# AI Transparency Statement

How urvo's AI systems operate, how data is used, and how users are protected

This AI Transparency Statement is published to fulfill disclosure obligations under Article 50–54 of the EU Artificial Intelligence Act (EU AI Act). urvo is a voice AI platform enabling businesses to deploy AI agents that interact with users via natural voice conversations.

## 1. Introduction

This document discloses key facts about how urvo's AI systems operate, how data is used, and how users are protected.

## 2. Provider Identity

- **Organization:** urvo LLC
- **Entity Type:** Delaware Limited Liability Company
- **Website:** [https://urvo.io](https://urvo.io)

## 3. Nature and Purpose of the AI System

urvo provides AI-powered voice agents that automate inbound and outbound calls. These agents perform tasks such as appointment scheduling, lead qualification, data collection, reminders, and customer support triage. The system operates 24/7 with real-time audio generation, capable of supporting multiple languages.

The purpose of this system is to automate repetitive call-center tasks, improve operational efficiency, and reduce wait times for customers.

## 4. AI System Functionality

- Powered by proprietary orchestration and models from providers including OpenAI, ElevenLabs, Anthropic (Claude), and Google (Gemini)
- Responds to voice input using real-time text-to-speech and ASR (automatic speech recognition)
- Supports multilingual interactions
- Integrated with telephony via Twilio and SIP trunking
- Uses fallback logic, sentiment detection, and escalation triggers to transfer calls to a human when appropriate

## 5. Human Interaction Disclosure

Users interacting with urvo's system may hear a synthetic voice. This voice is not a human operator. We recommend that the system disclose this at the start of every interaction with a statement such as:

> "Hello, this is an AI assistant from [company name]. I'm not a human but I'm here to help you."

This satisfies EU AI Act Article 50, which requires transparency of machine interactions.

## 6. Data Categories and Processing

urvo processes the following types of personal data in accordance with applicable data protection laws:

| Category | Examples |
|---|---|
| **Audio & Voice Data** | Call recordings, speech-to-text transcripts |
| **Identifiers** | Phone number, caller ID, user metadata |
| **Usage Data** | Response times, interaction logs, sentiment signals |
| **Optional CRM Data** | Names, appointment preferences, customer status (if integrated) |

Data is used only for providing and monitoring the AI system. Retention periods depend on your agent's configuration settings. You can disable recordings or set the conversations retention period to 0 in your agent's **Configure → Advanced** section.

## 7. Subprocessors & Hosting

urvo uses the following subprocessors to deliver its service:

- **Supabase** — Backend infrastructure, database, and storage
- **OpenAI** — Large language models (LLM)
- **Anthropic (Claude)** — Large language models (LLM)
- **Google (Gemini)** — Large language models (LLM)
- **ElevenLabs** — Text-to-speech
- **Twilio** — Telephony infrastructure

For a full list of service providers, see our [Privacy Policy](/privacy-policy).

## 8. Model Limitations & Risk Information

While urvo aims for high accuracy and low latency, limitations may exist:

- Responses may be incomplete or misleading in rare cases ("hallucinations")
- Voice agents cannot provide legal, financial, or medical advice
- Unexpected behavior may occur with non-standard speech, sarcasm, or noisy environments
- Misclassification of tone or intent may lead to incorrect routing

### Mitigations

- Rule-based fallbacks
- Call handoff to human agent
- Testing under adversarial conditions

## 9. Consent and Legal Compliance for Outbound AI Calls

Outbound AI calls are only placed in jurisdictions where the customer has confirmed user consent is collected, per applicable data protection and national telecom laws. urvo provides guidance for opt-out management and consent requirements.

Example disclosure at call start:

> "This is an AI-powered call from [company name]. If you prefer to speak with a human, please say so."

For more details on consent requirements, see our [TCPA Consent Guide](/tcpa-consent) and [Recording Consent Guide](/recording-consent).

## 10. Rights of Individuals

Data subjects interacting with urvo's AI systems can exercise their rights under applicable data protection laws (including GDPR), including:

- Right to access data
- Right to erasure
- Right to restrict or object to processing
- Right to lodge a complaint with a supervisory authority

Requests should be directed to: support@urvo.io

## 11. Updates and Versioning

This statement is reviewed and updated regularly. The current version (v1.0) is effective from March 1, 2026.

Changes will be published at: [https://urvo.io/ai-transparency](https://urvo.io/ai-transparency)