AI Usage & Transparency

Ziantrix AI Technologies Private Limited ("Ziantrix", "we", "us", or "our") provides an AI-native HRMS for employee operations, payroll calculation support, HR service delivery, workforce workflows, and compliance coordination. This AI Usage & Transparency document explains how AI is used in Ziantrix, where human oversight is expected, how customer and employee data is handled, and what limitations customers should understand before using AI-assisted workflows in production.

We publish this document because HR and payroll systems operate in sensitive contexts. Employee data, salary information, workplace decisions, and compliance workflows require clear governance, careful access control, and responsible review.

AI in Ziantrix is used to assist HR, payroll, finance, IT, managers, and employees with repetitive operational work. Its purpose is to help retrieve relevant information, classify requests, summarize records, draft responses, identify potential anomalies, route work to the right human owner, and maintain more consistent process visibility.

Ziantrix does not position AI as a replacement for HR judgment, payroll ownership, legal interpretation, statutory decision-making, or manager accountability. The platform is designed to support operational teams, not remove responsible human review from sensitive decisions.

• Employee self-service support for policy, leave, onboarding, payslip, reimbursement, attendance, and HR process questions.
• Knowledge retrieval from customer-approved HR policies, documents, workflows, and configured knowledge sources.
• Payroll calculation support, validation assistance, exception surfacing, and explanation of payroll inputs or outputs.
• Case triage, intent classification, priority suggestions, routing, and escalation support for HR and payroll teams.
• Document review assistance, including extraction, summarization, classification, and workflow routing where enabled.
• Compliance workflow support, such as task tracking, reminder generation, evidence organization, and policy-based workflow prompts.
• Administrative summaries, internal notes, suggested responses, and operational reporting assistance.
• Customer-configured notification drafting and workflow messaging across supported communication channels.

• AI does not autonomously execute salary disbursements, initiate bank transfers, or move funds.
• AI does not make final payroll, tax, statutory, disciplinary, termination, hiring, promotion, compensation, or benefits decisions.
• AI does not replace customer review, HR approval, payroll sign-off, finance controls, or legal review.
• AI does not act as a lawyer, tax advisor, auditor, medical professional, or statutory authority.
• AI does not intentionally bypass tenant boundaries, user permissions, approval workflows, or customer-configured access controls.
• AI does not use customer data to train public AI models.

Ziantrix is designed around human accountability. AI can help prepare, explain, classify, summarize, or route work, but customer-authorized users remain responsible for reviewing outputs and making final decisions in sensitive workflows.

• Customers are responsible for configuring roles, permissions, approval chains, escalation paths, and knowledge sources.
• Customers are responsible for validating payroll outputs before payroll lock, filing, disbursement, or downstream processing.
• Customers are responsible for ensuring that workplace decisions are made in accordance with their policies, employment contracts, and applicable law.
• Customers should train authorized users to treat AI output as assistance, not as an unquestioned source of truth.

Ziantrix may support payroll calculation, validation, anomaly surfacing, workflow approvals, reimbursement context, salary component explanations, and audit-ready documentation. These capabilities are intended to improve visibility and reduce repetitive operational work.

Operational responsibility for payroll remains with the customer. Customers decide whether payroll inputs are complete, whether outputs are correct, whether exceptions are resolved, whether statutory treatment is appropriate, and whether payroll is approved for execution in their chosen payroll, ERP, or banking systems.

Customers may configure documents, policies, help content, workflow data, and other approved sources for AI-assisted search, summarization, extraction, and response generation. This helps employees and administrators find relevant information without manually searching through disconnected records.

• AI responses are grounded in customer-configured knowledge sources where available.
• Document-derived text may be used for search, retrieval, summarization, and workflow routing within the customer environment.
• Access to documents and retrieved content is governed by customer tenant, role, permission, and workflow controls.
• Customers should keep policies and source documents current, because AI-assisted answers depend on the quality and freshness of available information.

• Purpose limitation: AI features are used to provide and improve customer-specific HRMS workflows, support, retrieval, routing, reliability, and security.
• Tenant separation: customer data is logically separated by tenant, and access is governed by customer configuration and role-based permissions.
• Least privilege: access to sensitive data is limited to authorized users, workflows, and operational needs.
• No public model training: customer data is not used to train public AI models.
• Data minimization: customers should configure only the data and knowledge sources needed for their HRMS workflows.
• Auditability: important workflow, access, configuration, and AI-assisted activity may be logged for review, security, and compliance support.

Ziantrix applies privacy and security safeguards appropriate for enterprise HRMS use. These safeguards are designed to help protect sensitive workforce, payroll, document, and workflow data while allowing authorized users to operate the platform.

• Role-based access controls and customer-managed permissions.
• Tenant isolation for customer environments.
• Encryption for data in transit and at rest, including additional protection for highly sensitive personal data where applicable.
• Authentication and enterprise identity capabilities on eligible plans.
• Security monitoring, operational controls, and audit logging for important platform activity.
• Privacy workflows that support access, export, correction, deletion, and retention processes where applicable.

These safeguards are described at a public level. Detailed security architecture, monitoring logic, internal detection rules, and infrastructure topology are not disclosed in this document.

• AI output is presented as assistive content rather than an automatic final decision.
• Sensitive workflows can be routed to human reviewers based on customer configuration and operational context.
• AI-assisted responses can include source-aware context where configured, helping users understand the basis of an answer.
• Permissions and tenant boundaries apply to AI-assisted retrieval and response workflows.
• Known limitations are communicated to customers and users in trust documentation and product experience where appropriate.
• Customer administrators can configure knowledge sources, access, escalation behavior, and workflow approvals.

AI-generated output may occasionally be incomplete, outdated, ambiguous, or inaccurate. This can happen when source documents are outdated, inputs are incomplete, context is missing, policies conflict, permissions are misconfigured, or a question requires professional judgment.

• Customers should review AI-generated responses before relying on them for critical HR, payroll, compliance, or financial decisions.
• Employees should be given clear escalation paths when an answer appears incorrect, unclear, incomplete, or sensitive.
• Administrators should maintain current knowledge sources and review unresolved or escalated cases.
• AI should not be used as the sole basis for decisions that materially affect an employee.

Ziantrix supports escalation-to-human workflows so that sensitive, uncertain, unresolved, or policy-dependent matters can be routed to authorized teams. Escalation behavior depends on customer configuration, enabled modules, user roles, and operational process design.

• Unresolved employee queries can be routed to HR or payroll teams.
• Payroll exceptions can be surfaced for review before customer approval.
• Document or onboarding issues can be assigned to appropriate owners.
• Compliance workflow tasks can be tracked, assigned, and reviewed by accountable users.
• Audit logs may record relevant workflow steps, approvals, escalations, and changes.

Ziantrix should be used in ways that support fair, explainable, and accountable HR operations. Customers should avoid using AI-generated content as the sole basis for employment decisions or in ways that conflict with labor law, anti-discrimination requirements, employee notices, workplace policies, or contractual obligations.

• Use AI to assist operational consistency, not to remove accountability from HR and management.
• Review outputs for bias, missing context, policy mismatch, or inappropriate tone before sending sensitive communications.
• Provide employees with human escalation paths for sensitive or contested matters.
• Avoid configuring AI workflows that infer protected characteristics or make decisions based on inappropriate attributes.
• Maintain clear internal policies for how AI-assisted HR workflows may be used.

Customers control how Ziantrix is configured for their organization. Available controls may vary by plan, enabled modules, integrations, and contractual configuration.

AI-assisted workflows may process prompts, retrieved context, documents, workflow metadata, user feedback, system-generated outputs, and audit events. Retention depends on the data type, customer configuration, legal obligations, security needs, contract terms, and backup lifecycle controls.

Customers should configure retention, deletion, export, and access policies according to their privacy obligations and internal governance requirements. Where Ziantrix acts as a processor, we support customers in responding to valid privacy requests as described in our Privacy Policy and applicable agreements.

We evaluate AI-assisted experiences to improve usefulness, reliability, safety, and clarity. Improvement activities may include reviewing aggregate performance, support feedback, error patterns, escalation outcomes, and product telemetry. Customer data is not used to train public AI models.

As AI features evolve, we may update this document to describe new capabilities, changed safeguards, additional customer controls, or revised limitations. We aim to keep public disclosures accurate without exposing sensitive operational details.

For AI governance, privacy, or trust questions, contact privacy@ziantrix.com. For legal notices, contact legal@ziantrix.com. General inquiries may be sent to contact@ziantrix.com.

Ziantrix AI Technologies Private Limited, Plot No. 68, E/Part, Sri Ram Nagar Colony, Pasmamala, Rangareddy, Hyderabad, Telangana 501505, IN.

We may update this document to reflect changes in product capabilities, legal expectations, customer controls, or responsible AI practices. Material changes will be reflected by updating the version or last updated date.