Technology drives changes, and ChatCampaign Research Lab is dedicated to be part of it

Through our consistent dedication on innovation and research, we have made significant deep tech R&D progress since 2017

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
or send an email to info@chatcampaign.io

Our History

HKSTP

Incubated by Hong Kong Science Park

HKAI LAB

Incubated by HKAI LAB (funded by Alibaba Hong Kong Entrepreneurs Fund and SenseTime)

HK GOV

Received R&D grants (RTH) from the Hong Kong government on generative AI R&D

AI Agent

Build flexible workflow system AI Agent to facilitate AI automation with accuracy modulation

Research Journey

Mockup

HKSTP

Incubated by Hong Kong Science Park (2019)
Mockup

HKAI Lab

Incubated by HKAI LAB (funded by Alibaba Hong Kong Entrepreneurs Fund and SenseTime) (2021)
Mockup

HK GOV

Received R&D grants (RTH) from the Hong Kong government on generative AI R&D (2021)
Mockup

ChatCampaign Research

Total accuracy modulation applying on AI Agent system, through multi-model data distillation with checking, and more researches going on (2025)

About Us

ChatCampaign was founded in 2017, headquartered in Hong Kong. We support our clients through cutting edge workflow AI and customer engagement technology. We focus on growing the value of their processes and workflows at internal engagement and customer engagement, by making wise use of AI with real-time visualisation of complex data and information.

Since 2021, we expanded our effort from a single SaaS product, to a diverse of AI R&D projects, with deep tech research funds supported by Hong Kong Government. Our research focus includes affective computing, generative AI on body language, AI Agent system accuracy modulation and various AI experiments.

Trusted by reputable companies and growing
Product

ChatCampaign AI Workflow

No-code workflow platform, with built-in CRM. Can quickly deploy AI + RPA, for internal processes, or customer engagement. Dedicate for chat type interactions.

Especially handy - bringing offline customers' data back to online database, for re-engagement and cross selling.
Can manage and capture lifetime value of end customers, through fully automatic customer engagement.
All data are automatically store in tabular format, easy for excel export on any user response/actions
Seamless interaction with user through chatbot automation, and switch to human agent when customised conditions met
Flexible integrations with any external systems include but not limited to external CRM, LLM (e.g. GPT, DeepSeek), OCR, etc.
Dashboard mockup
Product

Navatar - Live Avatar • Video Call

Navatar, is a web-based application that revolutionizes video calls by protecting the identity of participants and allowing users to express themselves through avatars that capture their real-time facial expressions.

With its advanced technology, Navatar is transforming the way people connect and communicate online.
Connects people by allowing them to use a virtual identity - a real time animated 1:1 unique avatar, to socialize and keep their real identities private
Allows users to fully express themselves through their avatars in video calls and video social space, which captures users’ real time facial expression to animate the avatars
Empowered by computer vision technology on facial landmarks capturing.
Dashboard mockup
Product

AI Agent HQ - with accuracy modulation

AI Agent workflow system with customisable total accuracy modulation through multi-model data distillation with checking, combining with flexible API modules to grab data from different source, and data management modules.

It provided different enterprise solutions which cover any automation tasks required by enterprises.
Drag and drop interface to quickly deploy different accuracy modulation strategies and underneath settings, for the most accurate content generation
Simple automation tasks which needs accuracy modulation, can be done by ChatCampaign - Workflow alone
Complex automation tasks which needs multi-layer and multi-model data distillation with checking, can be quickly test run and deployed through AI Agent HQ
Our blog

AI Experiments

Exploring the Uncharted Limits of AI
AI Experiment

Exploring Manus AI: Revolutionizing Research with Autonomous AI Agents

In the age of rapidly advancing artificial intelligence, Manus AI stands out as a groundbreaking system, redefining how research and data analysis are conducted. Recently, we tested Manus AI's potential with two experiments: one analyzing the economic impact of tariffs on Hong Kong and another examining a significant market shift in the Hang Seng Index. These experiments showcased how an AI Assistant look like
AI Experiment

Pushing the Limits of AI: Reflections on Vibe Coding and a New Minigame Experiment

In the ever-evolving landscape of AI research and development, we’re continually seeking tools and methodologies that not only streamline processes but also amplify creativity. Today, I’d like to share some thoughts on the potentials and limitations of AI development tools, specifically through my recent experiment: creating a minigame called 《Mind Control Simulator》 using the Poe App Creator.
Our blog

AI Articles

Explore cutting-edge insights into AI advancements, methodologies, and real-world applications.
Exploring the Architecture Behind Google’s Gemini Diffusion

Although its design is undisclosed, hypotheses like Latent Diffusion Models (LDMs) and task-specific distillation provide potential insights into its efficiency.

READ Article
Unveiling the Creation of Large Language Models: DeepSeek-R1 Technical Document Overview

Discover how the DeepSeek-R1 series uses reinforcement learning, multi-stage training, and model distillation to advance reasoning in large language models (LLMs). Explore key innovations and insights from its technical document.

READ Article
The Ethical Dilemma of AI's Persuasive Power – A Controversial Reddit Experiment

AI bots on Reddit changed user opinions using dark persuasion techniques. Methods included impersonation, fake data, and personalized replies. Raises ethical concerns about manipulation and deception.

READ Article
WHY WE ARE DOING THIS

Founder's words

I once bought a EEG device and the first thing it does is to send my brainwave info to the internet. Luckily my firewall blocked it that time. Because its just brainwave band signal strength data, I don't think it takes cloud computing, a local simple program should do. Intuitively I never tried any EEG device that I don't know how to control it's hardware since then.

I'm thinking what all these tech advancement means to us. We have software and hardware so advance that our bio-data can be easily accessible(not just by the owner). The AIs are becoming more intelligent, accepting more data, faster, more powerful, every second. And there is human brain cells being put into a chip that act like computer.

With all these technology going so fast, we know that those who adapt and find a new equilibrium point earlier, can have a greater influence that favour their causes (our stance: privacy for the user). That's why we keep working on it, and keep sharing on it

Claudia is also the founder of ScienceX deep tech knowledge platform (in collaboration with The Chinese University of Hong Kong)