Jewellink AI Features Addendum

Effective date: 5/01/2026

This AI Features Addendum applies to AI-assisted features in the Jewellink Services.

1. AI Features

AI Features may include suggested replies, AI chat, CRM insights, customer summaries, sales analytics, roleplay, coaching feedback, voice transcription, text-to-speech, avatar/video generation, image or media analysis, lead prioritization, generated templates, and other machine learning or generative AI functionality.

AI Features may be provided by Jewellink models, third-party AI providers, or a combination of both.

2. Inputs and Outputs

"Inputs" means Retailer Data, prompts, audio, messages, files, images, transcripts, instructions, and other content submitted to AI Features.

"Outputs" means text, audio, transcripts, summaries, recommendations, scores, generated media, replies, scripts, classifications, or other results generated by AI Features.

As between Retailer and Jewellink, Retailer retains ownership of Retailer-provided Inputs. Subject to the Master Terms, Retailer may use Outputs for its internal retail jewelry business purposes.

3. Human Review Required

AI Features are assistive tools only. Retailer and Authorized Users are responsible for reviewing, validating, editing, approving, and deciding whether to use Outputs.

Retailer must not rely on AI Features as the sole basis for legal, compliance, credit, employment, disciplinary, pricing, appraisal, insurance, medical, safety, or similarly significant decisions.

4. No Accuracy Guarantee

AI Outputs may be inaccurate, incomplete, offensive, biased, outdated, non-unique, or unsuitable. Jewellink does not warrant that Outputs are accurate, compliant, non-infringing, unique, complete, or appropriate for any particular purpose.

Retailer is responsible for verifying jewelry, inventory, pricing, discounts, financing, appraisal, warranty, provenance, scarcity, availability, customer preference, sales claims, customer records, POS data, transaction history, message history, usage data, sales analytics, and source data before using or relying on an Output.

AI Outputs and AI-assisted analytics depend on Retailer Data, source systems, integrations, mappings, configuration, and Retailer's instructions. Retailer is responsible for supplying accurate and authorized data, auditing AI Outputs, correcting source data and mappings, and deciding whether any Output is appropriate for customer-facing or business use.

5. AI Product Improvement

Jewellink may use Usage Data, Aggregated Data, De-Identified Data, feedback, evaluation data, and non-identifying telemetry to operate, secure, analyze, improve, and develop the Services and AI Features.

Jewellink will not use Retailer Customer Data containing personal information to train shared or generally available AI models unless Retailer has expressly authorized that use in writing.

If Jewellink offers an optional AI Product Improvement Program, the terms of that program should describe:

• what data may be contributed;

• whether data will be identifiable, de-identified, or aggregated;

• whether data may improve features for other retailers;

• whether third-party AI providers may receive the data;

• how Retailer may opt out or revoke participation;

• any limits on deletion after model training; and

• any additional security or confidentiality commitments.

6. Third-Party AI Providers

Jewellink may use third-party AI providers, such as language model, transcription, voice, avatar, media, or analytics providers, to provide AI Features.

Jewellink will use commercially reasonable efforts to configure third-party AI providers so that Retailer Data submitted through business/API services is not used by those providers to train their generally available models unless Jewellink discloses otherwise or Retailer authorizes such use.

Third-party AI providers may process Inputs and Outputs for abuse monitoring, security, service delivery, temporary retention, legal compliance, debugging, or other purposes permitted by their terms and applicable law.

7. Voice, Image, and Likeness

Retailer must obtain all rights and consents required to record, upload, transcribe, synthesize, clone, generate, or use any person's voice, image, likeness, name, performance, or identity.

Retailer must not use AI Features to impersonate a person, create deceptive endorsements, or suggest a person said or did something they did not say or do.

8. Prohibited AI Uses

Retailer must not use AI Features to:

• deceive, manipulate, defraud, harass, or exploit consumers;

• generate phishing, malware, credential harvesting, or social engineering content;

• make unlawful or misleading jewelry, financing, appraisal, warranty, investment, or authenticity claims;

• generate discriminatory or unlawful decisions;

• infer sensitive personal traits where unlawful or inappropriate;

• create fake reviews, fake testimonials, fake endorsements, or undisclosed synthetic media where disclosure is required;

• violate intellectual property, privacy, publicity, or confidentiality rights; or

• violate the Acceptable Use Policy or applicable third-party AI provider policies.

9. Similar Outputs

AI systems may generate similar or identical outputs for different users. Retailer does not receive exclusive rights in general ideas, concepts, formats, suggestions, or outputs independently generated for others.

10. Suspension

Jewellink may suspend, limit, disable, or remove AI Features if Jewellink reasonably believes use creates legal, security, privacy, platform, vendor, operational, reputational, or consumer risk.