AI & DATA USAGE TRANSPARENCY

Buzzer.ai is committed to transparency in how we use artificial intelligence and process data within our platform. This document describes how AI operates within the Buzzer Services, how your data is handled, and the safeguards we have in place. It supplements our Terms of Service and Privacy Policy.

1. HOW AI OPERATES IN THE BUZZER PLATFORM

The Buzzer platform deploys AI capabilities across telephony and omnichannel communication workflows. AI is used in the following capacities:

  1. AI Voice Agents: Autonomous AI agents that conduct voice conversations with callers, handle inquiries, collect information, and perform actions within connected systems (e.g., opening tickets, retrieving records).
  2. AI Text Agents: Agents that operate across text-based Supported Channels (WhatsApp, SMS, email, Messenger, Instagram, TikTok, Telegram, Line, and others) to handle customer interactions.
  3. Agent Copilot / Assist: AI that operates alongside human agents during live interactions, providing real-time transcription, suggested responses, data retrieval, and decision support.
  4. Transcription and Summarization: Real-time conversion of voice communications to text, and generation of structured summaries for record-keeping and workflow automation.
  5. Language Translation: Real-time translation capabilities to bridge language gaps during live communications.
  6. Compliance Monitoring: AI-powered analysis of communications to flag potential regulatory or policy compliance issues in real time.

2. DATA FLOW AND PROCESSING

2.1 Voice Data

Buzzer does not store audio recordings of calls.

Voice data is streamed and processed in real time for the purposes described above (transcription, AI agent interaction, summarization, translation). Once processing is complete, the audio stream is discarded. Only derived outputs (transcripts, summaries, structured data) may be retained, as configured by the Customer.

2.2 Text-Based Communications

Messages exchanged through Supported Channels are processed by AI agents or copilot systems as part of the Customer's configured workflows. Message content is retained within the Customer's account for the duration configured by the Customer or as required for service operation.

2.3 Transcripts and Summaries

AI-generated transcripts and summaries are derived outputs created during voice processing. These are stored within the Customer's account and are accessible to the Customer. Retention periods are determined by Customer configuration and applicable data retention policies.

2.4 Metadata

Operational metadata (call timestamps, durations, channel identifiers, routing decisions, agent assignments) is collected for service operation, quality monitoring, and troubleshooting. Metadata does not include the content of communications.

3. USE OF CUSTOMER DATA FOR AI

We do not use Customer Data to train general-purpose or shared AI models.

Your data is not made available to other customers, used to improve services for third parties, or contributed to shared model training datasets.

3.1 What We Do

  1. Process Customer Data in real time to deliver the Services (transcription, AI agent responses, summaries, translations).
  2. Use aggregated, anonymized, and non-identifiable usage metrics to monitor system performance and reliability.
  3. Where specifically agreed with the Customer, fine-tune existing AI models to improve accuracy and performance for that Customer's specific use case. Such fine-tuning is performed under contractual and technical safeguards.

3.2 What We Do Not Do

  1. Train shared or general-purpose AI models on Customer Data.
  2. Share Customer Data with other customers or third parties for model training purposes.
  3. Retain audio recordings of voice communications.
  4. Use Customer Data for advertising, profiling, or purposes unrelated to service delivery.

4. THIRD-PARTY AI PROVIDERS

The Buzzer platform may utilize third-party AI models and services for specific capabilities, including but not limited to:

  1. Speech-to-text (automatic speech recognition)
  2. Text-to-speech (voice synthesis)
  3. Natural language understanding and generation
  4. Language translation
  5. Sentiment and intent analysis

When third-party AI providers are used:

  1. Data is transmitted to these providers solely for the purpose of processing the specific request (e.g., transcribing a speech segment).
  2. Our agreements with these providers prohibit them from using Customer Data for their own model training or for any purpose other than fulfilling the processing request.
  3. We select providers based on their security practices, data handling policies, and compliance posture.
  4. For On-Premise Deployments, the Customer may have the option to use locally-hosted models, eliminating the need for data transmission to third-party providers.

5. AI OUTPUT LIMITATIONS AND HUMAN OVERSIGHT

AI-generated outputs within the Buzzer platform are produced by machine learning models and are inherently probabilistic. This means:

  1. Transcription: Automated transcripts may contain errors, particularly in noisy environments, with heavy accents, or with domain-specific terminology. Transcripts should not be treated as verbatim legal records without human review.
  2. Summaries: AI-generated summaries may omit details, mischaracterize nuance, or introduce inaccuracies. They are intended as aids, not replacements for human review of source material.
  3. AI Agent Responses: AI agents operate based on configured instructions and available data. They may produce responses that are incorrect, incomplete, or contextually inappropriate. Customers are responsible for configuring appropriate guardrails, escalation paths, and human oversight.
  4. Compliance Monitoring: AI-based compliance flagging is a support tool, not a guarantee of regulatory compliance. It may produce false positives or miss violations. It does not replace human compliance review.

Customers operating in regulated industries (healthcare, finance, insurance, telecommunications) are advised to implement human review processes proportional to the risk and sensitivity of the AI-assisted workflows.

6. DEPLOYMENT MODELS AND DATA RESIDENCY

6.1 SaaS (Cloud-Hosted)

In our SaaS deployment model, Customer Data is processed and stored in cloud infrastructure located in Israel and the European Union. Specific data residency requirements can be discussed during onboarding.

6.2 On-Premise

In On-Premise Deployments, the Buzzer software operates within the Customer's own infrastructure. Customer Data remains within the Customer's network and is not transmitted to Buzzer or third-party cloud services, except where the Customer explicitly configures integrations that require external connectivity (e.g., third-party AI model APIs). The Customer is responsible for infrastructure security, access controls, and compliance in On-Premise Deployments.

7. SECURITY AND COMPLIANCE

  1. Encryption: Data is encrypted in transit (TLS 1.2+) and at rest using industry-standard encryption algorithms.
  2. Access Controls: Role-based access controls restrict access to Customer Data to authorized personnel and systems.
  3. Audit Logging: Access to Customer Data is logged for security monitoring and audit purposes.
  4. ISO 27001:2022: Our information security management system is aligned with ISO 27001:2022 standards.
  5. Incident Response: We maintain incident response procedures and will notify affected Customers of security incidents in accordance with applicable law and contractual obligations.

8. CUSTOMER CONTROLS

Customers have the following controls over how AI and data processing operate within their Buzzer deployment:

  1. AI Agent Configuration: Customers define the behavior, scope, and limitations of AI agents deployed in their workflows.
  2. Data Retention Settings: Customers configure retention periods for transcripts, summaries, and communication logs.
  3. Channel Selection: Customers choose which Supported Channels to enable and configure.
  4. Human Oversight Workflows: Customers configure escalation rules, human-in-the-loop requirements, and agent supervision settings.
  5. Data Export and Deletion: Customers may export their data or request deletion in accordance with our Terms of Service.

9. THIRD-PARTY PLATFORM CONSIDERATIONS

When the Services operate through Third-Party Platforms (e.g., WhatsApp Business API, telephony carriers, messaging platforms), the following applies:

  1. Each Third-Party Platform has its own terms of service, privacy policies, and data handling practices that apply to communications on that platform.
  2. Buzzer does not control and is not responsible for the data practices of Third-Party Platforms.
  3. The Customer is responsible for ensuring compliance with the terms and policies of any Third-Party Platform they use in conjunction with the Services.
  4. Availability of AI features on a given channel may be subject to limitations imposed by the Third-Party Platform.

10. UPDATES TO THIS DOCUMENT

We may update this AI & Data Usage Transparency document as our technology, practices, or regulatory environment evolves. Material changes will be communicated via email or through the Services. The "Last updated" date below indicates when this document was last revised.

11. QUESTIONS AND CONTACT

If you have questions about how AI and data processing work within the Buzzer platform, or wish to discuss specific data handling requirements for your deployment, please contact us at:

Buzzer.ai Ltd., 2 HaShonit, Herzliya, Israel

Last updated: March 2026