Liquid AI and Qualcomm Bring Personal AI Context to Snapdragon Devices
Liquid AI and Qualcomm Bring Personal AI Context to Snapdragon Devices,The collaboration gives device manufacturers a foundation for developing more personalized, proactive and context-aware AI experiences. Alongside Liquid Context, OEMs can also evaluate Liquid Agent, Liquid AI’s efficient embedded agent designed to operate on-device.
Bringing Personal Context to AI Agents
For AI agents to provide useful assistance, they need to understand not only what a user is asking, but also the broader context surrounding that request. Liquid Context is designed to provide this persistent layer of understanding.
With user permission, Liquid Context learns from relevant device signals and builds an evolving understanding of routines, preferences and needs. Rather than requiring each AI agent to independently reconstruct this information, Liquid Context provides a shared context layer that can be made available to a user’s selected agents, including third-party agents and Liquid Agent.
The system separates context from action. Liquid Context provides relevant information about the user and their current situation, while the connected agent uses that information to reason, recommend next steps and perform actions with the user’s permission.
Personal context is created and maintained locally, with relevant information shared with connected agents according to the permissions established by the user.
Making Personal AI More Relevant
Ramin Hasani, CEO and co-founder of Liquid AI, said personal AI begins with understanding how people live and what they need at particular moments. By maintaining that understanding directly on the device, It Context can help the agents users choose provide assistance that is more relevant to their circumstances.
The use of the Hexagon NPU is particularly important for continuously maintaining context without requiring every update to be processed by a cloud-based model.
It Context is designed to operate in the background, processing permitted device signals locally and transforming them into useful context. AI agents can then use this information to identify situations where assistance may be useful and respond in ways that better reflect the user’s preferences and priorities.
Examples of Context-Aware AI Experiences
It Context can support a range of everyday scenarios in which persistent personal context could help AI agents provide more useful assistance.
Responding to a family emergency: If a user receives a message that their child is sick and needs to be picked up early from school, Liquid Context can provide an authorized agent with relevant family information and the user’s priorities. The agent could review the user’s calendar, identify meetings that can potentially be moved while preserving important commitments, and prepare rescheduling emails for the user to review and approve.
Creating a personalized conference recap: After attending a conference, It Context could provide an agent with relevant information from the day, along with the user’s preference for creating a LinkedIn recap and their preferred writing style. With permission, the agent could identify appropriate photos, draft a post in the user’s style and prepare it for review before publication.
Connecting activity across devices: A user might leave their phone at home while going for a run using a smartwatch. The watch could locally capture information about the activity, such as a new personal record. When the user returns to their vehicle, the authorized context could be synchronized locally. The vehicle’s agent could then congratulate the user, adjust the cabin environment according to their preferences and suggest a nearby recovery stop.
These scenarios illustrate how persistent context can allow AI agents to connect information across devices and situations while keeping users in control of what information is shared.
Optimized for Snapdragon Platforms
It AI and Qualcomm Technologies are exploring additional opportunities to enhance Liquid Context on Snapdragon platforms. The goal is to make personal context a built-in device capability that OEMs can integrate into their products and make easier for users to activate.
Compatible on-device, cloud-based and hybrid agents could use the context layer to provide more proactive and personalized assistance.
For OEMs developing agentic experiences, the combination of Snapdragon hardware and Liquid Context software provides a foundation for turning device signals into useful information for AI applications. Manufacturers can integrate personal context while retaining flexibility in choosing the agents, services and user experiences that best fit their products.
Durga Malladi, executive vice president and general manager of Technology Planning, Edge Solutions and Data Center at Qualcomm Technologies, Inc., said Qualcomm Technologies is focused on advancing AI across devices, edge solutions and data centers. He highlighted the collaboration with Liquid AI as part of efforts to accelerate the development of personalized and increasingly intelligent agentic AI experiences.
Hexagon NPU Enables Efficient On-Device Intelligence
Continuous context processing and responsive AI agents require efficient computing resources, particularly on battery-powered devices.
It AI has optimized Liquid Context to take advantage of the processing capabilities, memory access and low-latency characteristics of Snapdragon processors. Separately, the company has optimized its LFM2.5-2.6B agentic model, which powers Liquid Agent, for execution on Snapdragon platforms.
The next-generation Qualcomm Hexagon NPU combines scalar, vector and matrix processing with dedicated transformer hardware to accelerate AI workloads.
This hardware-software combination is designed to support continuous context updates and capable on-device agent execution while operating within the power constraints of everyday consumer devices.
Liquid Agent Provides an Embedded AI Option for OEMs
In addition to using Liquid Context with their own preferred AI agents, OEMs can work with Liquid AI to evaluate Liquid Agent, an efficient embedded agent powered by the company’s LFM2.5-2.6B agentic model.
Liquid Agent is designed to operate efficiently on the Hexagon NPU and can be customized for a manufacturer’s hardware, services, user interface and brand identity. It can also consume information provided by Liquid Context to deliver personalized and proactive assistance.
It Context serves as the shared context layer, maintaining relevant understanding and task state with user permission. This enables different agents to access appropriate context and continue tasks without requiring the user to repeatedly provide the same information.
OEMs can therefore use Snapdragon hardware and Liquid Context as a foundation for their own selected agents, while evaluating Liquid Agent when they require an efficient embedded or hybrid AI solution.
Building the Future of Device-Native AI
The collaboration between Liquid AI and Qualcomm Technologies highlights the growing role of on-device intelligence in the development of personalized AI experiences.
By processing context locally, devices can maintain a more continuous understanding of user preferences and activities while reducing reliance on cloud processing for every interaction. Combined with efficient embedded agents, this approach can enable AI experiences that are more responsive, personalized and adaptable across phones, laptops, vehicles, wearables and other connected devices.
About Liquid AI
It AI is a device-native foundation model company focused on bringing advanced AI capabilities to processors outside traditional data-center environments. The company is developing software and foundation models designed to deliver efficient intelligence across a broad range of devices, including smartphones, laptops, vehicles, aircraft, robots and wearables.
It AI’s open-weight Liquid Foundation Models (LFMs) are designed to deliver efficient AI performance directly on devices where latency, privacy and computational efficiency are important. Through its device-native approach, Liquid AI aims to enable advanced AI capabilities across the physical world.
Source Link: https://www.businesswire.com/