A new patent on the United States Patent and Trademark Office website reveals that Google Assistant may soon act as an “intermediary” between you and “3P” or third-party services on your Assistant-enabled smart speakers, displays, and more. Google Assistant can perform over a million actions utilizing 3P services that have programmed interactions with Google’s platform, but the existence of this patent may reveal that most people probably aren’t utilizing these enough – I’ll bet it’s because they don’t know they exist and because the human brain can’t actively decide which 3P service would be best suited to complete a task or provide the right information to them at their time of need.
Outside of a few die-hard Google fans, no one is going to browse the Assistant Explore website and commit even a handful of the million actions there to memory and then recall them on their own when needed. Usually, people just remember the “hot words” for popular services used to buy movie tickets, order food, and so on.
These implementations may enable a user to engage a 3P agent via interaction with an automated assistant (Google Assistant), without necessitating the user to know “hot words” to explicitly trigger the 3P agent and/or without necessitating the user even initially know the 3P agent exists.parenthetical text added for clarity
Basically, Google Assistant will decide for the user, based on their request, which service can complete it and connect them to that service automatically. The Assistant may also be looking to lend its voice to 3rd party services to make the interaction feel more natural. More accurately, third-party services with Assistant integration will receive a user’s response from Google Assistant, generate an appropriate response based on the request, and send that data back to Google Assistant so that it can then give the answer to the user in its own voice. The patent calls this “dynamic dialog”.
The 3P agent is configured to receive (…)content from the automated assistant. In response to receiving the content, the 3P agent generates 3P responsive content based on the received content and transmits the 3P responsive content for the provision of output that is based on the 3P responsive content.A fancy way of saying it’s going to let Google Assistant tell you what it said
Let’s say for example that you want to book movie tickets, hail an Uber, order takeout, or do anything that these third-party services are capable of – Assistant may soon be able to do it all in its own voice with their help. This should make the experience of tapping into large storehouses of third-party interactions feel less jarring. Normally, you’d have to say “let me talk to (third party service name)”, wait for it to connect you, and then interact with it, receiving answers in a different voice based on what the developers have chosen.
These services are also not always programmed to handle fallback intents with the same quality as Google Assistant. This means that if you say something that the third-party voice assistant or service didn’t expect you to say and has no response to it, it often just breaks or quits instead of reverting back to saying something natural like “sorry, I didn’t understand, try that again” or “I don’t know how to help with that, but you can do this or that instead”. Routing all of this third-party data through Google Assistant as an intermediary not only gives it the appearance of being much smarter than it is on its own, but it also cleans up a lot of these issues and provides a better experience to the user with one and only one voice assistant. It’s a win-win for everyone.
According to the patent, for each request you make to Google Assistant, it can dynamically connect you to a different third-party service instantaneously and behind the scenes if it determines that the one you first interacted with can not do what you ask of it. So, instead of having to quit your interaction with a third-party integration and start an interaction with another, Google will rapidly and automatically do this on your behalf as well. The user won’t notice, of course, because everything will be done with the normal Assistant voice and without a delay! The whole things – start to finish – will more than likely feel like a conversation with a human when and if this eventually rolls out, especially when paired with the ‘continuous conversation’ feature.
With Google Assistant’s voice and the data of over a million third-party Assistant-enabled interactive experiences, your Assistant across all of your devices may soon feel a lot more human in the way it handles conversation with you. While some may say by quickly glancing at the patent photos that we’re looking at Google Duplex integration, I don’t think that’s the case here – not yet. I definitely see Duplex coming to your homes eventually, but this is something else entirely. It’s exciting, it’s innovative, it’s clever, and takes the whole “make Google do it” slogan to a new level. It is, of course, important to keep in mind that just because Google has filed a patent outlining this technology, that doesn’t mean that it will come to market. Even still, it could be very interesting and useful and I can see this making its way all the way through to implementation.