Reimagine financial service experience with LLM
summary
Generative LLMs promise conversational, personalized financial experiences but require hard engineering and big money: China already hosts 40+ LLMs, ICBC deploys 10 digital personas, and Xiaomi plans an offline LLM on phones. Running pretrained models costs upfront hardware in the hundreds of millions to billions and electricity about 700 million per year in some locations. Practical builds include an open‑source LLM assistant for a bank call center that measurably raises success rates (ongoing), while GPT‑4/5 multimodal capabilities drive the push to rethink product/agent design.
AI-written summary · about
chapters (13)
- 0:00Financial service AI intro
- 1:09AI 2.0 defined
- 5:01GPT beyond search
- 7:08AI evolution reflection
- 11:21AI breakthroughs
- 13:56AI 2.0 strategies
- 19:19LLM in finance
- 23:37Human-centered AI
- 28:11Big data mistakes
- 30:45AI teams
- 32:12LLM security threats
- 33:39AI predictions
- 34:24Call to action
related talks
10mThe Impact of LLMs on the Circular Economy and Beyond
12mKeynote: How LLMs Have Changed the Way Organizations Do AI- 12mKeynote: How LLMs Have Changed the Way Organizations Do AI
47mOwn Your LLM: Running Generative and Agentic AI Locally
46mBuilding an LLM Infrastructure from Scratch: Lessons from X5 Tech
36mLLMs and AI agents demystified
40mKeynote
30mLLMs: Beyond the Hype - A Practical Journey to Scale