They're supposed to be good a transformation tasks. Language translation, create x in the style of y, replicate a pattern, etc. LLMs are outstandingly good at language transformer tasks.
Using an llm as a fact generating chatbot is actually a misuse. But they were trained on such a large dataset and have such a large number of parameters (175 billion!?) that they passably perform in that role... which is, at its core, to fill in a call+response pattern in a conversation.
At a fundamental level it will never ever generate factually correct answers 100% of the time. That it generates correct answers > 50% of the time is actually quite a marvel.
This is the correct way to manage a budget.
The budget software I use ("you need a budget") follows a similar philosophy; based on the way you describe you financial circumstances, paying for that software is probably not in your budget, but I suspect you will find the "rules"/method/philosophy described on their website helpful.