AI & Automated Processing Policy

How Azmarq's AI Agent Studio and automated features may be used, and the limits that apply.

Draft for legal review. Template for Azmarq Technovation Pvt. Ltd. (India). Complements the Terms of Service, Service-Specific Terms, Privacy Policy, and DPA. Not legal advice.

1. Scope

What this Policy covers.

This Policy applies to AI Agent Studio and other Azmarq features that use machine learning, large language models, rules-based automation, or similar techniques to draft, classify, route, summarise, or send communications, or to assist agents in Unified Inbox and related products.

2. Roles and responsibility

Who is accountable for AI outputs.

You (the customer) decide when to enable AI features, which workflows they join, and what data they may access. Messages and actions taken via AI toward end users are your Customer Content and communications. You must supervise AI use appropriate to the risk (for example, financial advice, health, legal, or account takeover scenarios require human review).

Azmarq provides the tooling as a Processor for customer-configured AI processing of end-user data, and as a service provider for platform operation.

3. Permitted and prohibited uses

Acceptable AI use on the platform.

You may use AI features to assist customer service, FAQ handling, routing, summarisation, and drafting within your lawful business purpose, subject to the Acceptable Use Policy and Messaging Compliance Policy.

You must not use AI features to:

• generate or send spam, phishing, fraud, or deceptive content;
• impersonate individuals or authorities without clear disclosure where required;
• make fully automated decisions with legal or similarly significant effects on individuals where prohibited without human involvement and lawful basis;
• upload or process special-category data except as permitted by law and agreed in writing;
• attempt to extract model weights, circumvent safety filters, or abuse rate limits.

4. Training and model improvement

How we treat your data for AI.

Unless a written agreement says otherwise, Azmarq does not use your Customer Content (including end-user message content) to train public foundation models for unrelated third parties. [PLACEHOLDER: confirm whether anonymised/aggregated telemetry may be used to improve Azmarq's own safety or quality features — counsel to approve exact wording.]

Third-party model providers engaged as sub-processors (if any) are listed on the Sub-processors page and process data only to provide the feature.

5. Accuracy and disclaimers

Limits of AI.

AI outputs may be incomplete, outdated, or incorrect. Azmarq does not warrant that AI-generated content is fit for any particular purpose. You are responsible for verifying outputs before relying on them for consequential actions.

6. Transparency and human oversight

Disclosure and control.

Where law or channel policy requires disclosure that a recipient is interacting with an automated agent, you must configure and provide that disclosure. You should maintain human escalation paths for Unified Inbox and AI Agent Studio in customer-facing workflows.

7. Logging, retention, and deletion

Records of AI activity.

Prompts, transcripts, and tool actions may be logged to operate, secure, and debug the feature, and are retained per the Data Retention Policy and your account settings.

8. Changes and contact

Updates and where to reach us.

We may update this Policy; the "Last updated" date reflects the current version. Questions: dpo@azmarq.com / legal@azmarq.com.