Terms of Service

These Terms govern your use of this website and any evaluation access to the SecureVoice platform (the "Service") operated by SecureVoice Technologies FZ-LLC ("SecureVoice", "we"). By using the Service you accept these Terms. Production use by a financial institution is governed by a separate signed agreement (MSA + DPA) that takes precedence over anything on this page.

SecureVoice is a real-time voice fraud-intervention platform: when a transaction crosses a bank's risk threshold, the platform places an outbound call, verifies identity within guardrails, and executes protective actions such as freezing a card or holding a transfer. The public demonstration on this website replays representative scenarios with synthetic data; the conversation-turn endpoint is the same auditable decision surface used in production deployments.

Names, card digits, case references and amounts shown in the demo are synthetic. The demo is provided "as is" for evaluation. While we work hard to keep it available, we do not warrant uninterrupted or error-free operation of the public demonstration, and availability figures shown on the status endpoint apply to the reference deployment only.

Do not attempt to identify real individuals from demo data, probe or disrupt the Service, submit another person's personal information through our forms, or use automated tooling to spam the pilot intake (we rate-limit and honeypot; abuse is logged). Security research is welcome within our Responsible Disclosure policy on the Security page.

Fraud decisions produced by the platform are operational actions taken on behalf of the contracting bank under its own policies. Nothing on this website constitutes financial, legal, or regulatory advice, and no customer relationship is created by using the demonstration.

The Service uses an AI voice agent for customer contact and, where configured, a language model to phrase replies. No model ever decides a financial outcome. Intent classification, identity verification and every protective action are deterministic server-side code governed by the bank's policy; a language model may only rephrase a line that policy has already approved, and its output is scanned before it is spoken.

A freeze or hold requested by the agent is staged and reversible, and a second actor — the bank's system or a human specialist — commits it. You can ask at any point during a call whether you are speaking to an automated system, and the agent will tell you plainly.

Because our website demonstration runs on synthetic data and is for evaluation only, no automated decision made by it affects any real person. Production terms on this point are governed by the signed agreement with the contracting bank.

The Service, including its guardrail policy format, agent architecture, voice personas and this website, is the property of SecureVoice Technologies FZ-LLC and its licensors. You may reference and link to this site; you may not copy its design, content or code for competing purposes without written permission.

To the maximum extent permitted by law, SecureVoice is not liable for indirect, incidental or consequential damages arising from use of the public website or demonstration. Nothing in these Terms limits liability that cannot be limited by law.

These Terms are governed by the laws of the United Arab Emirates, and the courts of Dubai have exclusive jurisdiction. We may update these Terms; material changes will be announced on this page with a new version and effective date.