Generative AI Chatbot Info Card
Published in compliance with the Infocomm Media Development Authority (IMDA) of Singapore's Transparency Guidelines for Generative AI Chatbots (released July 2026). This document serves as our standardized disclosure of system capabilities, limitations, safety guardrails, data protection practices under the Singapore PDPA, and consumer feedback mechanisms.
System Identity & Deployment Profile
Standardized technical disclosure following IMDA Annex B format
Let's Chat Autonomous Revenue & Resolution Agent
Domain-locked website chatbot and workflow automation agent
Digital Futures Consultancy LLP
Registered in Singapore (UEN: T17LL1937H)
Website Visitors & Consumers
Prospects, buyers, and existing customers seeking service info
Hybrid Origin-Locked RAG Architecture
Retrieval-augmented grounding with verified domain crawler index
Singapore IMDA & AI Verify
Agentic AI Governance Framework (v1.5) & PDPA 2012
legal@letschat.asia
4-hour response SLA for security & safety notices
Capabilities & Limitations
What the Chatbot CAN Do
- Provide Instant Verified Answers: Retrieves factual information strictly grounded in the client website's published pages, knowledge sources, and product documentation.
- Multi-Lingual Customer Support: Supports natural conversation across 49 locales (including English, Simplified/Traditional Chinese, Malay, Tamil, Indonesian, Tagalog, Vietnamese, Thai, Japanese, Korean, Hindi, Bengali, Urdu, Arabic, and European languages).
- Conversational Lead Qualification: Captures qualified prospect interest and structured contact requirements directly into CRM systems.
- Appointment Scheduling: Coordinates meeting bookings via direct calendar links or conversational appointment requests.
- Live Support Escalation: Hands off complex or sensitive discussions to human customer service representatives in real time.
Known Limitations & What It CANNOT Do
- No Professional Advice: The chatbot does NOT provide legal, medical, accounting, tax, or regulated financial advice. All conversational responses are indicative and informative.
- No Legally Binding Contracts: The chatbot cannot execute contracts or commit to contractual terms outside of the pre-approved published terms of service.
- No Direct Financial Transactions: The chatbot does not execute payment charges, bank transfers, or credit processing in conversation.
- Non-Deterministic Generative Outputs: As a Generative AI system, outputs are probabilistically synthesized. While grounded in site knowledge, phrasing may vary and subtle discrepancies can occur. Users should verify high-impact specifications with staff.
Prohibited Uses & Intended Audience
To maintain safety, ethical alignment, and regulatory compliance under Singapore law, the following interactions with the Let's Chat engine are strictly prohibited:
Users may not attempt system prompt exfiltration, delimiter hijacking, jailbreaking, or bypass of retrieval grounding guardrails.
The chatbot is not a 995/999 emergency service or suicide prevention hotline. For life-threatening emergencies, contact local emergency services immediately.
Transmitting defamatory, obscene, harassing, racist, or abusive content will result in immediate session termination and IP rate-limiting.
Designed for adult business users and general consumers aged 18 and older. Minors must use the platform under parental or guardian guidance.
Safety, Reliability & Technical Safeguards
Unlike legacy widgets that only check snippet possession, Let's Chat requires domain owners to complete cryptographic DNS TXT verification. Copied snippets deployed on unauthorized origins fail initialization immediately.
Every prompt is processed through a deterministic RAG gate. Responses cite verified source paths with confidence percentages. Out-of-distribution queries trigger fallback mechanisms rather than hallucinations.
If an administrator revokes a domain or modifies permissions, global in-memory edge caches invalidate authorization tokens in under 5 milliseconds across edge networks worldwide.
High-consequence actions (e.g. appointment approvals, customer record updates, dispute escalations) are routed through authenticated human review workflows before commitment.
Data Practices & Singapore PDPA Compliance
In accordance with the Singapore Personal Data Protection Act 2012 (PDPA) and IMDA Generative AI guidelines:
End-user chat queries and personal data are strictly never used to train, retrain, or improve foundational AI models.
Encrypted in transit via TLS 1.3 and at rest via AES-256 in Singapore-hosted sovereign cloud infrastructure.
Chat session logs are automatically purged after 30 days unless legal retention or client SLA dictates otherwise.
Data Subject Rights (Access, Correction & Erasure)
Under the Singapore PDPA, you have the right to request access to and correction or deletion of your personal data held in our systems. To exercise these rights, please contact our Data Protection Officer at legal@letschat.asia.
Feedback Channels & Incident Redress
In-Session User Feedback
Every message synthesized by the chatbot includes interactive thumbs up and thumbs down buttons. Users can flag inaccurate statements, report hallucinations, or commend helpful answers directly within the active chat window.
Formal Safety & Inaccuracy Reporting
To report potential hallucinations, security concerns, or harmful outputs, email our dedicated compliance desk. Reports are acknowledged within 4 hours.
Singapore IMDA Generative AI Chatbot Info Card • Document Version 2.4 (Updated October 2026)