AI Transparency & Responsible AI Policy

1. About Evercrew

Evercrew Venture Pte Ltd is a Singapore-based AI workflow automation and local LLM solutions company focused on helping SMEs adopt practical, secure, and business-ready AI systems.

Company: Evercrew Venture Pte Ltd
UEN: 202551047C
Address: 18 Kaki Bukit Road 3 #05-16, Singapore 415978
Website: https://evercrew.ai/
Email: support@evercrew.ai

2. Purpose of This Policy

The purpose of this Policy is to explain how Evercrew.ai approaches responsible AI usage, AI system governance, data handling, transparency, human oversight, and risk management.

This Policy applies to our website, AI demonstrations, consultation materials, AI workflow automation services, local LLM deployments, AI assistants, prototypes, and client-facing AI solutions.

3. Our Responsible AI Principles

Evercrew aims to design and deploy AI systems that are practical, explainable, secure, human-supervised, and suitable for real business use.

Our responsible AI principles include:

• Human oversight: AI should support human decision-making, not blindly replace it.
• Transparency: Users should understand when AI is being used and what it is intended to do.
• Security: Business data should be handled with reasonable technical and organisational safeguards.
• Accuracy review: AI outputs should be reviewed before being used for important business decisions.
• Purpose limitation: AI systems should be used only for the agreed business purpose.
• Practical governance: AI workflows should have clear ownership, monitoring, escalation, and maintenance processes.
• Continuous improvement: AI systems should be reviewed and improved where appropriate.

4. Alignment with AI Standards and Good Practices

Evercrew.ai seeks to align its AI governance approach with recognised AI system life cycle and data life cycle practices, including principles reflected in ISO/IEC 5338 and ISO/IEC 8183.

ISO/IEC 5338 provides guidance on AI system life cycle processes, including the definition, control, management, execution, and improvement of AI systems across their life cycle.

ISO/IEC 8183 provides a data life cycle framework for AI systems, covering data-related activities across stages such as acquisition, creation, development, deployment, maintenance, and decommissioning.

Unless expressly stated in writing, references to ISO/IEC standards in this Policy describe Evercrew’s alignment with relevant good practices and do not represent a claim of formal ISO certification.

5. AI System Life Cycle Approach

Where appropriate, Evercrew applies a structured AI system life cycle approach when designing, deploying, and maintaining AI solutions.

This may include:

• Understanding the client’s business problem and intended AI use case.
• Assessing whether AI is suitable for the workflow.
• Defining system objectives, scope, limitations, and expected outputs.
• Designing workflow logic, prompts, integrations, and escalation paths.
• Selecting suitable AI models, tools, or local LLM deployment options.
• Testing AI outputs before production use.
• Monitoring deployed AI systems for performance, errors, and user feedback.
• Maintaining, updating, or retiring AI systems where appropriate.

6. AI Data Life Cycle Approach

Data is an important part of AI system quality. Evercrew aims to handle data used in AI workflows with reasonable care across its life cycle.

This may include:

• Understanding what business data is required for the AI workflow.
• Avoiding unnecessary collection of sensitive or irrelevant data.
• Using client data only for the agreed purpose.
• Reviewing data sources for relevance and quality where practical.
• Applying reasonable access controls and security measures.
• Supporting updates to knowledge bases, FAQ documents, or workflow rules.
• Helping clients remove, archive, or decommission data where appropriate.

7. Human-in-the-Loop Review

Evercrew encourages human-in-the-loop review for AI workflows that affect customers, finances, legal matters, compliance matters, employment matters, or important business decisions.

Depending on the project, AI workflows may include approval queues, escalation rules, confidence checks, audit logs, staff review steps, or manual override options.

AI should assist the business team by reducing repetitive work, drafting responses, summarising information, and organising tasks. Final responsibility for business decisions remains with the client or authorised human user.

8. AI Output Accuracy and Limitations

AI systems, including large language models, may produce outputs that are inaccurate, incomplete, outdated, biased, or unsuitable for a specific situation.

Users should not treat AI-generated outputs as guaranteed facts or professional advice. AI outputs should be reviewed before being used for:

• Legal or regulatory decisions.
• Financial or accounting decisions.
• Medical, health, or safety decisions.
• Employment or HR decisions.
• Customer commitments, quotations, contracts, or official company communications.

9. Cloud AI, Local LLM, and Hybrid Deployment

Evercrew may design AI workflows using cloud AI models, local LLM systems, or hybrid deployments depending on client needs, cost, performance, privacy, security, and maintenance requirements.

Cloud AI systems may provide strong performance and scalability but may involve third-party platforms or external processing. Local LLM systems may offer greater control over certain data flows but may require suitable hardware, maintenance, updates, and technical support.

Evercrew will aim to explain the practical trade-offs of cloud, local, and hybrid AI deployment options during consultation or project planning where relevant.

10. Client Data and Confidentiality

Evercrew may receive business data, workflow details, internal documents, customer enquiries, email samples, FAQs, policies, or operational information during consultations or projects.

Such information will be used only for the relevant enquiry, consultation, project, service delivery, troubleshooting, or agreed business purpose.

Specific confidentiality, data retention, access control, and deployment arrangements may be further defined in separate agreements, project scopes, service terms, or technical documents.

11. Security and Access Control

Evercrew aims to apply reasonable security measures when designing AI systems and automation workflows.

Depending on the solution, these measures may include:

• Role-based access control.
• Restricted administrative access.
• Separation of client data where appropriate.
• Secure API key handling.
• Use of reputable hosting, cloud, and software providers.
• Logging and monitoring where appropriate.
• Human approval for sensitive actions.

12. Responsible Automation

Evercrew’s AI workflow automation services are designed to reduce repetitive work, improve response speed, support staff productivity, and streamline business processes.

We do not encourage the use of AI automation for deception, impersonation, unlawful surveillance, spam, unauthorised data extraction, harassment, fraud, or any activity that violates applicable laws or platform policies.

13. Prohibited Uses

Users and clients must not use Evercrew’s website, AI demonstrations, AI workflows, or related systems for unlawful, harmful, or abusive purposes.

Prohibited uses include:

• Generating or distributing fraudulent, misleading, or deceptive content.
• Impersonating individuals or organisations without permission.
• Unauthorised access, scraping, extraction, or misuse of data.
• Creating malware, phishing content, spam, or harmful automation.
• Using AI outputs to make high-impact decisions without human review.
• Violating privacy, intellectual property, employment, consumer protection, or other applicable laws.

14. Monitoring, Maintenance, and Improvement

AI systems may require ongoing review, maintenance, and improvement after deployment.

Depending on the project scope, Evercrew may support clients with monitoring, issue review, prompt refinement, workflow updates, knowledge base maintenance, usage reports, and performance improvements.

Clients are encouraged to report inaccurate outputs, unexpected behaviour, security concerns, or workflow issues so that the system can be reviewed and improved where appropriate.

15. AI Governance Roles

Responsible AI deployment works best when responsibilities are clearly defined.

Depending on the project, Evercrew and the client may define roles for:

• Business owner or project sponsor.
• System administrator.
• Human approver or reviewer.
• Data owner or document owner.
• Technical support contact.
• Escalation contact for sensitive or uncertain cases.

16. Third-Party AI Providers and Tools

Some AI workflows may involve third-party AI providers, cloud platforms, software vendors, hosting providers, communication tools, or business applications.

Evercrew is not responsible for the independent policies, pricing, availability, security practices, or performance of third-party platforms. Where relevant, clients should review the applicable third-party terms, privacy policies, and data processing arrangements.

17. Transparency to End Users

Where AI systems interact with customers, employees, or external users, Evercrew encourages appropriate transparency.

This may include informing users when they are interacting with an AI assistant, providing escalation to a human staff member, and avoiding misleading representations that suggest an AI system is a human.

18. No Formal Certification Claim

Evercrew.ai follows and references recognised AI governance and life cycle practices, including ISO/IEC 5338 and ISO/IEC 8183, as guidance for responsible AI system design and data handling.

Unless expressly stated in a signed document or official certification notice, Evercrew does not claim that any specific system, workflow, product, or service is formally certified under ISO/IEC 5338, ISO/IEC 8183, or any other ISO standard.

19. Relationship with Other Policies

This Policy should be read together with our Privacy Policy, Terms of Use, project agreements, quotations, service scopes, and any other applicable documents.

If there is a conflict between this Policy and a signed client agreement, the signed client agreement will apply to the extent of the conflict.

20. Updates to This Policy

We may update this AI Transparency & Responsible AI Policy from time to time to reflect changes in technology, regulations, standards, business practices, or service offerings.

The updated version will be posted on this page with the latest revision date.

21. Contact Us

If you have questions about this Policy or Evercrew’s approach to responsible AI, please contact:

Evercrew Venture Pte Ltd
Email: support@evercrew.ai
Website: https://evercrew.ai/
Address: 18 Kaki Bukit Road 3 #05-16, Singapore 415978

Questions about our AI practices?

If you have questions about how Evercrew.ai uses AI, local LLM systems, workflow automation, data governance, or responsible AI practices, please feel free to contact us.

← Return to Homepage Email Us