Key Takeaways
- G. Scott Paterson believes the next generation of personal AI agents will become more useful by working directly with documents people already own.
- FutureVault provides a digital document management platform for financial institutions and businesses, with more than 4,000 permissioning configurations.
- FUTR brings document-based AI to consumers, allowing users to retrieve information from insurance policies, financial records, contracts, and everyday documents.
- FutureVault can use different language models for different tasks, including reasoning and document summarization, while maintaining document access controls.
- FUTR introduces a token-based rewards system that connects personal data management with potential commercial value when users consent to sharing their information.
G. Scott Paterson is betting that the next generation of personalized digital agents will depend on something most people already have in abundance: documents.
Through FutureVault and FUTR, Paterson is developing systems that connect artificial intelligence with financial records such as insurance policies, tax documents, contracts, and business records. The goal is to make those documents immediately usable while keeping them inside a controlled digital environment with detailed access permissions.
FutureVault serves businesses and financial institutions, giving clients a digital vault that can hold records tied to an individual, a family, a holding company, a charitable foundation, or other related entities. FUTR takes the concept directly to consumers through an app and web-based platform.
Together, the companies point toward a personal AI agent built around financial information the user has stored in one location.
FutureVault Puts AI Inside a Permissioned Data System
FutureVault’s architecture is a multi-tier construct: head office, operations, front office, and end clients. A financial institution can offer the service under its own branding, with documents organized around the customer instead of scattered across separate products and departments.
Someone applying for a car loan may provide tax documents and proof of income, then face another request for similar records when applying for a mortgage. A digital vault gives the institution a central location for those records, subject to the rules governing access and reuse. The same document management approach can help companies organize contracts, financial records, and important business insurance documents.
FutureVault has more than 4,000 permissioning configurations. A financial advisor may be able to access certain documents in a client’s vault, another employee may have a different level of access, and an outside estate planner may only see files relevant to that work. Privacy requirements and internal business rules can be incorporated into those permissions.
This focus on controlled access also reflects the importance of protecting sensitive information when introducing AI into financial services. The National Institute of Standards and Technology (NIST) identifies privacy, security, transparency, and accountability as important considerations for trustworthy AI systems.
AI sits on top of that structure. FutureVault received its first AI-related patent in 2021, before ChatGPT became publicly available. According to FutureVault’s official patent announcement, its patented technology uses machine learning to help identify and organize incoming documents. Its system can also connect stored documents with large language models and use the information inside those files to answer specific questions.
FutureVault can also route different tasks to different language models. G. Scott Paterson said in an appearance on In Good Company with Jay McCauley that the company uses Claude for certain reasoning requests, while Llama has been used for summarization work. That setup allows the platform to select models by task while preserving its own document permissions and access controls.
FUTR Moves the Model Into Everyday Life
FUTR takes the same document-centered approach into a consumer product, with the app designed to hold records that are focused on financial services but extend to everyday matters. Potential files include insurance policies, medical documents, children’s report cards, certifications, investment agreements, payment records, and presentations. G. Scott Paterson’s personal vault contains roughly 20,000 documents.
The value of that archive becomes clear when a user needs a specific detail buried in a file.
A cracked windshield could trigger a request to the FUTR agent about insurance coverage. The system could locate the relevant policy, identify the covered vehicle, retrieve the coverage amount and deductible, and help with the next step with the insurer. This could simplify how users access their insurance policies and coverage details, particularly when they need information quickly.
