Responsible AI at Arena

Effective Date: 12/2025

Last Modified Date: 12/2025

Arena AI is a suite of intelligent capabilities designed to enhance productivity, decision making, and operational efficiency across the Arena platform. These features will continue to expand and currently include the Arena AI Assistant, AI file actions, and automated risk and compliance, powered by machine learning and natural language models.

This document explains the design of Arena AI, details how it handles customer data, and describes the ways customers can manage and control its features.

1. AI Services and Governance

Arena AI adheres to the PTC AI Governance Policy. All AI sub-processors and tools used within Arena are reviewed and approved by the PTC Governance Committee to ensure transparency, trust, and compliance.

2. Opt-In Design

Arena AI features that require customer data to be processed by large language models (LLMs) are disabled by default. Workspace Admins have full control over which AI capabilities are enabled. Features can be selectively enabled/disabled in Workspace Settings.

NOTE: Base level Supply Chain Intelligence actions and the Arena AI Assistant for Help are enabled by default because they do not share customer data with LLMs.

3. Data Usage and Privacy

Customer Data Protection: Arena AI keeps your data secure and does not use it to train models outside your instance.
No Cross-TenantSharing: Customer data is never shared across tenants or used to improve models globally.
Ephemeral AI Requests: Inputs and outputs are logged only for audit and performance tuning—not for training.
No Autonomous Actions: AI Assistants and Actions only suggest or summarize; they do not modify records or trigger workflows without explicit user permission.

4. Hosting and Infrastructure

Hosted on AWS: All Arena AI processing occurs in-region using AWS infrastructure.
Third-Party Models: Arena may use LLMs from Anthropic housed by AWS under strict confidentiality and data processing agreements.
Optional Use: Arena AI is optional. If used, AWS will be listed as a sub-processor.

5. Auditability and Controls

All AI feature enablement and disablement actions are logged and reviewable by Workspace Admins.
Arena’s enterprise Access Policies govern AI usage. Users cannot access or perform actions via Arena AI that they are not otherwise permitted to do.
Admins can revoke access or disable AI features from within workspace settings at any time.
Companies can request to have Arena AI settings removed from their workspace settings via a support ticket or email to [email protected].

6. Known Limitations and Guardrails

User Review Required: Generative AI output should be reviewed before acting on suggestions.
Content Guardrails: Arena AI includes safeguards to prevent generation of hateful, explicit, or violent content and to prevent the accidental revealing of sensitive data.

7. Support Process

For questions or concerns about Arena AI, please contact your Customer Success Manager or email [email protected].

8. Warranty Disclaimer

Please note that Arena AI is provided without warranties regarding the accuracy or suitability of its outputs. Arena AI aims to provide helpful and accurate information, but as with any AI technology, results may vary. We recommend that users review and verify any designs, workflows, or outputs generated by Arena AI before acting on them.

9. Service Level Terms

Downtime caused by third-party AI services is excluded from Arena’s uptime guarantees.