Trust & legal

AI at Xylem

Clear language about how AI-assisted features work at Xylem — and where human judgment still leads.

Last updated August 6, 2026Version 1.0.0

Philosophy

AI at Xylem is meant to assist decision-making, not replace it. When features use automated or AI-assisted processing, Xylem Strategic, LLC labels them calmly in the product and keeps humans in control of consequential actions.

AI does not make its output inherently correct. Xylem does not represent generated recommendations as guaranteed marketplace outcomes. Generated content can contain inaccuracies and should be verified.

AI-generated analysis is not financial, legal, medical, or other professional advice.

Where AI may assist

Depending on the features you use, AI may help with:

  • Marketplace performance analysis and pattern spotting
  • Operational recommendations and summaries
  • Listing optimization drafts — titles, bullet points, descriptions, and keywords
  • Categorization suggestions
  • Image generation or editing workflows
  • Workflow assistance inside the product

Some diagnosis and prioritization experiences may also use transparent, rule-based logic rather than generative models. We still expect you to apply judgment before acting.

Human review before publishing

Marketplace-visible or materially consequential changes follow a deliberate lifecycle:

Generated → Review → Approved → Publish

AI produces drafts. Xylem is designed so that newly generated listing copy, product claims, pricing actions, advertising actions, or images are not pushed to a marketplace solely because an AI created them. Autonomous modes may come later; the foundation preserves human control for externally visible changes.

Product facts stay authoritative

Authoritative product data (from your catalog / PIM) should ground generation. AI listing copy is a draft — never the canonical source of truth for dimensions, certifications, compatibility, materials, quantities, warranties, or performance, health, safety, or environmental claims.

Third-party model providers

When a feature requires a third-party AI or model provider, relevant inputs may be processed by that provider to perform the function you requested. That processing is governed by our Privacy Policy and the provider's applicable terms.

We only name specific providers publicly when the product depends on them and disclosure is useful. Prefer in-product context labels over generic popups — we do not require a mandatory AI consent ceremony on every login.

Copyright and rights

Submit only content you have the right to use. You are responsible for ensuring generated and published materials do not violate third-party rights or marketplace policies. AI output is not automatically copyrightable everywhere, and it is not guaranteed free of third-party intellectual-property issues.

Questions? See Support or our Terms of Service.