Detecting fraud worth $1 billion is a really cool use case from the Treasury Department that leverages the power of machine learning to identify patterns and anomalies. You don't have to be the federal government to harness this power. There are tools available now for finance teams to leverage their data to identify trends and spot anomalies early. These tools could help with multiple use cases in the finance world, including detecting potential fraud, identifying out-of-policy transactions quickly, or predicting cash flow. 

It's exciting to see even the slowest technology adopters start to experiment with use cases that provide tangible value (although $1 billion in value won't always be the case). This isn't something to think about in the future. The time is now to be curious and experiment.