tl;dr
Machine learning is a branch of AI in which a system improves at a task by learning from data, rather than being explicitly programmed with rules for every case.
In finance, machine learning models learn what a correctly coded transaction looks like, what a fraudulent invoice pattern looks like, or what next quarter's cash position is likely to be, by training on historical examples.
The practical value is pattern recognition at a scale no person could match: spotting the invoice that doesn't match its usual vendor pattern, or the expense report with the anomaly. Machine learning underpins much of what makes modern AI accounting software more than a faster version of the old rules engine.
Back to the glossary