Yesterday morning we encountered an issue with OpenAI that prevented our requests from completing. Our engineers picked it up quickly and rerouted automations to our Azure models. While that switch completed you may have seen a short spike in errors and a few more handoffs to your agents than usual. It was contained and resolved yesterday, and nothing is needed from you. Sharing for visibility in case you see a small bump in tickets today. In response, we have added fallback handling for this error were we to encounter again in future, effectively expediting the same migration process.