TAIPEI (Taiwan News) — NexRetail has developed an AI system that analyzes store operations, providing insights for merchandising while preserving privacy.
The technology was born from NexRetail CEO Wu Cheng-han’s (吳承翰) early research in neural networks at UC San Diego and defense communication systems, per CNA. His MBA thesis at National Taiwan Normal University inspired the core idea: predicting consumer behavior more accurately while improving efficiency.
NexRetail’s system uses cameras that do not capture faces but analyze age, gender, and behavior, generating insights for product shelving and staffing. The AI combines image recognition and natural language processing, creating what Wu calls “storytelling through images.”
A 2019 TechCrunch Disrupt showcase helped the startup gain attention after its AI identified a theft at a South Korean booth without revealing personal identities. The incident secured LG as the company’s first client and demonstrated the system’s value.
The COVID-19 pandemic accelerated NexRetail’s growth, allowing brands to measure which products and endorsements resonate with consumers, improving digital advertising strategies.
NexRetail also developed human-like AI clerks for kiosks, helping customers find their products. Running on Qualcomm’s ARM-based Snapdragon 8550 chip, the system operates offline at a fraction of the cost of cloud-based AI like ChatGPT.
Wu credited government support for helping NexRetail expand overseas. He urged startups to move beyond Taiwan’s contract manufacturing mindset and build trust with clients.
“It’s a long road, but collaboration between industry and government can make AI solutions truly impactful,” he said.
