As AI continues to advance, cellular experiences are evolving quickly past easy command execution. A brand new paradigm of agentic AI is rising, one which understands intent and context to autonomously take motion for customers.
Amid this shift, Samsung Electronics has positioned Bixby on the forefront. With its official launch on March 31, Bixby has developed from a voice assistant right into a “device agent” — able to understanding gadget context, connecting features and executing complicated duties on customers’ behalf. With intuitive pure language management, Bixby offers customized options primarily based on gadget standing, together with seamless entry to web-based info inside a single conversational circulate.
So what goes into making Bixby greater than only a voice assistant? Jisun Park, Corporate Executive Vice President and Head of Language AI Team at Samsung Electronics’ Mobile eXperience (MX) Business breaks it down.
Q. What has modified with the brand new Bixby in comparison with earlier than?
Bixby has developed right into a extra highly effective gadget agent, going past a standard assistant. Optimized for every consumer’s gadget, it deeply understands gadget standing and capabilities to offer extra related responses and tailor-made options. With enhanced pure language understanding, it additionally permits extra intuitive and seamless gadget management.
Q. What are a few of the key experiences customers can anticipate from the brand new Bixby?
The most noticeable enchancment is how intuitive gadget management has grow to be.
Bixby understands consumer intent and recommends probably the most applicable settings or options, eliminating the necessity to navigate menus or know precise function names. Users can merely describe what they need in pure language.
For instance, if a consumer says, “Make my screen visible only to me,” Bixby prompts the Privacy Display function.
Bixby may reply questions concerning the gadget and supply customized options primarily based on present settings — basically a service middle in your pocket. For instance, asking “My eyes are tired — how can I make the screen easier to look at?” will immediate Bixby to suggest and activate the Eye consolation defend function proper then and there.
Users can get solutions and options just by asking questions throughout a dialog, with no need to go looking via settings or open separate apps similar to a browser or maps.
In addition, Bixby is now not restricted to device-related queries. It now can analyze real-time net info and supply related solutions. For instance, customers can ask, “Recommend three Korean restaurants in Seoul for a family of four,” and obtain outcomes straight inside the dialog.
This permits customers to ask follow-up questions naturally and get the data they want with out interrupting their circulate or switching contexts.
Q. What was probably the most difficult a part of the Bixby replace course of?
The largest effort went into redesigning Bixby’s structure from command-based to agentic, enabling it to higher perceive consumer intent and ship optimum outcomes.
Previously, Bixby labeled consumer enter and executed duties primarily based on preset situations. Now, with an LLM at its core, it may well interpret intent extra flexibly and generate its personal execution plans.
More particularly, we remodeled particular person features into callable brokers and outlined them in a means that permits the LLM to invoke them as wanted. This permits the system to mix a number of features and APIs to finish duties extra meaningfully, going past easy pure language understanding.
As a outcome, Bixby now handles complicated, multi-step requests extra naturally with larger contextual consciousness, together with situations that had been beforehand tough to course of.
