Trust & AI policy
ContractVector is designed to help small businesses understand, prepare, and make better federal-market decisions—without inventing evidence or taking control away from the customer.
THE OPERATING PROMISE
MINIMUM NECESSARY CONTEXT
If customer-facing AI is activated, ContractVector may send the customer’s question, relevant saved profile and project facts, and reviewed official-source guidance. It will not send the entire account when a smaller context is sufficient.
NO TRAINING BY CONTRACTVECTOR
ContractVector will not use customer questions, profiles, projects, documents, or generated answers to train, fine-tune, or test AI models. We will not voluntarily opt customer content into a provider’s model-training program without separate, explicit customer consent.
EVIDENCE BEFORE CONFIDENCE
Guidance must distinguish saved facts, official-source information, assumptions, and missing evidence. ContractVector must not invent past performance, certifications, eligibility, capacity, requirements, or award probability.
CUSTOMER CONTROL
The customer reviews the underlying facts and decides whether to rely on a draft or recommendation. ContractVector provides educational and decision-support guidance; official notices, agency records, and authorized officials control.
Human approval boundary
These actions always require the customer or another properly authorized human. Product automation cannot silently cross this boundary.
Submitting a proposal, registration, certification, or other binding representation
Making a payment, purchase, refund, or financial commitment
Contacting an agency, buyer, partner, or other third party
Making final legal, eligibility, certification, or award decisions
Data boundary
The initial ContractVector product is not authorized to receive or process CUI. Customers should use official and appropriately authorized systems for sensitive government or identity information.
Login.gov or SAM.gov passwords and one-time security codes
Banking credentials, payment-card details, or API keys
Social Security numbers or military-service documents
Classified, export-controlled, CUI, CDI, or protected health information
Provider transparency
The prepared integration uses the OpenAI API. OpenAI states that API data is not used to train or improve its models unless the account explicitly opts in. ContractVector’s policy is not to opt customer content in.
ContractVector requests no response storage. OpenAI may still retain customer content in abuse-monitoring logs for up to 30 days by default, subject to its policies and legal or safety exceptions. ContractVector does not claim zero retention.
ContractVector’s prepared usage meter records operational metadata such as status, model, request ID, and token totals. It does not store the customer’s question, AI prompt context, or generated answer in those usage records.
Daily request and token limits, a provider-side budget confirmation, and automatic reviewed-guidance fallback are required before billable AI can run. Reaching a limit must not create uncontrolled spending.
ContractVector has not completed SOC 2 or FedRAMP authorization and must not present itself as authorized for CUI. Local export and deletion controls are prepared; complete identity deletion remains inactive until Clerk reverification and signed webhook synchronization are configured and approved.
This page describes founder-approved operating commitments, not final legal terms. Legal review is still required before customer-facing AI or this policy is published.