Assistant for Writing a Response to a Customer

How does it work?

1. Connect OpenAi Token - to send requests and process them by the robot.

   Currently, it works through the use of a token specified in the Fast Line.

2. In an active dialogue, enter a preliminary response to the customer in the text input field.

To use the assistant, process the entered text to bring it to the required format:

1. Call the assistant by button.

1. Choose what needs to be done:
   1. Enter arbitrary text that needs to be done with the text, for example - "correct the grammar in the text" or "translate to language....".
   2. Use quick "Prompts" - these are pre-prepared phrases, what needs to be done with the entered text.

It is used so that you do not have to enter the same phrase every time. It is configured by the supervisor or administrator of the portal. Adding through the "Quick Answers" table - for the quick answer to work as a "Prompt", you need to tick the box.

When using a quick response as a "Prompt" - it will not be displayed in the list with quick answers.

1. After sending a request or selecting a request from the list with "Prompts", wait for a response from the robot to view the result. After that, you need to make a decision: replace the entered text with the text from the robot or delete it to continue working with the AI assistant.

Optionally:

There is also the possibility to add a "System Prompt" - by default it is empty. It is usually used for the robot to consider an additional parameter when generating a response. For example: "When forming a response, take into account the business theme...."

To add/change the system Prompt, you need to contact support.