DeepSeek integrated into WeLab Bank's system

DeepSeek has been integrated into WeLab Bank’s local system, saving time and costs in writing programs and developing new products, starting from tomorrow.
Tat Lee Ka-tat, the chief executive of the Hong Kong-based virtual bank, said it achieved a breakeven point in a single month last December and he has strong confidence in achieving profitability in the first half of this year.

AI application in Hong Kong's banking industry is "not progressing fast enough"

Lee pointed out that the use of large models like ChatGPT in Hong Kong's banking industry is actually quite limited. Accessing them through VPNs can lead to network instability and comes with high costs, with annual fees reaching millions of US dollars. Even if applied, they may not be "fast enough." However, DeepSeek is the only "open-source large model" that can be integrated into local systems, creating a closed-loop environment that ensures information security and prevents data leaks. At the same time, it allows for the creation of more tailored content based on the company's needs and policies, reflecting a more distinct corporate style.

Aspiring to develop AI agents

He mentioned that the bank aims to create AI agents to serve customers and will make an official announcement once the product is ready. Additionally, he emphasized that the use of AI will not lead to staff reductions, as employees can shift their focus to more creative tasks. He joked that it’s not about AI replacing humans, but rather "those who know how to use AI replacing those who don’t."