SCHEDULE D

AI SERVICES ADDENDUM

(EU AI Act-Aligned) Version 1.0 · Effective Date: 4 August 2026

This AI Services Addendum ("AI Addendum") forms an integral part of the Terms of Service (the "Agreement") accepted electronically by the Customer (including by clickwrap acceptance during account registration or online checkout). This document is published at a permanent URL on ONLYAI’s website and is incorporated into the Agreement by reference. It applies between Dostart Ventures Limited ("ONLYAI", "Provider") and the Customer identified in the applicable electronic Order Form ("Customer").

Capitalized terms not otherwise defined herein shall have the meanings assigned to them in the Agreement.

1. PURPOSE

1.1 Objective

The purpose of this AI Addendum is to establish the contractual framework governing Customer's use of the artificial intelligence functionalities incorporated into the Services.

This AI Addendum is intended to:

• promote trustworthy AI deployment;
• establish contractual governance principles for AI-assisted communications;
• allocate responsibilities between the Parties regarding the operation of AI-enabled functionality;
• support compliance with the Regulation (EU) 2024/1689 laying down harmonized rules on artificial intelligence (the "EU AI Act"), where applicable;
• supplement the Agreement with AI-specific operational requirements.

Nothing in this AI Addendum shall be interpreted as modifying the Parties' respective roles under the Agreement, the Data Processing Agreement or Applicable Law.

1.2 Scope

This AI Addendum applies to all AI-enabled functionality made available through the Services, including:

• AI Chat Engine;
• AI-assisted communication workflows;
• AI-generated recommendations;
• AI-generated message suggestions;
• AI-assisted sales automation;
• AI-assisted operational analytics;
• AI classification and prioritization functions;
• AI moderation and safety systems;
• any future AI functionality designated by ONLYAI as subject to this AI Addendum.

1.3 Relationship with the Agreement

This AI Addendum supplements the Agreement.

In the event of any conflict concerning AI-specific matters:

• Mandatory Applicable Law (including the EU AI Act);
• This AI Addendum;
• Master SaaS Agreement;
• Data Processing Agreement;
• Service Level Agreement;
• Remaining Schedules.

2. DEFINITIONS

For the purposes of this AI Addendum:

"AI System"

means the software functionality provided by ONLYAI that incorporates machine learning models, large language models, natural language processing, retrieval-augmented generation, orchestration logic, prompt engineering, classification systems or other computational techniques capable of generating outputs influencing digital communications.

"AI Output"

means any text, recommendation, response, classification, prioritization, summary, insight, workflow recommendation or other output generated by the AI System.

"Human Oversight"

means meaningful supervision by appropriately authorized personnel of the Customer, including the ability to review, intervene in, modify, suspend or override AI-generated outputs or automated workflows.

"Deployer"

shall have the meaning assigned under the EU AI Act.

"Provider"

shall have the meaning assigned under the EU AI Act, where applicable.

"Supported Platform"

means any third-party digital platform supported by the Services in accordance with the Agreement.

3. NATURE OF THE AI SERVICES

3.1 AI-Assisted Communication

The Parties acknowledge that the Services are designed to facilitate AI-assisted communication rather than fully autonomous communication.

The AI System operates as a configurable software tool intended to support Customer personnel in carrying out digital communication activities.

The Services are not designed to replace Customer's managerial, commercial or legal decision-making.

3.2 Configurable Technology Platform

Customer acknowledges that the AI System is configured and operated based upon parameters determined by Customer, including:

• Creator profiles;
• communication style;
• escalation thresholds;
• automation rules;
• monetization workflows;
• AI activation settings;
• prohibited topics;
• platform-specific restrictions.

ONLYAI does not independently determine these business parameters.

3.3 No Independent Decision-Making

The AI System does not independently:

• determine Customer's commercial objectives;
• determine whom Customer should communicate with;
• determine pricing strategies;
• determine legal compliance obligations;
• determine retention periods;
• determine lawful processing purposes.

The AI System executes Customer-defined configurations and generates probabilistic outputs intended to assist Customer personnel.

4. ALLOCATION OF RESPONSIBILITIES

4.1 ONLYAI Responsibilities

ONLYAI shall:

• provide and maintain the AI-enabled software platform;
• implement commercially reasonable AI governance measures;
• maintain AI safety mechanisms;
• maintain documented operational procedures;
• monitor the technical performance of the AI System;
• maintain AI security controls;
• implement updates and improvements where reasonably necessary.

ONLYAI does not supervise Customer's operational use of the AI System.

4.2 Customer Responsibilities

Customer remains solely responsible for:

• determining whether AI functionality should be enabled;
• configuring AI workflows;
• determining the scope of automation;
• supervising AI-assisted communications;
• reviewing escalated conversations;
• ensuring compliance with Applicable Laws;
• ensuring compliance with Supported Platform terms;
• maintaining meaningful Human Oversight;
• determining whether AI Outputs should be relied upon in any particular circumstance.

Customer acknowledges that AI Outputs are decision-support tools and not substitutes for human judgment.

4.3 Shared Responsibility Model

The Parties acknowledge that the AI functionality operates under a shared responsibility model.

ONLYAI is responsible for the secure operation of the AI platform and its underlying technical infrastructure.

Customer is responsible for the operational deployment of the AI functionality within its own business processes.

Nothing in this AI Addendum transfers Customer's regulatory, contractual or commercial responsibilities to ONLYAI.

5. HUMAN OVERSIGHT

5.1 Human Oversight Principle

The Parties acknowledge that meaningful Human Oversight constitutes a fundamental operational principle of the Services.

The AI System is designed to support, rather than replace, human decision-making. Customer shall ensure that appropriately trained personnel retain the ability to supervise, review, intervene in, suspend, modify or override AI-assisted activities where necessary.

Nothing in the Services is intended to eliminate human judgment in circumstances requiring legal, ethical, commercial or reputational assessment.

5.2 Human Intervention

Subject to the functionality of the subscribed Services, the Platform shall enable Customer to implement one or more Human Oversight mechanisms, which may include:

• review-before-send workflows;
• conversation escalation to human operators;
• manual takeover of AI-assisted conversations;
• approval workflows;
• AI suspension controls;
• configurable automation thresholds;
• conversation reassignment;
• emergency interruption mechanisms.

The availability of any particular functionality shall depend upon the Customer's subscription plan and the features purchased under the applicable Order Form.

5.3 Escalation Events

Customer acknowledges that the AI System may be configured to escalate interactions to human operators where predefined conditions are satisfied.

Examples may include:

• suspected underage users;
• suspected platform policy violations;
• suspected prohibited content;
• requests involving legal issues;
• requests involving financial matters;
• requests outside configured AI capabilities;
• suspected AI disclosure events;
• repeated failure of AI to satisfy user intent;
• confidence scores below configured thresholds.

The selection and configuration of escalation events remain Customer's responsibility.

5.4 Human Accountability

Customer shall remain solely responsible for all decisions made by its personnel following review of AI Outputs.

Human review does not transfer responsibility for Customer's business operations to ONLYAI.

6. AI OUTPUTS

6.1 Nature of AI Outputs

Customer expressly acknowledges that AI Outputs:

• are generated through probabilistic computational processes;
• are non-deterministic;
• may differ for substantially similar inputs;
• may contain factual inaccuracies;
• may contain incomplete information;
• may contain reasoning errors;
• may require independent verification.

AI Outputs should be considered informational assistance only.

6.2 No Reliance

Customer shall not rely exclusively upon AI Outputs when making:

• legal determinations;
• regulatory decisions;
• tax decisions;
• employment decisions;
• financial decisions;
• contractual commitments;
• other decisions involving significant legal or commercial consequences,

without appropriate independent review.

6.3 No Representation of Accuracy

ONLYAI does not represent or warrant that AI Outputs shall:

• be factually correct;
• be complete;
• be lawful in every jurisdiction;
• satisfy Customer's commercial objectives;
• comply with all Supported Platform policies;
• produce consistent results.

Customer acknowledges that AI technology continues to evolve and that output quality may improve or change over time.

6.4 Customer Validation

Customer remains solely responsible for determining whether AI Outputs are appropriate for use in Customer's particular operational environment.

7. AI SAFETY FRAMEWORK

7.1 AI Safety Measures

ONLYAI shall maintain commercially reasonable AI safety measures designed to reduce the likelihood of generating unlawful, harmful or platform-non-compliant outputs.

Such measures may include:

• safety prompts;
• moderation layers;
• content classifiers;
• prohibited content detection;
• confidence scoring;
• risk-based routing;
• conversation termination logic;
• escalation mechanisms;
• output filtering.

7.2 Prohibited Content Controls

The AI System may be configured to identify or restrict content involving, among other matters:

• minors;
• child exploitation;
• incest;
• coercion;
• non-consensual conduct;
• human trafficking;
• sexual exploitation;
• bestiality;
• extreme violence;
• self-harm;
• terrorism;
• hate speech;
• prohibited substances;
• illegal financial activity;
• sanctions violations;
• platform circumvention;
• off-platform payment solicitation.

ONLYAI reserves the right to expand, modify or improve prohibited content controls in response to changes in Applicable Law, Supported Platform policies or evolving AI safety standards.

7.3 Platform Integrity

The AI System may implement safeguards intended to support Customer's compliance with Supported Platform requirements.

Such safeguards may include restrictions relating to:

• external payment methods;
• prohibited contact sharing;
• competitor promotion;
• prohibited solicitation;
• platform-specific communication restrictions.

Customer acknowledges that such safeguards do not constitute legal advice or guarantee compliance with any third-party platform requirements.

8. AI TRANSPARENCY

8.1 Transparency Principles

ONLYAI is committed to promoting responsible and transparent deployment of AI technologies.

Accordingly, ONLYAI shall maintain documentation describing, where appropriate:

• the intended purpose of AI functionality;
• significant AI capabilities;
• operational limitations;
• known risks;
• Human Oversight mechanisms;
• supported configurations.

8.2 AI Documentation

Technical documentation may include, where appropriate:

• feature descriptions;
• AI architecture summaries;
• release notes;
• supported functionality;
• operational guidance;
• known limitations.

ONLYAI may update such documentation from time to time.

8.3 Customer Transparency Obligations

Customer acknowledges that Customer may have independent obligations under Applicable Laws or Supported Platform policies relating to disclosure of AI-assisted communications.

Customer remains solely responsible for determining whether, when and how any AI-related disclosures should be made to end users.

ONLYAI does not determine Customer's transparency obligations.

9. AI GOVERNANCE

9.1 Governance Programme

ONLYAI shall maintain an internal AI Governance Programme appropriate to the nature, scale and complexity of the Services.

The programme may include:

• AI risk management;
• model governance;
• change management;
• incident management;
• testing procedures;
• monitoring procedures;
• documentation standards;
• security controls;
• periodic governance reviews.

9.2 Continuous Monitoring

ONLYAI may continuously monitor the operational performance of AI functionality for purposes of:

• improving service quality;
• detecting abnormal behaviour;
• identifying operational issues;
• improving safety;
• maintaining platform reliability.

Operational monitoring shall not be used to make independent commercial decisions on behalf of Customer.

9.3 AI Risk Management

ONLYAI shall maintain commercially reasonable processes for identifying, assessing and mitigating material operational risks associated with AI functionality.

Risk assessments may consider:

• model performance;
• hallucination risk;
• prompt injection;
• abuse attempts;
• security threats;
• operational resilience;
• regulatory developments.

10. AI MODEL LIFECYCLE MANAGEMENT

10.1 Continuous Improvement

Customer acknowledges that AI technology develops rapidly.

Accordingly, ONLYAI may periodically:

• replace supported models;
• introduce new models;
• retire obsolete models;
• improve orchestration logic;
• modify routing strategies;
• improve prompt engineering;
• optimize retrieval mechanisms;
• update moderation systems.

Such updates are intended to improve the Services and shall not require Customer approval unless they materially reduce subscribed functionality.

10.2 Model Independence

ONLYAI does not guarantee the continued availability of any specific third-party AI model.

Where reasonably practicable, ONLYAI may substitute an alternative model providing substantially similar functionality.

10.3 AI Versioning

ONLYAI may maintain internal version control for AI components.

Customer acknowledges that AI outputs generated by different model versions may differ in wording, reasoning, tone or structure without constituting a Service Defect.

10.4 AI Routing Infrastructure

Customer acknowledges that ONLYAI may utilize third-party AI routing, orchestration or gateway infrastructure to facilitate access to one or more underlying artificial intelligence models.

Where such infrastructure is utilized:

(a) AI Inputs may be transmitted through an intermediary routing provider;

(b) the routing provider may facilitate selection of, or transmission to, an underlying AI model;

(c) ONLYAI may not maintain a direct technical integration with every underlying model provider;

(d) model availability may vary based upon routing configuration, provider availability, technical performance, cost, safety considerations or operational requirements; and

(e) ONLYAI may modify routing configurations or substitute supported models where reasonably necessary to maintain or improve the Services.

ONLYAI does not warrant the continued availability of any particular routing provider, underlying model provider or AI model.

Any Processing of Personal Data through such infrastructure shall remain subject to the applicable Data Processing Agreement and Subprocessor Register.

11. CUSTOMER RESPONSIBILITIES

11.1 Responsible Deployment

Customer shall deploy and use the AI System responsibly and in accordance with:

• the Agreement;
• this AI Addendum;
• Applicable Laws;
• the terms and conditions of Supported Platforms;
• internal governance policies adopted by Customer.

Customer shall ensure that the AI System is configured in a manner appropriate to Customer's intended business purposes and risk profile.

11.2 Human Oversight

Customer shall establish and maintain appropriate Human Oversight procedures proportionate to the nature of its use of the AI System.

Without limitation, Customer shall:

• designate appropriately trained personnel responsible for supervising AI-assisted communications;
• review escalated conversations;
• monitor AI performance within Customer's operational environment;
• periodically review AI configuration settings;
• implement internal approval procedures where appropriate.

11.3 Training

Customer shall ensure that Authorized Users responsible for supervising AI-assisted communications receive appropriate training regarding:

• operation of the Services;
• Human Oversight procedures;
• AI limitations;
• escalation procedures;
• applicable Supported Platform requirements;
• internal governance policies.

ONLYAI may provide training materials for informational purposes but remains under no obligation to certify Customer personnel.

11.4 Compliance with Applicable Law

Customer remains solely responsible for ensuring that its deployment and use of the AI System complies with all Applicable Laws, including any obligations applicable to Customer as a deployer, user or operator of AI systems under the EU AI Act or other applicable legislation.

12. PROHIBITED USES

12.1 General Restrictions

Customer shall not use, or permit any third party to use, the AI System:

(a) for any unlawful purpose;

(b) in violation of Applicable Laws;

(c) in violation of Supported Platform requirements;

(d) in any manner likely to infringe the rights of third parties;

(e) in any manner inconsistent with this Agreement.

12.2 High-Risk Activities

Unless expressly agreed in writing by ONLYAI, Customer shall not use the AI System:

• to make legally binding decisions concerning individuals;
• as the sole basis for employment decisions;
• as the sole basis for creditworthiness assessments;
• for law enforcement activities;
• for immigration or asylum determinations;
• for biometric identification;
• for biometric categorization;
• for emotion recognition in regulated contexts;
• for medical diagnosis or treatment decisions;
• for critical infrastructure management;
• for any use classified as prohibited under Applicable Law.

12.3 Circumvention

Customer shall not intentionally attempt to:

• circumvent AI safety mechanisms;
• disable moderation systems;
• bypass prohibited content filters;
• manipulate confidence thresholds;
• interfere with model routing;
• exploit known vulnerabilities;
• conduct prompt injection attacks;
• extract proprietary prompts;
• reverse engineer AI workflows.

12.4 Unsupported Automation

Customer shall not use the AI System in combination with unauthorized automation tools where such use materially increases security, regulatory or operational risks.

13. AI INCIDENTS

13.1 AI Incident Management

ONLYAI shall maintain documented procedures for identifying, assessing, recording and managing material AI-related operational incidents.

AI Incidents may include:

• unexpected model behavior;
• safety control failures;
• abnormal output patterns;
• AI system degradation;
• material AI security events;
• prompt injection attacks;
• model misuse;
• unauthorized AI behavior.

13.2 Customer Notification

Where ONLYAI determines that an AI Incident materially affects Customer's use of the Services, ONLYAI shall use commercially reasonable efforts to notify Customer without undue delay.

Notifications may include:

• description of the Incident;
• affected functionality;
• mitigation measures;
• available workarounds;
• expected remediation activities.

13.3 Customer Reporting

Customer shall promptly notify ONLYAI where Customer becomes aware of:

• suspected AI malfunctions;
• unexpected AI behavior;
• attempted circumvention of AI safeguards;
• material safety concerns;
• security vulnerabilities affecting AI functionality.

13.4 Continuous Improvement

Customer acknowledges that information relating to AI Incidents may be used by ONLYAI solely for purposes of improving:

• platform security;
• AI safety;
• operational resilience;
• system reliability;

provided that such use remains subject to the Agreement, the Data Processing Agreement and Applicable Data Protection Laws.

14. REGULATORY COOPERATION

14.1 Regulatory Developments

The Parties acknowledge that the legal framework governing artificial intelligence continues to evolve.

The Parties shall cooperate in good faith where amendments to Applicable Law materially affect the operation of the Services.

14.2 Information Requests

Where reasonably necessary to support Customer's compliance obligations, ONLYAI shall make available information reasonably necessary regarding:

• the intended purpose of AI functionality;
• technical characteristics of the AI System;
• implemented governance measures;
• Human Oversight functionality;
• AI safety controls.

Processor shall not be required to disclose trade secrets, proprietary algorithms, confidential prompt engineering methodologies, source code or other confidential intellectual property except where required by Applicable Law.

14.3 Regulatory Audits

Where a competent regulatory authority lawfully requires information relating to the AI System, the Parties shall cooperate in good faith while preserving the confidentiality of proprietary information to the maximum extent permitted by law.

15. LIABILITY

15.1 AI Outputs

Customer acknowledges that AI Outputs are generated probabilistically and remain subject to human review.

ONLYAI shall not be liable for decisions made by Customer or its personnel based upon AI Outputs.

15.2 Operational Decisions

Customer remains solely responsible for:

• communications sent through Customer accounts;
• commercial strategies;
• pricing decisions;
• monetization;
• Creator management;
• Supported Platform compliance;
• business decisions.

The AI System serves solely as a decision-support and communication-assistance tool.

15.3 Regulatory Compliance

Nothing contained in this AI Addendum transfers Customer's legal or regulatory obligations to ONLYAI.

To the extent permitted by Applicable Law, ONLYAI shall not be responsible for Customer's failure to comply with:

• the EU AI Act;
• GDPR;
• consumer protection legislation;
• Supported Platform terms;
• industry-specific regulations.

15.4 Application of Agreement

The limitation of liability contained in the Agreement shall apply equally to this AI Addendum.

16. CHANGES TO AI FUNCTIONALITY

ONLYAI continuously develops the AI System.

Accordingly, ONLYAI may:

• improve models;
• replace models;
• modify orchestration;
• enhance safety mechanisms;
• improve moderation;
• optimize routing;
• improve inference efficiency;
• introduce new AI capabilities.

Such changes shall not require Customer approval provided they do not materially reduce the subscribed functionality.

17. SURVIVAL

The following provisions shall survive termination of the Agreement:

• Confidentiality;
• Liability;
• Regulatory Cooperation;
• Intellectual Property protections;
• Governing Law;
• Dispute Resolution.

18. ORDER OF PRECEDENCE

In the event of conflict concerning AI-specific matters:

• Mandatory Applicable Law;
• This AI Addendum;
• Master SaaS Agreement;
• Data Processing Agreement;
• Service Level Agreement;
• Remaining Schedules.

ANNEX I – AI GOVERNANCE PRINCIPLES

The Parties acknowledge the following operational principles governing the Services:

• Human Oversight – AI-assisted communications remain subject to meaningful human supervision.
• Human Agency – Customer retains ultimate control over operational decisions.
• Transparency – ONLYAI provides documentation describing intended AI functionality and known limitations.
• Technical Robustness – ONLYAI maintains commercially reasonable testing, monitoring and resilience measures.
• Safety – The AI System incorporates safeguards intended to reduce unlawful or harmful outputs.
• Security – AI functionality is supported by information security measures proportionate to the risks presented.
• Continuous Improvement – ONLYAI continuously evaluates and improves AI functionality.
• Accountability – The Parties acknowledge their respective responsibilities under the Agreement and Applicable Law.

ANNEX II – AI SAFETY CONTROLS

Without limitation, the AI System may include configurable safeguards relating to:

• detection of underage users;
• prohibited content filtering;
• violence and exploitation controls;
• platform policy compliance;
• confidence-based escalation;
• conversation termination logic;
• prompt injection detection;
• anomaly detection;
• abuse prevention;
• human escalation mechanisms.

ONLYAI may modify these safeguards to reflect evolving legal requirements, platform policies and industry best practices.

ACCEPTANCE

This AI Services Addendum is incorporated into and forms part of the Terms of Service.

Customer’s electronic acceptance of the Agreement or electronic submission of the applicable Order Form constitutes acceptance of this AI Addendum. No signature is required.