# AI Disclosure Policy
ClientBuddy
A Netherlands-registered company (KvK: 97483672)
Operating agents.miami
Business Activity: Ontwikkelen, produceren en uitgeven van software
Effective Date: February 1, 2026
Last Updated: January 21, 2026
## 1. Purpose
This AI Disclosure Policy explains how agents.miami uses artificial intelligence to provide phone reception services and ensures transparency with all parties who interact with our technology.
## 2. AI Operations Statement
### 2.1 All Calls Handled by AI
agents.miami operates entirely through artificial intelligence. When callers contact restaurants using our Service:
- All calls are answered by an AI voice assistant
- No human agents are employed to handle calls
- The AI processes natural language in real-time
- Reservation details are extracted automatically
### 2.2 No Human Employees in Call Handling
We want to be completely transparent:
- Zero human operators answer calls on behalf of restaurants
- The AI voice you hear is generated by ElevenLabs technology
- Responses are generated by AI language models
- Human supervision exists only for system oversight, not call handling
## 3. Call Recording Disclosure
### 3.1 Florida Two-Party Consent Compliance
Florida law (Fla. Stat. § 934.03) requires consent from all parties to a recorded conversation. agents.miami complies by:
1. Mandatory Disclosure at Call Start:
Every call begins with the following notice:
> "Thank you for calling [Restaurant Name]. This call may be recorded and is handled by an AI-generated voice assistant powered by agents.miami. By continuing, you consent to recording, AI processing, and interaction with synthetic voice technology. This is not a live human agent. Say 'human' at any time to request a callback from the restaurant. agents.miami does not provide live human call handling."
2. Implied Consent Through Continuation:
- Continuing the call after the disclosure constitutes consent
- Callers who do not consent may hang up immediately
- The disclosure is given before any substantive conversation
3. Clear Identification as AI:
- The AI identifies itself as an "AI-generated voice assistant" using "synthetic voice technology"
- The disclosure explicitly states "This is not a live human agent"
- No attempt is made to deceive callers into believing they are speaking to a human
- If asked directly whether it is human, the AI will always truthfully identify itself as AI
### 3.2 Recording Retention
- Call recordings are retained for 180 days
- Recordings may be used for quality assurance and AI improvement
- Callers may request deletion of their recordings
## 4. Automated Decision-Making
### 4.1 What the AI Decides
The AI makes the following automated decisions:
| Decision | Description | Human Review Available |
|----------|-------------|------------------------|
| Reservation Extraction | Identifying date, time, party size from conversation | Yes - via restaurant dashboard |
| SMS Confirmation | Sending booking details to caller | Yes - restaurant can disable |
| Call Routing | Determining if escalation needed | Yes - caller can say "human" |
| Quality Scoring | Evaluating call success | No direct impact on callers |
### 4.2 Limitations of AI Processing
The AI may:
- Misunderstand accents or unclear speech
- Extract incorrect reservation details
- Miss nuanced requests or special accommodations
- Not understand every language or dialect
All reservations should be verified by the restaurant.
### 4.3 Right to Human Review
Callers and restaurants have the right to:
- Request human review of any AI decision
- Override automated reservation details
- Escalate issues through human support channels
- Opt out of AI processing entirely (by canceling service)
- Bias Mitigation: AI training data is regularly audited for biases (e.g., in speech recognition for accents/dialects). Human review includes checks for fairness, with corrective actions documented per EU AI Act Article 50.
## 5. AI Technology Components
### 5.1 Technology Stack
agents.miami uses the following AI components:
| Component | Provider | Function |
|-----------|----------|----------|
| Voice Synthesis | ElevenLabs | Generating natural-sounding AI voice |
| Speech Recognition | ElevenLabs | Converting caller speech to text |
| Language Understanding | Anthropic (Claude) | Processing natural language |
| Reservation Extraction | Proprietary | Identifying booking details |
### 5.2 Continuous Learning
Our AI systems improve over time by:
- Analyzing call patterns and outcomes
- Updating speech recognition models
- Refining reservation extraction accuracy
- Learning from quality evaluations
Note: Individual caller data is aggregated and anonymized for training purposes where feasible.
## 6. Caller Rights Regarding AI
### 6.1 Right to Know
Callers have the right to know:
- That they are speaking with an AI, not a human
- That their call is being recorded
- How their data will be used
This disclosure is provided at the start of every call.
### 6.2 Right to Human Escalation
Callers may request human assistance at any time by:
- Saying "human" during the call
- Saying "real person" or "speak to someone"
- Saying "stop calling" or "don't call me" (for follow-up prevention)
When triggered, the AI will:
- Log the request
- Offer to have the restaurant call back
- Provide alternative contact methods
### 6.3 Right to Opt Out
Callers who do not wish to interact with AI may:
- Hang up before or immediately after the disclosure
- Request to be placed on a do-not-call list
- Contact the restaurant directly through other means
### 6.4 Right to Data Access and Deletion
Callers may:
- Request a copy of their call recording and transcript
- Request deletion of their data
- Opt out of SMS communications by replying STOP
Submit requests to: hi@agents.miami
## 7. Client (Restaurant) Responsibilities
### 7.1 Transparency Obligation
By using agents.miami, restaurant clients agree to:
- Not modify the AI disclosure script to hide AI nature
- Not instruct the AI to claim it is human
- Inform staff about AI call handling
- Verify AI-extracted reservations for accuracy
### 7.2 Caller Complaints
If callers complain about AI handling, restaurants should:
- Acknowledge the AI nature of the service
- Offer human callback or alternative contact
- Report persistent issues to agents.miami support
## 8. Regulatory Compliance
### 8.1 Federal Communications Commission (FCC)
agents.miami complies with FCC regulations regarding:
- Automated calling systems
- Caller ID accuracy
- Recording disclosure requirements
### 8.2 State Laws
We comply with all applicable state laws regarding:
- Two-party consent states (strictest standard applied)
- AI disclosure requirements
- Consumer protection regulations
### 8.3 FCC AI Voice Ruling Compliance
In accordance with the FCC's 2024 declaratory ruling treating AI-generated voices as "artificial voices" under the Telephone Consumer Protection Act (TCPA):
- All calls handled by our AI system clearly identify the use of AI-generated synthetic voice technology
- Callers are informed at the start of each call that they are interacting with an AI, not a human
- For outbound automated calls (callbacks, reminders), we require documented prior express consent that explicitly mentions AI-generated voice technology
- Callers may opt out of AI interaction at any time by saying "human"
### 8.4 Industry Standards
We adhere to:
- CTIA messaging guidelines
- Carrier requirements for A2P messaging
- Telephony industry best practices
## 9. EU AI Act Compliance
As a Netherlands-based provider, ClientBuddy complies with the EU AI Act (Regulation (EU) 2024/1689).
### Transparency Obligations (Article 50)
Effective August 2, 2026, the EU AI Act mandates that users interacting with AI systems must be clearly informed of this fact. Our compliance measures include:
- Clear AI Disclosure: All callers are informed at the start of each call that they are interacting with an AI-generated voice assistant
- Synthetic Content Marking: Audio outputs generated by our AI system are identified as AI-generated where technically feasible and required by law
- No Deception: Our AI systems never impersonate humans or deceive users about their artificial nature
- Human Escalation: Users may request human assistance at any time during AI interactions
### Risk Classification
Our AI voice assistant is classified as a "limited risk" AI system under the EU AI Act, subject to transparency obligations but not the stricter requirements applicable to high-risk AI systems.
### Potential Regulatory Updates
See Privacy Policy §14.3 for details on potential Digital Omnibus updates that may affect EU AI Act timelines.
## 10. Grievances and Complaints
### 10.1 Reporting Issues
If you believe AI processing has caused an error or harm:
1. For Callers: Contact the restaurant directly or email hi@agents.miami
2. For Restaurants: Contact hi@agents.miami
### 10.2 Response Process
We will:
- Acknowledge complaints within 2 business days
- Investigate the AI decision in question
- Provide a response within 10 business days
- Take corrective action where appropriate
## 11. Updates to This Policy
This AI Disclosure Policy may be updated to reflect:
- Changes in AI technology
- New regulatory requirements
- Feedback from users and regulators
Material changes will be communicated to clients and posted on our website.
## 12. Contact Information
For questions about AI disclosure:
ClientBuddy
A Netherlands-registered company (KvK: 97483672)
Operating agents.miami
- Email: hi@agents.miami
- Website: https://agents.miami/ai-policy
ClientBuddy, operating agents.miami, is committed to transparent, ethical AI operations. We believe callers deserve to know when they're speaking to AI, and we design our systems to be honest and helpful.