We show that a monetary policy in which the central bank commits to a randomized inflation target allows for potentially faster-expectations convergence than with a fixed target. The randomized target achieves faster convergence in particular in transition environments: those demonstrating either particularly high or low inflation.