Smart reply

Smart reply is an AI-powered feature designed to assist support agents in responding to customer requests more efficiently. By analyzing the issue’s details, including its summary, description, and comments, smart reply suggests up to 3 potential responses based on your existing Canned Responses templates.

How to use smart reply

Smart reply button is enabled by default. Unless the Jira admin disables it, each user will see the button on the issue view. To start using the feature, every user must approve the consent form.

Follow these steps to start using smart reply:

1. Navigate to the Jira Service Management issue you want to respond to.

2. Find the Smart reply button within the Canned Responses Pro Templates section. Click the button to initiate the process.

   Open

   

3. Enable this feature for your user by clicking Start.

   Open

   

4. Click the Smart reply button again to trigger AI to analyze the issue’s summary, description, and comments and provide up to 3 potential response options. Use the chevrons to switch between options.

5. (Optional) Rate the AI’s suggestions by giving a thumbs up or thumbs down. Your feedback helps improve future responses.

6. Once you're satisfied with the response, click Select to apply the template to the text box. This replaces the text in the editor.

7. Customize the text if needed and reply to a customer.

How to manage smart reply

Jira admins have full control over enabling or disabling the smart reply feature across the organization. If you're a Jira admin and want to disable the feature:

1. Navigate to Apps > Canned Responses Pro.

2. From the side menu, open Settings.

3. From the Global settings section, turn off the Enable smart reply toggle.

How it works

We take data security seriously and we are committed to being transparent about how the smart reply feature utilizes AI.

• AI-powered response suggestions: We leverage technology from Microsoft’s Azure OpenAI Service to provide you with this service. The data sent to the Azure OpenAI service includes the text from issue summaries, descriptions, and comments, along with the embeddings of Canned Responses templates from your instance. Embeddings are numerical representations of text or documents. For more information, check out an example of an embedding from OpenAI.

• No training on your data: Appfire or Microsoft does not store or use your data to train AI models. For more information, refer to Appfire’s AI consent form.

• Data protection: All data is encrypted during transmission, and individual customer accounts are kept securely separate in our production environment. We ensure that data from different customers is never mixed or processed together during AI operations.

• Monthly credits: The usage of AI-powered responses is governed by a monthly credit system:

  • Users with the Standard plan have 100 monthly credits for Smart Reply.

  • Users with the Advanced plan have 1,000 monthly credits, allowing for expanded use of the feature.

  If your credits run out, the feature will be temporarily disabled until the next renewal period. For more information about your credit allocation or to explore increasing your limit, please contact our Support team.

• Permissions: We respect user’s permissions. The LLM (Large language model) used to generate AI responses for a user cannot see or use any information to which that user does not already have access. The visibility scope of templates is also taken into account while generating responses.

• Control: Jira administrators have full control over the feature, enabling or disabling it as needed through the Canned Responses Pro app settings.

Keep in mind that content quality may vary because of the nature of AI-powered response generation. AI models can sometimes behave in inaccurate, incomplete, or unreliable ways.

We hope that you find this feature helpful to your support workflow. We appreciate your feedback and suggestions on how to improve it.