Two different questions about correctness
Precision and recall each ask a distinct question about a classifier's positive predictions. Precision asks, of everything the model flagged as positive, how much was actually positive. If a fraud model flags 100 transactions and 80 of them are truly fraudulent, precision is 80%, the other 20 were false alarms. Recall asks a different question entirely, of everything that was actually positive, how much did the model actually catch. If there were 200 real fraud cases total and the model only caught 80 of them, recall is 40%, meaning 120 real fraud cases slipped through undetected.
Notice that both numbers can look reasonable in isolation while telling very different stories about the model's behavior. A model with high precision but low recall is cautious, it rarely cries wolf but misses a lot of real cases. A model with high recall but low precision is aggressive, it catches almost everything real but buries you in false alarms along the way.