon’s Echo device but also with better-branded personal assistants like Siri, Cortana, and Alexa.”
Google Assistant can link into third-party services including Ticketmaster, Spotify, Uber and Whatsapp. But Alexa already works with many more.
“Amazon should take note of Google Home given Google’s search and artificial intelligence capability, but Alexa has an early lead in third-party integration,” noted Geoff Blaber from the CCS Insight consultancy.
Google Assistant also plays a role in a new chat app called Allo, in which it can make suggestions based on what the user is talking about with their friends.
One executive demonstrated the AI tool proposing a restaurant to visit when two users discussed wanting to eat Italian food.
Allo’s other unusual feature is that it can provide replies on its user’s behalf.
This includes commenting on pictures sent by friends, thanks to its use of image recognition algorithms.
The company acknowledged that the process would not be perfect but said the suggestions should improve the more people used the app.
A partner video chat app called Duo provides a live view of the caller before the recipient decides whether or not to answer.
In the past, Google IO has provided a first look at the next version of Android.
But many details about Android N were released weeks ahead of this year’s event.
However, there were new details about a forthcoming feature how virtual reality would be integrated into the OS via a platform called Daydream.
It introduces a dedicated VR app store and provides a set of specifications that devices must meet in order to provide a lag time of less than 20 milliseconds between a movement of the user’s body and an on-screen response.
Daydream also includes a reference design for a VR headset mount, into which phones can be fitted, and a controller – a handheld device featuring a trackpad and two buttons.
Google is not making either, but hopes that manufactures will put its designs into production.
“One of the worst things about Samsung’s Gear VR [headset] is the control system – Daydream looks much better,” said Mr Dawson.
The firm also announced an upgrade to its smartwatch OS, Android Wear.
It now supports “standalone” apps, meaning the software can link directly to the cloud via wi-fi or a watch’s own 3G/4G chip, rather than having to link up via a smartphone.
This should make apps run faster and addresses a problem that had also plagued apps on the original version of Apple’s rival Watch OS.
In addition, the updated Android Wear provides ways to reply to messages beyond speaking into the device’s microphone.
Users will able to:
select from a choice of “smart reply” suggestionsdraw letters on the screen that then get converted into texttype in words via a small keyboard
How practical this will be is unclear, particularly as an on-stage demo went wrong.
“Android Wear 2.0 is a reset for Google’s wearables platform,” said Mr Blaber.
“Having learned lessons from the first generation, Google and partners will hope 2.0 will kick start wearables adoption and usage.”
Google rounded off IO’s first event with a look at a feature that could shake up one of Android key principles.
Rather than making users install an additional app each time they want to access a new service, the firm wants to make it possible to quickly download only the code needed to carry out a specific task.
“We are evolving Android apps to run instantly without installation,” explained Google executive Ellie Powers.
“We call this Android Instant Apps.
“As a developer… you’ll modularise your app and Google Play will download only the parts that are needed on the fly.”
Examples given included a parking meter app that automatically presents its payment interface when a motorist leaves their car at a stand, and a news app that streams its videos when a user clicks a link.
The idea echoes Scopes found in Ubuntu’s smartphone OS and Microsoft’s App-V facility, which both aim to provide access to third-party services without making users download full apps.
Analysts had mixed feelings about the proposition.
“Android Instant Apps [are the] first real signs of the post-app era,” tweeted Gartner’s Brian Blau.
But Creative Strategies’ Carolina Milanesi wrote: “Is it like a smart preview that will get me to install the apps I don’t have?”
And Moor Insights’ Patrick Moorhead added: “When I think about Android Instant Apps, why is the first thing I think of [a] security panic?”