In May this year, Google outlined its plan for redefining instant messenger apps by showing off a future vision which included Duo and Allo. The former is a simple-to-use video call which, according to the numbers Google shared with us, has crossed 10 million users on Android alone (there is an iOS app as well for the Apple iPhone). This is, perhaps, at least in terms of sheer traction for the two new apps, the perfect time to unleash Allo into a world of instant messenger apps, which already has WhatsApp, Facebook Messenger, WeChat, Google’s own Hangouts and others. The big difference, though, is machine learning.
The process of setting up Allo is very similar to that of setting up WhatsApp. For example, Allo is linked to your phone number, which means you can also use it to send text messages to anyone in your phone book, irrespective of whether or not they are using the Allo app.
However, there is an added advantage—it also connects to your Google account. At present, it seems as though Allo uses your Google account to personalize the information it provides via Google Assistant in chats.
However, there is the potential of multi-device support in the future, allowing users to sync messages across devices and perhaps even a desktop client in the mould of what WhatsApp does.
Read more: Why is Google making so many messaging apps?
On the face of it, Allo looks like just any instant messenger app. But there are three primary elements which make Allo unique—the machine learning-based Google Assistant, something Google refers to as Expression, and of course, security.
Allo uses machine learning and natural language processing, both of which work in the background, to suggest smart replies as you are in a conversation with a friend. It is essentially understanding the latest sent message and anticipating the replies based on context and content. Initially, you may not always get the perfect reply suggestion, but as you converse more and more, Allo will understand your style of conversing and register some specific words that you use and offer enhanced suggestions. Also, these suggestions will be unique to you. For example, if Allo understands over time that you prefer to reply with “yeah” instead of “yes”, that will start up showing in the smart reply options more often.
Also, the algorithms can detect content in photos that a friend may share with you, and based on that, can suggest replies. For example, a photo of the Golden Gate bridge in San Francisco will get suggested replies that include “beautiful”, but a photo of food will instead get “delicious”. These fine differences may not always be noticed, but this sort of smartness makes a difference over time.
The smartness extends beyond smart reply suggestions.
The new Google Assistant, is always watching, and also constantly learning. It integrates with Google’s wide range of services. It can be a part of your chats with another friend, or you can also chat with it like you would with any friend (it can be summoned within a conversation by typing “@google” and then the question or request). It will offer everything from the weather to sports scores, reply to a question on, say, suggestions for an Italian restaurant in your city and even share photos with friends.
You can also ask Assistant to set calendar entries, or even wake you up with an alarm the next morning. However, these deep integration features require access at an operating system level, which is currently only available in Google’s own Android phones and not Apple’s iOS-based iPhones yet. If you are on the latter, some features may remain unavailable.
Then there is the entire expression bit. Text messages are no longer just limited to simple words. Most of us regularly use stickers and emoji. And Allo has both bases covered. Next time, a “Thug Life” sticker will convey your feelings about someone else’s achievements than words probably ever would. You can click a new photograph through Allo, and scribble your message on it and send to a friend (at the time of writing, this option was unavailable in the iOS version, but present in the Android version).
You’ll even see smart reply options when someone sends you a photo. Allo can understand the content and context of photos, thanks to Google’s computer vision capabilities. If someone sends you a photo of pasta, you will see smart replies that include mentions of pasta, yummy, or whatever. The idea here is that assistive technology can help you communicate with little to no effort.
There has been a massive debate recently around the security and privacy of instant messenger apps, after WhatsApp confirmed that Facebook will be able to access users’ WhatsApp account details. The app has an incognito chat mode, which allows the conversation starter to set the time when the messages will be automatically erased from the app of all parties involved in the chat.
At present, Google isn’t offering any sort of message back-up feature in Allo, unlike WhatsApp.
First impressions with the Google Allo app seem to suggest that this is indeed a refreshed take on the entire ecosystem of instant messenger apps. There is smartness under the hood which makes everything seem better.
However, the biggest challenge isn’t with artificial intelligence or machine learning—it really boils down to whether you can convince your friends to shift to another IM app, which would perhaps not be as easy as it sounds.