AI Use Policy
Published Winch legal document. This version is the current AI Use Policy for the Winch Legal / Trust platform.
This AI Use Policy explains how AI is used in Winch and what customers should expect when configuring AI-assisted commerce workflows.
AI-assisted service
Winch uses AI systems to help answer buyer questions, interpret documents, summarize context, route workflows, and support commerce operations. AI output is generated from prompts, configuration, business rules, available tools, and customer-provided data.
Human responsibility
Customers are responsible for deciding when AI output requires human review. Important commercial, legal, financial, safety, or customer-impacting decisions should be reviewed by an authorized human before being relied on.
Configuration
Customers are responsible for configuring accurate product data, policies, pricing rules, shipping rules, tax rules, knowledge content, and operating instructions. Incorrect configuration can lead to incorrect AI output.
Prohibited use
Users must not use Winch AI features to deceive people, violate laws, infringe rights, generate unlawful content, bypass safety controls, or process data they are not authorized to process.
Data processing
AI processing may involve customer content, buyer messages, catalog data, documents, and operational context. Details are described in the Privacy Policy, Data Processing Agreement, and Subprocessors document.
Limitations
AI systems may make mistakes, omit details, or respond unexpectedly. Winch may add safeguards, monitoring, logging, and evaluation workflows to improve quality, reliability, and safety.
Changes
Winch may update this AI Use Policy as the product, AI providers, safety practices, or legal requirements evolve.
What changed
Starter English legal documents and cookie inventory for the Winch Legal / Trust platform.
Previous versions
No archived versions are available yet.
Version metadata
- Content hash
- 2af21c44d6e41eff1e414b1890eec802e7ea82a8afc4b7c789a1db52bfb2900e
- Source path
- legal/documents/ai-use/2026-05-12.en.md
- Change level
- editorial
- Reacceptance required
- No