We study the use of foreign exchange (FX) intervention as an additional policy instrument in an environment with learning, where agents infer the central bank policy rules from its policy actions. Under full information, a central bank focused on stabilizing output and inflation can achieve better outcomes by using FX intervention as an additional policy tool. Under policy uncertainty, where agents perceive that monetary policy may also have exchange rate stabilization goals, the use of FX intervention entails a trade-off, reducing output volatility while increasing inflation volatility. While having an additional policy tool is always beneficial, we find that the optimal magnitude of intervention is higher in monetary policy regimes with lower uncertainty. These results indicate that the benefits of using FX intervention as an additional stabilization tool are greater in regimes where monetary policy is credibly focused on output and inflation stabilization.