These were the earlier attempts at decision making systems that could replace intellectual jobs.
Their main problem is that it is very costly to acquire data, which is kind of the main issue that large language models address with their ability to consume natural language input.