CHAT4U LEGAL
AI Services Addendum
Last updated: November 4, 2026
1. Scope and relationship to the Agreement
This Addendum forms part of the Agreement between פרסום מי סטודיו לעיצוב גרפי והפקות דפוס בע"מ, trading as Chat4U, and the Customer. It covers selected model generation, retrieval, embeddings, ranking, classification, summaries, conversation insights, owner assistance, supported image/audio understanding and AI-supported tools. A reference to a capability is not an announcement that it is available in every plan, integration or channel.
The Terms govern general commercial matters and the common liability treatment, the AUP governs prohibited uses, and the DPA governs personal data processed on your behalf. This Addendum supplies AI-specific terms without overriding those protections. A supplier's own contract is not automatically a promise by Chat4U to give you the same warranty or indemnity.
2. Inputs, permissions and processing
Inputs include customer or visitor messages, authorized source passages, prompts, settings, supported media and relevant tool results. You must have the necessary rights and authority to provide and use them. Chat4U orchestrates processing through selected providers for the enabled purpose, including retrieval and generation, customer-specific analysis and appropriate support/security. The Privacy Policy and Subprocessor List explain data and recipient roles.
We do not use Customer Content or Output to train general-purpose or shared models and require that restriction from AI suppliers processing it for us. Customer-specific indexing, retrieval and agreed analysis remain permitted service functions. This commitment does not promise zero provider retention, exclusive models or processing in one country. Temporary files, traces and provider security records need their own controls and retention.
Customer-managed provider accounts or keys do not authorize a use contrary to this Agreement. Their settings and separate terms must be compatible with the selected processing. We remain responsible for our own handling and our subprocessor obligations; a customer key is not a blanket disclaimer of those duties.
3. Limitations and review
AI outputs are probabilistic. They can be false, incomplete, outdated, biased, misleading or inappropriate even when confident, supported by a citation or derived from a relevant source. Retrieval can miss or misread information; summaries and sentiment or interest labels can incorrectly characterize a person. Similar outputs can be generated for others.
You must evaluate the suitability of your deployment and use review proportionate to the consequences. Independently check factual assertions before relying on or publishing Output where an error could have significant legal, financial, safety or other consequences. Routine automated answers, including FAQs, may be delivered without individual human approval or separate pre-delivery checking of each answer; use proportionate testing and review of sources and configuration, with appropriate controls and escalation. A human review step does not itself permit high-impact processing prohibited by the AUP.
Guardrails and quality checks can be optional or configuration-dependent and can miss problems or block legitimate content. They do not guarantee correctness, compliance, detection of sensitive material or prevention of prompt injection. We do not represent that every response is checked or corrected by a human before delivery.
4. Output rights and liability
You retain rights in your inputs. Output use and assignment of any transferable Chat4U rights are governed by Terms section 6, subject to applicable law and underlying third-party rights. No rights are transferred in the underlying hosted platform, models, protected system materials or another person's content.
We do not guarantee that Output is unique, copyright-protectable, accurate, non-infringing or suitable for a particular use. We do not give a separate AI intellectual-property defense undertaking. Qualifying customer-caused third-party claims follow Terms section 18, including its absence of a contractual monetary ceiling, mixed-fault protections and defense and settlement procedure; not every claim concerning Output is automatically your responsibility. Terms section 17 supplies Chat4U's common aggregate ceiling and the applicable mandatory exceptions across the Agreement. This Addendum creates neither an additional AI cap nor immunity from a non-excludable duty.
5. Tools, actions and uncertainty
A suggested action, an authorized action, execution and a confirmed result are different stages. Use the supported approval flow for the displayed scope. An approval does not guarantee success; a timeout does not establish failure or permit automatic repetition of an action that may already have occurred. Check current state or follow the supported recovery before repeating an uncertain action.
Browser tools must not collect or transmit passwords, one-time authentication codes, card data, bank credentials or comparable secrets. This absolute rule is not relaxed by consent or an enterprise contract. Technical redaction and protected selectors reduce risk but do not prove that all sensitive information will be detected.
Native commerce tools check current eligible product options and cart state. The shopper's final informed action authorizes the displayed change; a changed executable product, quantity or option requires renewed authorization. Store text, notes and shipment descriptions are data, not trusted instructions to override safeguards. Unsupported extensions or purchase requirements are completed in the native store. Order creation, final charges and payment stay in native checkout.
6. Private order assistance
Supported WooCommerce order assistance requires a current signed-in shopper or secure proof for one guest order. A model cannot choose an unverified identity or infer authorization from an email, order number or cart token. Dedicated guest verification fields go to the native store and its mail provider, outside model context and durable Chat4U history.
Minimized authorized results may enter the configured AI processor to answer the current request. A fresh proof is checked before releasing the private answer, and a neutral verification message replaces that answer in durable Chat4U history. Information independently typed into ordinary chat is not automatically removed by this mechanism. Store-supplied shipment or tracking information can be incomplete or stale and is not a delivery guarantee.
These WooCommerce terms do not represent that Shopify or another integration grants the same private-order access; actual supported scope controls.
7. Channels, media and human help
Supported channels can provide messages, images or audio to the corresponding analysis service and retain derived descriptions or transcripts for the instructed conversation. An image description or transcription may be wrong. Do not present it as an authenticated identity, a sensitive-trait determination or permission for prohibited surveillance.
A human-help request is routed according to the Customer's available support arrangement. It does not guarantee that a person has joined, that a message reached the channel or that Chat4U will staff the conversation. The Customer must provide required AI and privacy information in its deployment; Chat4U remains responsible for its own disclosures and supplied behavior. Ordinary chatting is not blanket agreement to undisclosed data uses or marketing.
8. Restricted uses and changes
The AUP's high-impact, child-directed, sensitive-data and specifically excluded-use rules apply. The standard Service is not a qualified professional adviser or emergency service. A separately signed specialist agreement is possible only where expressly offered, lawful and supported; it cannot override a supplier prohibition or the secrets restriction.
Models and providers may change for compatibility, performance, availability, security or other legitimate service reasons. Such changes remain subject to the DPA's supplier authorization and objection rules, the Terms' major-service-change remedies and required agreement changes. No supplier change creates a new shared-training permission. Contact support@chat4u.ai about AI behavior or privacy@chat4u.ai about personal data.
English is the controlling language of this document. Version: 2026-11-consolidated-v1. Permanent link to this version. SHA-256: 81440d3a87b90b588ed4c225283d302c62253dc061681105e27d9298cc44a08e.