What should a company configure to automate AI-generated responses while allowing agent review?

Configuring Einstein Service Replies with manual approval enabled is the optimal choice for a company looking to automate AI-generated responses while still allowing for agent review. This feature enables the use of AI to generate responses to customer inquiries, which can help streamline support operations and reduce response times. However, by incorporating a manual approval step, it ensures that a human agent can review and modify the AI-generated responses as necessary before they are sent to the customer. This approach effectively combines the efficiency of AI with the nuanced understanding that human agents bring to customer interactions, ultimately enhancing the quality of service.

The other options do not meet the criteria of balancing automation with the need for human oversight. Fully automated AI-generated responses without human oversight would leave no room for adjustments or corrections by agents, potentially leading to misunderstandings or incorrect information being sent to customers. Only manually written responses would eliminate the benefits of automation entirely, likely increasing response times and work volume for agents. Disabling AI-generated responses altogether takes away the advantages of leveraging AI technology, which can significantly enhance customer service efficiency and effectiveness.