ParallelChain Lab is a pioneer in the research and development of voice-based age classification AI models. These models power our compliance solutions that target age-restricted businesses (such as gaming, gambling, and tobacco businesses) and age-appropriate content access.

The genesis of ParallelChain Lab’s voice AI models was the collaboration with the Personal Data Protection Commission of Singapore in 2024, in an initiative aimed at enhancing privacy in age estimation. Our sophisticated models leverage various unique characteristics of the human voice to accurately identify an individual's age group.

Our innovation is a timely one, given the increasing need for privacy-preserving age verification methods, particularly in scenarios involving young users. Importantly, our solution does not require the collection of sensitive biometric or personal data, such as images of faces and IDs.

Anti-spoofing is a critical security measure in face-based authentication, which is very vulnerable to spoofing attacks. Just consider how easily deepfakes nowadays can deceive even human eyes!

To this end, most systems rely on active anti-spoofing, which requires users to perform multiple actions in front of the camera, such as blinking or moving their heads in a specific way.

In contrast, ParallelChain Lab has developed a faster, more secure, and user-friendly passive face anti-spoofing model. This technology works silently in the background, confirming liveness in less than a second without requiring user action, providing a truly seamless user experience.

After liveness confirmation, ParallelChain Lab's face matching AI meticulously compares the user's selfie with the photo on their ID, verifying that the user is indeed the rightful owner of the presented ID.

One critical challenge in face matching technology is the dynamic nature of human faces—faces inevitably look different across time, which makes accurate identification difficult. Our model adeptly overcomes this challenge—it is able to manage various changes in appearance, including ageing, facial hair, glasses, and makeup.

ParallelChain Lab's document verification system employs a state-of-the-art anti-spoofing model, a breakthrough capability in the field of regulatory technology.

This feature allows our Know Your Customer (KYC) solution to instantly confirm whether the document presented is an original copy—a task that is challenging even for the human eye—by analysing textures, patterns, security features, and other document-specific characteristics. This provides an additional layer of security and greatly reduces the need for manual checking.

ParallelChain Lab’s ID document verification system is a universal solution for worldwide users, supporting global passports and a wide range of local IDs.

Our system employs the passport security mechanisms of the United Nations’ International Civil Aviation Organisation. On top of that, our AI models also analyse the security elements unique to different document types.

Optical Character Recognition (OCR) converts an image of a document to text. ParallelChain Lab's OCR system is purpose-built for ID verification, hence performs significantly better than other general-purpose OCR systems in the market.

Our OCR system accurately extracts and interprets information from IDs, even in conditions such as poor lighting and low image quality. Additionally, ParallelChain Lab’s KYC solution eliminates the need for manual data entry, which significantly enhances user experience and streamlines the verification process.

Our Large Language Models (LLMs) models are at the heart of MailCat, a customer relationship management (CRM) solution.

Traditional CRM systems rely on keyword matching or pre-defined scripts, leading to generic responses that are inappropriate in context. In contrast, our LLM models enable MailCat to understand the nuances and intents in natural language. With this, MailCat can automatically categorise, delegate, and respond to customer enquiries with remarkable context sensitivity.

While our LLM models lay the groundwork by understanding human language, our in-house Generative AI models take this understanding to the next level by producing contextually relevant and personalised responses.

The design philosophy of ParallelChain Lab’s Generative AI adopts a business-native approach, which manages the risks inherent to using AI to communicate directly with customers.

This approach integrates business-specific and ethical protocols into the system’s operational parameters.