AI disclosures and intended use
Last updated August 19, 2026. What each AI feature is for, what it cannot do, and the controls around it.
How AI use is described
Zennvue describes AI as assistance inside a human-operated workflow, not as independently certified or as a guarantee of an outcome. Publicly described safeguards are human review before external use, reviewable activity records, and workspace access controls. The public site does not claim that those records are tamper-evident or that the current authorization model is attribute-based. Prohibited uses and consequences are set out in the Acceptable Use Policy.
Content and model training
Customer content, including couple-to-vendor messages, contracts, guest information, and event details, is not represented as training material for a shared cross-customer model. Tenant-scoped adaptation, where contractually enabled, must remain inside that tenant. A generally available self-service disablement control is not currently claimed. Details are in the Privacy Policy.
Provider transparency
The current foundation-model provider and its retention, training, and processing-location terms have not yet been cleared for publication on this site. Until the provider register is complete, do not place regulated, highly sensitive, or third-party confidential information into an AI feature unless your signed agreement identifies the approved processing boundary.
Automated processes that order outcomes
Two AI features can order outcomes: lead scoring can order a vendor's inbox, and matching can order the vendors a couple sees; paid placement may also affect position. Neither is represented as accepting, rejecting, or blocking an inquiry automatically. Request human review, correction, or disablement of AI lead scoring for your workspace by emailing hello@zennvue.com. A self-service opt-out is not currently claimed.
Disclosure to recipients
Zennvue does not currently claim that every AI-assisted draft carries a persistent technical marker after export. The person sending or publishing a draft remains accountable for it and should disclose material AI involvement when the recipient could otherwise reasonably believe the content was independently authored or professionally reviewed.
AI vendor matching
Intended use. Ranks marketplace vendors against a couple's date, style, budget, and location so the couple sees a fit-ordered shortlist. Paid placement can affect position. The marketing site does not independently verify the authenticated product's placement-label behavior.
Known limitations. Matching works from the information couples and vendors provide; incomplete profiles or preferences reduce match quality. A high match score is a relevance estimate, not a guarantee of availability, price, or service quality.
Not for. Not a background check, license verification, or endorsement of any vendor. Do not treat match order as a safety or quality certification.
AI drafting (proposals, replies, descriptions, emails)
Intended use. Drafts text in the vendor's workspace from the inquiry and workflow context. Any tenant-scoped brand adaptation and its disablement controls are deployment-dependent and are not represented here as generally available. Nothing should be sent until the responsible human reviews it.
Known limitations. Drafts can be incomplete, generic, or wrong about details like dates, prices, and inclusions. The sender is responsible for every word that reaches a couple or client.
Not for. Not for generating deceptive claims, fake credentials, or messages that hide who is sending them; see the Acceptable Use Policy.
AI lead scoring
Intended use. Estimates how likely an inquiry is to become a booking so vendors can prioritize their inbox. Scores order the inbox; they do not answer or reject anyone automatically.
Known limitations. Scores are probabilistic and reflect patterns in inquiry data; they can be wrong about any individual inquiry. Vendors decide whom to answer, and a human can always review an inquiry regardless of score.
Not for. Not a basis for unlawful discrimination in deciding whom to serve, and not an automated decision that blocks any inquiry from being seen.
AI contract review
Intended use. Flags potentially risky clauses and missing terms in a wedding or event contract so the reader knows where to look before signing.
Known limitations. It can miss issues and flag non-issues. It is not legal advice, it does not know your full situation, and it does not replace a lawyer for significant agreements.
Not for. Not a substitute for professional legal review, and not a determination that a contract is safe to sign.
AI planning assistant
Intended use. Builds budgets, timelines, and checklists for couples from their event details and preferences.
Known limitations. Estimates are starting points calibrated to typical weddings, not quotes; real costs and timing vary by market, season, and vendor.
Not for. Not financial advice and not a commitment that any vendor can deliver at an estimated price or time.
The rule that applies to every feature
AI output can be incomplete, incorrect, or unsuitable, and it is not legal, financial, tax, insurance, medical, or safety advice. Review it before sending, signing, paying, booking, or relying on it. Questions: hello@zennvue.com.