Throughout the history of automated reasoning, mathematics has been viewed as a prototypical domain of application. It is therefore surprising that the technology has had almost no impact on mathematics to date and plays almost no role in the subject today. This article presents an optimistic view that the situation is about to change. It describes some recent developments in the Lean programming language and proof assistant that support this optimism, and it reflects on the role that automated reasoning can and should play in mathematics in the years to come.