Portal Admin

Portal Admin is where you can configure your generated response with different setting. Which Portal Admin, you can also feed the AI with your knowledge centers and files so that the AI Actions can use it to generate the most accurate and appropriate responses for each ticket.

Manage Workspace

Each Portal Admin account can manage multiple HubSpot accounts, which will be identified by the workspace domain (HubSpot account ID). You will able to setting the response for each help desk in the Portal Admin.

To change workspace:

  1. Click on Change Workspace button

  2. A modal will be opened. Here, select the workspace that you want to change

General Tab

At the General Tab of the AI Actions application, you will find a range of settings that you can utilize to customize and enhance your experience. Some of the settings available include:

  • Tone of AI: Change your response tone.

  • Language: Select default language for your responses.

  • Auto draft on Ticket load: Start drafting a response as soon as a ticket is loaded.

  • Auto draft using Template: Enable drafting a response using a template.

  • Shopify URL: Coming soon

  • Template: Define the template for your response.

    • Example template:

Hello {first_name_of_requester},
Thank you!

Template is a sample that helps AI respond in a predefined structure. The Template can contain variable values enclosed in curly braces { }, and these values should be nouns representing attributes related to a ticket. For example: {requester_name}, {first_name_of_client}, {address}, {response}, etc. These variables can be customized.

Train AI Tab

Caffeinated CX has the capability to learn from any website or knowledge center. You can submit a URL or SLAs (PDF, CS, and Plain text files are supported) to provide additional context to your AI.

The training process will includes 3 steps:

  1. Receiving: Once you submit your data, we receive it for further processing.

  2. Training: Our system initiates the training process using the data you provided. This step involves the analysis and integration of the data to enhance the AI's understanding and performance.

  3. Done: After completing the training process, your data is ready to be utilized by the AI. The AI has assimilated the information and is now equipped to provide more accurate and insightful responses.

The training process typically takes approximately 24 hours, although the duration may vary depending on the size and complexity of the submission. Rest assured, once you submit the URLs and SLAs, our system takes care of all the necessary steps involved in the training process.

Our team will handle the data processing, analysis, and integration, ensuring that the AI receives the required information to enhance its capabilities. You can trust us to manage the training process efficiently, allowing you to focus on other aspects of your work.

During this time, we recommend checking back after 24 hours to access the trained AI, which will be ready to provide more accurate and informed responses based on the data you provided.

If you have any further questions or need assistance throughout the training process, our support team is always available to help you.

Rules Tab

In addition to integrating with the Knowledge Center, the AI Actions application provides a powerful feature that allows you to define specific business rules for the generated responses. These rules ensure that the responses align with your organization's guidelines and requirements.

For instance, if you want the generated response to include an apology when a user expresses dissatisfaction or complaint about your product, you can easily define this rule. Simply include the following line in the input: "If the user is complaining about something, apologize to the user first." Then, press Enter to save the rule.

By incorporating such business rules, you can customize the behavior of the AI-generated responses to match your desired tone, style, and specific guidelines. This flexibility empowers you to ensure that the responses generated by the AI align with your organization's standards and meet customer expectations.

Feel free to experiment with different rules and conditions to fine-tune the AI's behavior according to your requirements.

You don't need to follow any specific format for your rules; simply input them in your natural language.

Last updated