AI and Automated Systems Disclosure

What it does: When your AI agent initiates a transaction, our policy engine automatically evaluates it against the spending policies configured for that agent. The current wallet rail checks agent status, available budgets, per-payment caps, velocity limits, recipient and network support, and approval rules. Merchant and MCC controls apply only to a future card rail if card services are activated.

How decisions are made: The policy engine uses deterministic, rules-based logic. It applies the exact parameters you set — not predictive models, machine learning, or behavioral profiling. If a wallet request satisfies its hard rules and does not require human review, it may proceed. If it meets an approval rule or threshold, it is held for manual review. If it exceeds a budget or payment cap, violates velocity limits, or targets an unsupported recipient, chain, or asset, it is declined.

What data it uses: Transaction amount, recipient address, chain, asset, agent status, spending history against configured limits, and policy settings. The current wallet path does not use merchant category codes. The engine does not use personal demographics, credit history, or unrelated profile data.

How to get a human review: If you believe a transaction was incorrectly handled, contact us at [email protected]. You can also adjust your spending policies at any time through the dashboard.

What it does: The Reinx Assistant is an AI-powered conversational assistant available in the dashboard. It can read permitted agent, transaction, balance, approval, and policy data to answer questions and navigate you to the relevant workflow. It cannot create agents or change configuration.

Important: When you interact with the copilot, you are communicating with an artificial intelligence system, not a human. The copilot does not provide financial, legal, or investment advice. Responses are generated by AI and may not always be accurate. The copilot cannot approve or deny transactions, change your spending policies, or take financial actions on your behalf.

What data it accesses: The copilot can access your dashboard data (agent information, transaction history, spending policies) to answer your questions. It uses the same permissions as your logged-in session — it cannot see data you don't have access to.

Third-party AI providers: The Assistant's responses are currently generated by OpenAImodels in the GPT-5 family. A Google Gemini integration is retained as an inactive configuration option, but it is not part of the current generation path. To produce a response, your messages and the dashboard data the Assistant reads on your behalf (such as agent, transaction, and spending information) are sent to OpenAI. Inputs are normalized and gated locally. The generated response is then sent to OpenAI's content-moderation service before display. The inactive Gemini configuration does not receive this data. We run the Assistant on our own self-hosted runtime, so the request is not routed through any third-party copilot cloud. OpenAI processes the request under its own terms. This is the one place Reinx sends your financial context to an external AI provider; it does not happen elsewhere on the platform.

Current status: Agent Health Score is a planned capability. Reinx does not currently calculate or display a score; the dashboard shows a neutral placeholder.

What factors influence the score: If introduced, Reinx expects to use operational signals such as transaction regularity, policy compliance history, funding consistency, and usage patterns. Reinx will disclose the final methodology before activation.

How it is used: No current product decision, eligibility rule, or agent-to-agent trust level relies on an Agent Health Score.