1. Introduction

Wisesight (Thailand) Co., Ltd. (hereinafter referred to as the "Company" or "Wisesight") develops and uses artificial intelligence (AI) systems as a key component of its social media data analytics products and services, its software services (SaaS), and its data-driven research.

The Company is aware of the opportunities and risks that accompany AI technology, particularly in the context of processing social media data involving the personal data of large numbers of persons. This Statement is prepared to:

Describe the nature of the AI systems the Company develops and uses.
Set out the Company's AI principles and governance measures.
Identify limitations, risks, and capabilities the Company prohibits itself from developing or using.
Establish complaint channels relating to the use of AI.

This Statement is intended to describe the Company's use of AI systems at an overview level and should be read together with the Privacy Policy, the service-specific privacy notices, the service agreements, the data processing agreements, and the Company's other relevant documents. Where a specific provision exists in any contract or service document, that specific provision prevails.

2. The Company's AI Principles

The Company adheres to the following principles in the development and use of AI systems:

Accountability & Explainability — The Company is accountable for the AI systems it develops and uses. The Company works to ensure that the operation of its AI systems can be explained at an appropriate and meaningful level, to users, clients, and data subjects alike.
Privacy & Data Protection — The Company attaches importance to the protection of personal data in the development and use of AI systems, applying Privacy by Design, which includes the reduction of PII (data minimisation), anonymisation, and access control.
Human Oversight — The Company designs its AI systems as decision-support tools, not as stand-alone automated decision-making with significant effects on individuals. Users and clients retain a central role in oversight, verification, and final decision-making.
Transparency — The Company publishes information about its AI systems through this Statement and related documents. The Company indicates clearly when a user is interacting with an AI system or when content has been generated by AI.
Security — The Company implements appropriate cybersecurity measures for its AI systems, vulnerability testing, and anomaly monitoring, in each case under the Company's security management policy (IT Security Policy).

3. Overview of the Company's AI Systems

Background AI — AI systems operating in the background to process the public data the Company collects, in order to generate outputs useful to clients, comprising:

Sentiment Classification — classifying text as positive/negative/neutral.
Topic Classification — classifying text into topic categories (products, services, brands).
Intent Classification — identifying the intent of a message, such as a question, a complaint, or praise.
Image Classification — identifying the type of object, scene, or content in an image.
Named Entity Recognition — categorising key information.
Speech-to-Text & OCR — transcribing speech to text and extracting text from images.
Influencer Profiling — analysing the public accounts or profiles of social media influencers to produce statistical indicators.

User-Facing AI — AI systems that users of the Company's platforms can invoke and interact with directly, comprising:

Wisesight AI Assistant — answers questions, analyses, and generates outputs according to the user's instructions; the user types a prompt to request outputs, and can draw on data from multiple Wisesight products simultaneously within a single platform.
Virtual Persona — creation of, and interaction with, simulated consumer groups (synthetic personas); the user defines the parameters of the target group and interacts with the persona.
AI Summary — summarises content, reports, or threads selected by the user; the user selects content and asks the AI to summarise it.
AI Insights — generates conclusions and recommendations from the context of the data the user is viewing; the system generates outputs contextually and the user can request further generation.
Research Assistant — an interactive dashboard integrating AI, allowing users to examine analytical results and interactively query the collected data.

AI in Service Delivery — In addition to on-platform AI features, the Company may use AI and automation as tools supporting its teams' work in delivering services, such as Research services and Monitoring and Alert services, to assist in compiling, classifying, summarising, and preparing reports or analytical results. The Company adheres to human oversight (human-in-the-loop), with the Company's personnel reviewing and exercising judgement over the work products or analytical results delivered to clients.

Internal Development AI — The Company uses AI tools in its product development processes, testing, quality assessment, preparation and verification of model-training datasets (data labelling/annotation), and other internal operations. These tools are used principally to support the work of the Company's teams. Where such processes involve personal data, the Company acts under appropriate personal data protection measures consistent with personal data protection law.

User-Connected External AI — In some products, the Company provides a Connector (an API or a Model Context Protocol (MCP) server) that allows corporate clients and Authorised Users to connect their own External AI Services (such as Claude or ChatGPT) to the products, to retrieve and analyse product data through conversation. In this case, the External AI Service belongs to the corporate client or the user; it is not an AI system developed or controlled by the Company, and the processing occurring within such service is beyond the Company's control.

4. Data Used in AI Development and Processing

Data used to train AI models — The Company's AI models are developed from several categories of data according to the purpose of each model:

Public data collected by the Company — used to train models relating to the analysis of Thai-language and other-language content; such data comprises posts, comments, and other public content on social media and public sources.
Curated/Annotated Data — data reviewed, classified, or labelled by the Company's teams for use in model training (e.g., verified sentiment examples).
Synthetic Data — data the Company generates using synthesis techniques for training or testing models.
Third-Party Foundation Models — the Company uses foundation models from external AI providers, trained on the providers' own data.

In training AI models the Company develops itself, the Company will not use identifiable personal data before it is used for model training. Where it is necessary to process data that can still be linked to an individual (such as pseudonymised data) for model development or testing, the Company will act under an appropriate lawful basis and the personal data protection measures required by personal data protection law.

Inference Data — When the Company's AI systems operate, the data processed at the time of use (inference) comprises:

Content the user inputs (prompts, queries, uploaded files).
The user's context and the client's instructions.
Data retrieved from the Company's databases at the user's instruction.
AI-generated outputs.

5. External AI Providers

OpenAI, L.L.C. — large language model APIs; text processing, summarisation, translation, question answering.
Anthropic, PBC — large language model APIs (Claude); text processing, analysis, content generation.
Google LLC — Gemini API; cloud AI model services.

6. Use of Artificial Intelligence

The Company uses artificial intelligence (AI) as a component of its products and services, including in the Company's back-end data processing operations. This section describes the use of AI as it relates to the processing of personal data.

AI Used by the Company to Process Public Data

The Company uses AI to classify and analyse the public data it collects, in order to generate outputs useful to clients, comprising: sentiment classification, topic classification, intent classification, influencer ranking and attributes, summarisation, inference of demographic attributes, and generation of embeddings for semantic search. These outputs are used for aggregate-level analysis and report preparation, and are not used for automated decisions producing legal effects or similarly significant effects concerning the data subject.

AI Invoked Directly by SaaS Platform Users

The Company provides AI features that SaaS platform users may invoke directly, comprising:

Wisesight AI Assistant — the user types a command or question (prompt) for the AI to assist with analysis, answer questions, or generate outputs as instructed.
Virtual Persona — a tool with which the user may create simulated consumer groups (synthetic personas) for research and hypothesis testing.
AI Summary — summarisation of content, reports, or threads selected by the user.
AI Translation — translation of text between languages.
AI Insights — conclusions and recommendations generated by the AI from the context of the data the user is viewing.
Research Assistant — an interactive dashboard integrating AI, allowing users to examine analytical results and interactively query the collected data.

When you use the AI features above, the Company will process your prompts, search terms, input content, and relevant context in order to generate the outputs you request.

External AI Providers

In some cases, the Company uses AI model services of external providers acting as data processors (such as OpenAI, Anthropic, Google).

For clarity, the External AI Providers under this clause mean providers the Company itself uses in delivering the AI features of its products, as distinct from External AI Services that corporate clients or users connect themselves; in the latter case, the External AI Service provider is a data processor of the corporate client, not of the Company.

7. User-Connected External AI Services

In addition to the External AI Providers the Company itself uses, certain of the Company's products (such as Zocial Eye) allow corporate clients and Authorised Users to connect their own External AI Services to the products through a Connector (an API or an MCP server), on the following principles:

Procurement and control. The External AI Service in this case is a service that the corporate client or the user procures, subscribes to, pays for, and controls itself. The Company does not select or designate such provider, and is not a party to the relationship between the corporate client or the user and that External AI Service provider.
Scope of data access. Access to data through the Connector is limited to the scope of rights determined by the corporate client (such as the scope of access credentials and authorised campaigns), and is subject to the usage rates announced by the Company.
The Company's role. The Company acts solely to provide the Connector on the instructions of the corporate client. Once data has been sent or retrieved out to the user's External AI Service, the processing, storage, and use of the data occurring within or by that service is beyond the Company's control. The Company is neither a Data Controller nor a Data Processor for the processing occurring within that service.
Personal data protection. An External AI Service provider connected by the user is a data processor or sub-processor of the corporate client, not of the Company. The corporate client, as Data Controller, is responsible for determining the lawful basis, putting in place the necessary agreements, and establishing the cross-border transfer mechanism under Section 28 of the PDPA (if applicable).

8. Limitations and Risks of AI Systems

Outputs generated by AI systems may contain errors, be incomplete, be out of date, or be unsuitable for a specific context. Users should verify the accuracy of outputs before use, particularly in cases involving business, legal, financial, medical, safety, reputational, or third-party-rights decisions.

Technical limitations

Output inaccuracy — AI systems, especially Large Language Models, may produce incorrect outputs (hallucination), outputs of limited responsiveness, or contradictory outputs.
Model bias — outputs may reflect biases present in the training data.
Training-data limitations — models may lack current data or fail to cover specific contexts.
Context limitations — models may not understand cultural context, idiom, or sarcasm in the Thai language.
Output variability — AI systems may produce different outputs for the same input, depending on the parameters and context of processing.

Social and ethical risks

Over-reliance on AI — users may accept AI outputs without verification.
Privacy impact — although the Company applies protective measures, data processing at standard levels still carries privacy risks that must be managed.
Misuse of AI outputs — users may take AI-generated outputs, such as Virtual Persona or AI Summary, and use them to disseminate distorted or misleading information. The Company prohibits such use and accepts no responsibility for any outcome arising from such acts of users.

Package-based limits on access to and use of AI features

The artificial intelligence (AI) features the Company provides within its products and services may differ in level of access, scope of use, output size, number of prompts, number of requests, volume of data processed, processing speed, or available features, according to the service package, the contractual terms, or the entitlements your corporate client has agreed with the Company. In this regard:

Certain AI features may not be available in the package you are using, or may be restricted to specific packages.
The Company may impose quotas, rate limits, or other quantitative restrictions relating to the use of AI features according to your package.
The Company reserves the right to add, improve, reduce, change, suspend, or discontinue the AI features available in each package from time to time, at the Company's discretion and as required by law, contractual terms, or the requirements of external providers.
Your ability to access any AI feature at a given time does not give rise to any claim requiring the Company to continue providing that feature within the same package, save as expressly stated in the corporate client's service agreement.

Transmission of data out to User-Connected External AI Services — When a user connects an External AI Service through the Connector, data the user retrieves out is subject to the processing and data protection measures of that external provider, which the Company cannot control. Users and corporate clients should review the privacy policy and security measures of the external provider, and must not export sensitive personal data, data they have no right to process, or data whose export would breach the requirements of the Source Platform.

9. Disclaimer

USERS AND CLIENTS OF THE COMPANY ARE OBLIGED TO USE THE COMPANY'S ARTIFICIAL INTELLIGENCE (AI) FEATURES LAWFULLY AND IN ACCORDANCE WITH CONTRACTS, TERMS OF SERVICE, AND APPLICABLE POLICIES, AND ARE OBLIGED TO ENSURE THAT THE INPUT OF ANY DATA, PROMPTS, CONTENT, FILES, OR INFORMATION INTO THE AI SYSTEMS IS AN ACT FOR WHICH THEY HAVE THE RIGHT, LAWFUL BASIS, OR DUE AUTHORISATION.

USERS AND CLIENTS MUST NOT INPUT PERSONAL DATA, SENSITIVE PERSONAL DATA, CONFIDENTIAL INFORMATION, INFORMATION PROPRIETARY TO ANOTHER PERSON, OR THIRD-PARTY DATA INTO THE AI SYSTEMS, UNLESS THEY HAVE THE RELEVANT RIGHT, LAWFUL BASIS, CONSENT, OR LAWFUL AUTHORITY TO DO SO.

OUTPUTS GENERATED BY AI SYSTEMS MAY CONTAIN ERRORS, BE INCOMPLETE, BE OUT OF DATE, OR BE UNSUITABLE FOR A SPECIFIC CONTEXT. USERS AND CLIENTS SHOULD THEREFORE VERIFY THE ACCURACY, SUITABILITY, AND COMPLETENESS OF OUTPUTS BEFORE USE, PARTICULARLY IN CASES INVOLVING BUSINESS, LEGAL, FINANCIAL, SAFETY, REPUTATIONAL, OR THIRD-PARTY-RIGHTS DECISIONS.

THE COMPANY MAY IMPOSE ADDITIONAL RESTRICTIONS, WARNINGS, USAGE CONTROLS, OR GUIDELINES ON THE USE OF AI FEATURES AS APPROPRIATE, SO THAT USE IS SAFE, TRANSPARENT, AND CONSISTENT WITH APPLICABLE LAW AND POLICIES.

FURTHER, WHERE A USER CONNECTS HIS OR HER OWN EXTERNAL AI SERVICE THROUGH THE CONNECTOR, THE COMPANY ACCEPTS NO LIABILITY WHATSOEVER FOR THE PROCESSING, STORAGE, DISCLOSURE, ACCURACY, OR SECURITY OF DATA AND OUTPUTS OCCURRING WITHIN OR BY THAT EXTERNAL AI SERVICE.

10. Availability of AI Features and Failures Attributable to External Providers

You acknowledge that some of the Company's artificial intelligence (AI) features depend on the services of Large Language Model providers and external AI providers, including but not limited to OpenAI, L.L.C., Anthropic, PBC, Google LLC, and other providers the Company may use from time to time. Where an AI feature is unavailable, malfunctions, responds slowly, produces incomplete outputs, or ceases to operate due to any of the causes listed below, the Company shall not be liable for any loss, damage, delay, lost profit, or impact arising from or in connection with the foregoing. The Company does not warrant that AI features will be continuously available, uninterrupted, or error-free.

Failures, errors, defects (bugs), or malfunctions of the systems or APIs of external AI providers.
Failure, stoppage, or unavailability of the servers or infrastructure of external AI providers (server down, outage, downtime).
Changes to, improvement of, discontinuation of, or suspension of the models, features, APIs, or services of external AI providers.
Rate limits, quotas, or other technical restrictions imposed by external AI providers.
Changes to the policies, terms of use, or contractual requirements of external AI providers.
Failures of internet networks, telecommunications systems, cloud services, or other third-party services necessary for connecting to external AI providers.
Force majeure, or causes beyond the Company's reasonable control.

Where AI features are unavailable for a significantly extended period, the Company will make reasonable efforts to notify users and to restore service as quickly as practicable. Any remedy, refund, or service credit (if any) shall be as set out in the service agreement between the Company and your corporate client, and shall not exceed what is set out in that agreement.

11. Contact Channels

If you encounter usage problems or have questions relating to the Company's services, please contact the Company's staff responsible for your organisation, or contact support@wisesight.com or any support channel appearing on the platform. For contact relating to the exercise of personal data rights, please contact dpo@wisesight.com. To make a complaint relating to the use of AI, please contact po@wisesight.com.

AI Transparency Statement