If we were tasked with creating a system using LLM to enhance a bank’s decision-making regarding risk assessment or product approval for its clients, the initial thought might be to provide the LLM with client and product information, enabling it to make decisions on whether to grant the product to the client or not.
But let me tell you, this wouldn’t work. We can’t solve everything by providing information to an LLM and expecting it to generate an output based on the input data. ...