Data Integration and Intelligence in VSA

The Vold Smart Agent (VSA) architecture features a dualistic data approach, segregating user-generated data from application/auto-generated data, enhancing the AI’s analytical capabilities.

On one side, knowingly user-generated dataβ€”search history, emails, identity documents, and wallet transactionsβ€”are collected with user consent. This data is stored in a self-custodial data store, ensuring user sovereignty over their digital identity and activity footprint.

On the other side, the VSA taps into a diverse set of application and auto-generated data sources. By interfacing directly with APIs from platforms, the VSA bypasses traditional app interfaces to glean deeper behavioral insights. This method not only captures explicit user actions within these services but also the implicit data streams such as transaction patterns and social interactions.

The core AI agent utilizes this amalgamated data set for on-device behavior analysis, sidestepping the need for external data processing. This in-situ processing architecture turns unseen actions from closed applications into actionable insights, providing a more nuanced understanding of user needs and preferences. Consequently, the AI can proactively perform tasks or offer recommendations, all while maintaining a strict adherence to privacy and security standards set by the user.

Last updated


Vold 2023