Published On: May 15, 2018
Last Updated: March 25th, 2019
If Google I/O 2017 was all about Google’s restructuring effort and the resonance of artificial intelligence brewing in Google Labs for a long time, I/O 2018 is more about humanizing those AI efforts by integrating them into its line of products, particularly Google Assistant, which powers Google Home and is an integral part of a bunch of other Google and non-Google devices.
Google Assistant itself was the successor to Google Now. It enabled Android users, for the first time, to engage in a two-way conversation with their Android devices, something iOS users were already enjoying with SIRI.
An experience beyond devices: Google Assistant
During I/O 2016 while introducing Google Assistant, Pichai call it, “The future of computing as an ambient experience that extends beyond devices.” He was quite right. Two of the very big Google launches at the event: Google Allo, a messaging app with a resident chatbot, and Google Home, a smart speaker to rival Amazon Echo powered by Alexa, were both powered by Google Assistant, an outcome of Google’s continual efforts in natural language processing, deep learning, and text to speech over the years.
Regardless the device, Google Assistant can book tickets, order spaghetti on a Sunday from your favorite restaurant, remind you of a meeting with the boss, height of mount Everest and the temperature today and many days to come. All a person has to say is the hotword: ‘Hey Google’ followed by his command.
Continued Conversations with Google Assistant
To be a great conversation partner the Assistant needs to deeply understand the social dynamics of conversation. For example, let’s be honest it gets a bit annoying to say ‘Hey Google’ every time when user wants to get the Assistant’s attention.
Google realized it shouldn’t be so hard. Therefore, they introduced Continued Conversations to Google Assistant at the I/O 2018 event.
Now a person doesn’t have to say ‘Hey Google’ every time he has to wake the Assistant up. It now understands the context of conversation going on and can figure out pronouns. In the example below, it could understand that the user points to the Warriors when he says ‘they’.
A person can now have a natural back-and-forth conversation with the assistant without having to repeat ‘Hey Google’ for every follow-up requests. Even more helpful, the Assistant can understand when you are talking to it versus when you are talking to somebody else. Continued Conversation had been a top feature. Google would be rolling it out as an update in the coming weeks.
Continued Conversation, however, didn’t solve the problem entirely. The Assistant still couldn’t handle complex commands.
Multiple Actions in a single command
With Multiple Actions, you can ask the Assistant multiple things at once. Like you can ask it to raise temperature of you Nest and tell the temperature outside.
In this image above, Google Assistant could process multiple commands at a time: turn on warrior game and start popcorn machine.
While this sounds really natural to us humans, it’s very difficult for computers to understand. One might think: come on it’s easy just look for the end but it’s not always so straightforward. Even though it looks very similar, the assistant has to break it apart to form the two requests in linguistics. It’s called coordination reduction.
Engaging in a two-way conversation with the Assistant is one thing but engaging in an all-out telephonic conversation is another. Google attempted that successfully during the 2018 event.
Merging Multiple Actions and Continued Conversations to create perfect voice controlled coffee maker
In the fall of 2016, we collaborated with a coffee maker that wanted to create a coffee machine that could acknowledge voice commands via Google Assistant in Google Home. We agreed and in the week passing by, they sent us a couple of their Coffee Machines. This was the first time we were working on Internet of Things at hardware level.
We used a Raspberry Pi to connect the coffee machine with the Google Home. Everything worked fine. Here is what the series of commands looked like on Google Assistant to brew a cup of coffee via voice commands to connected Google Home.
If you look closely, the person had to say OK Google three times just to have a cup of coffee. So, I reimagined the entire setup of the Google Home-connected coffee maker with Multiple Actions and Continued Conversations in picture. Google Home with Multiple Actions could understand more in a single command. Like a person could tell it to brew a cup of coffee and dim the lights in a single command.
Continued Conversation ensured that the user doesn’t have to repeat Ok Google every time he has to activate the Google Home device.
This how I postulated the conversation flow to go.
Google Duplex: Google Assistant would fix an appointment over a phone call
Did you know in the U.S. 60% of small businesses don’t have an online booking system set up? To book an appointment with them, a person has to call them manually and talk to the front desk lady, which can get annoying at times.
Google is testing a feature where its Assistant can talk to a human and fix an appointment. The assistant can actually understand the nuances of conversation. Google has been working on this technology for many years. It’s called Google Duplex. Google Duplex is an outcome of Alphabet’s various investments over the years, particularly in natural language understanding, deep learning, and text-to-speech.
The best part is this that the assistant carries the call in the background without interrupting the user and sends a confirmation notification when the appointment has been taken care of.
Google is working hard to get this technology right. It is a technological breakthrough for businesses as well as users. It’ll save time for people and generate a lot of value for businesses.
Business benefits of Google Duplex
Let’s us assume you sell an app-based on-demand delivery solution to businesses. You receive a lot of leads every day who are interested to buy your product but want to see it in action before making the decision. You have a nice demo to show your solution in action but do not have the staff to call each lead and fix an appointment. This is leading to business losses.
With Google Duplex, you can simply upload the spreadsheet containing your leads for the day and ask it to call each of them and fix appointments. By the end of the day, you’ll have an updated spreadsheet with the time of appointment beside each lead. You can show them the demo while the assistant would be busy fixing appointment for the next set of leads.
Google is bringing IoT closer to AI
With the updated Google Assistant, Google is bringing its AI efforts closer to common people and small businesses as it continues to give its products a human touch. In future, we will see smarter IoT devices, just what Google Home has become. Google Assistant is emerging from a Google product to an entire device-independent platform that will power conversational IoT devices in the future.
The time is not far when we would have Google Assistant enabled ACs, thermostats, Refrigerator and other IoT devices. So what do you think of Google Assistant as the ultimate IoT platform? It’s already on our phones, TVs, Tablets, wearables and speakers.
Together, all our Google Assistant-enabled devices would work in unison and could be controlled from one another. Like I can ask the Assistant on my Pixel to turn on the AC in the living room. Likewise, I can ask the Google Assistant on my AC to ring my Pixel.
A brief overview of a smart home in action
A smart home improves the life of its residents by reducing the efforts required to carry out most tasks.
- Automate daily chores
- Control Smart devices
- Smart suggestions
- Kids friendly