Today, Samsung has officially announced its own take on the virtual assistant, and it’s calling it Bixby. The company believes that a new intelligent interface on our devices will change the way we interact with our phones placing it as a major differentiator from other voice agents or assistants in the market like Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana and Google’s Google Assistant.

Samsung has employed artificial intelligence, reinforcing deep learning concepts to the core of their user interface designs. Bixby will first appear on the Galaxy S8, expected to be announced later this month. The phone will have a dedicated button to launch the assistant, but Samsung says that its initial capabilities will be limited to a handful of preinstalled apps, with others added over time.

The new Virtual Assistant offers a deeper experience thanks to proficiency in these three key features:

1. Completeness

When an application becomes Bixby-enabled, Bixby will be able to support almost every  task that the application is capable of performing using the conventional interface (ie. touch commands).  Most existing agents currently support only a few selected tasks for an application and therefore confuse users about what works or what doesn’t work by voice command. The completeness property of the new Virtual Assistant will simplify user education on the capability of the agent, making the behaviors of the agent much more predictable.


2. Context Awareness

When using a Bixby-enabled application, users will be able to call upon Bixby at any time and it will understand the current context and state of the application and will allow users to carry out the current work-in-progress continuously.  Bixby will allow users to weave various modes of interactions including touch or voice at any context of the application, whichever they feel is most comfortable and intuitive.  Most existing agents completely dictate the interaction modality and, when switching among the modes, may either start the entire task over again, losing all the work in progress, or simply not understand the user’s intention.


3. Cognitive Tolerance

When the number of supported voice commands gets larger, most users are cognitively challenged to remember the exact form of the voice commands. Most agents require users to state the exact commands in a set of fixed forms. Bixby will be smart enough to understand commands with incomplete information and execute the commanded task to the best of its knowledge, and then will prompt users to provide more information and take the execution of the task in piecemeal. This makes the interface much more natural and easier to use.

Samsung knows it can’t compete with the big guys like Google, Microsoft, Amazon, and others when it comes to raw machine learning power and putting vast amounts of information at your fingertips, so it’s using Bixby to solve a simpler task that those companies have largely ignored. The company argues that its deeper integration of Bixby within apps will make it more useful and more reliable than prior efforts.

Still confused about how the whole Bixby thing will work? We’ll find in a matter of weeks when the Galaxy S8 is announced.