Blip Copilot Extension – Blip | Blip Help

Index:

Demonstration video

This extension is part of a closed beta and is not available to the entire user base. Features and the business model are subject to change. The new feature is not yet available in other languages.

Introduction

Blip Copilot is an extension that optimizes customer support on Blip Desk. With this extension, support agents have access to personalized suggested responses generated by a language model (generative AI) that takes into account the company's configured knowledge base and the conversation context to propose a response. Thus, Blip Copilot enables support agents to quickly and efficiently respond to customer messages without spending time searching for or formulating an appropriate response. This tool helps improve the quality and/or speed of human assistance for Intelligent Contacts using Blip Desk.

For example:

During a customer interaction, a support agent can request assistance from Copilot using the "Response suggestion" button. These suggestions are generated based on the conversation context and the knowledge base provided by the company using this extension. It is up to the support agent to validate the suggested responses and select the most appropriate one, either by editing it or sending it as is to the customer.

How does it work?

Blip Copilot functions as a co-pilot for support agents using Blip Desk for customer support. Blip Copilot offers the following features:

Suggested responses based on conversation context and the knowledge base provided by the company using the extension.
Supported channels: Blip Desk.

Installation and Configuration

When activating the Blip Copilot extension through the Blip Store, it should be configured in the overflow bot. To do this, send an email with the contract and the bot identifier where the extension will be installed.

Once this is done, our team will grant access to the extension, and it will be possible to configure the plugin in the bot to make the necessary adjustments for the Blip Copilot to function.

Configuring the Plugin

To configure the plugin, click on the gear icon in the upper right corner to access the settings. Then, go to "Advanced Settings" and click "Continue."

To create a new configuration, click on "Add."

Add a new configuration with the following values:

Domínio: postmaster@portal.blip.ai

Chave: Plugins

Valor: { "Copilot": {"name": "Configurações Blip Copilot", "url": "https://secure-backend-api.stilingue.com.br/blip-desk-gpt-ui/prod"}}

For example:

After adding the above configuration, refresh the page to make the plugin option available.

Go to the three dots in the upper right corner of the screen, and the "Blip Copilot Settings" button will appear for you.

Access "Blip Copilot Settings" to configure your company's assistance copilot.

Setting up Blip Copilot

After configuring and accessing the plugin, you will be redirected to the main page. Carefully fill out the form to configure Blip Copilot. It is of utmost importance that the information added to this form is correct to ensure the proper functioning of the copilot.

Variable	Definition	Mandatory
Nome da empresa (Company name)	Refers to the name of the organization that has contracted Blip's services. This variable will define the scope of response suggestions provided by Copilot, which will assume the role of the company's support agent.	Mandatory
Demanda(s) de atendimento no Blip Desk (Blip Desk support requests)	Specific requests that are directed to human agents when the bot is unable to provide adequate responses. For example, if the company is a chatbot company, a support request that may trigger a transfer to Blip Desk could be "chatbot sales".	Mandatory
Orientações adicionais (Additional guidelines)	This field allows the user to add relevant information for the behavior of Copilot. These instructions are used to assist Copilot in providing more suitable response suggestions for the company's demands.	Optional
Variable: Perfil (Profile)	Refers to the profile or tone of the response suggestions offered by Copilot. There are three options available: 1.Amigável (Friendly): Copilot adopts an empathetic and friendly tone, providing response suggestions that are welcoming and understanding. 2.Divertido (Fun): Copilot uses a slightly more informal tone, and may even include emojis to make the response suggestions more relaxed and entertaining. 3.Técnico (Technical): Copilot adopts a more specialized and technical tone, providing precise, concise, and detailed response suggestions.	Optional

Best Practices

The form fields are text fields, so please pay attention to the correct spelling of the responses. In the "Nome da Empresa" (Company Name) field, only enter the company name with the initial letters capitalized. In the "Demanda(s) de atendimento no Blip Desk" (Blip Desk support requests) field, provide a concise and direct response.

If you need to add additional instructions in the "Orientações adicionais" (Additional guidelines) field, here are some recommended best practices:

Clear and concise instructions: Formulate instructions in a straightforward manner, avoiding ambiguity and unnecessary complexity.

Limit the scope of responses: If necessary, instruct Copilot to provide response suggestions within a specific domain, topic, or context to avoid out-of-scope responses.
Avoid lengthy instructions: Copilot tends to work better with direct and short instructions, so avoid including unnecessary information.

After filling out the fields, click the "Save" button to save the changes.

Preparing the Knowledge Base:

For Blip Copilot to be able to suggest accurate responses to agents, it is essential that the knowledge base is well-stocked with relevant information. This base serves as a data source for the copilot to formulate response suggestions for human agents to address the demands that come to Blip Desk.

Therefore, it is important to include detailed information about products and services handled by agents, such as benefits, prices, technical specifications, and other relevant information. The base must be sent in .txt format, and the document's structure must be organized so that the NAME OF THE CONTEXT (in caps) | Texts representing that context are on the same line. It is important that the assembly and categorization of the base be done by someone who knows well what the main demands that human agents receive in the overflow are.

Below, we have included an example of how the base should be structured. Note that it follows the pattern NAME OF THE CONTEXT (in caps) | Texts representing that context on the same line. The base in question has 6 lines, and each line represents 1 context.

Knowledge Base Checklist

Check if the base contains detailed information about products and services handled by agents, such as benefits, prices, technical specifications, and other relevant information.
Check if the base is categorized and follows the following pattern:
- NAME OF THE CONTEXT (in caps) | Texts representing that context (both on the SAME LINE).
Verify that the knowledge base document is in .txt format.

Access the .txt of the knowledge base here.

Uploading the Database:

After ensuring that the base contains all the necessary information, upload it. Select the .txt file containing the knowledge base and then click "Upload."

It is important to note that to add, modify, or remove any information from the knowledge base, changes must be made locally in the .txt file and then uploaded. This new file will be the new knowledge base, as the previous .txt file will be replaced. Furthermore, the content of the base must be textual only; the copilot cannot extract information from images, links, videos, and the like.

Initial Messages

The screenshot below shows how to configure initial messages, which are sent by the human agent to the customer after the overflow. By default, two messages are added, but it is possible to modify them or add more.

The minimum number of initial messages is 1, and the recommended maximum is up to 3 messages.

To add a new message, fill out the field with the message, and then click "Add message."

To delete, simply click on the trash can icon located next to the message.

To save the changes after making adjustments to the initial messages, click "Save."

Note: Initial messages will be suggested when the conversation in Desk has only customer messages.

Advanced Settings

To access advanced settings, simply select the "Advanced Settings" tab.

ATTENTION: This is the default value (0.2); change this setting only if you are sure it will not interfere with the quality of the suggestions provided by Copilot.

Understanding the "Temperature" Parameter Better:

The temperature controls the degree of precision of the responses suggested by the copilot. A low value indicates that the responses generated by the model will be more accurate and predictable, being more faithful to the information contained in the knowledge base. A high value indicates that the copilot will be more likely to generate creative and unpredictable responses, possibly deviating a bit from the scope of the information contained in the base.

Blip Copilot Metrics Page:

To access the metrics page, go to the left-hand side menu and click on "Export Reports."

On this page, you can download Blip Copilot metrics, including suggested and selected responses, as well as a summary of activities that occurred within a period of up to 60 days before the current date.

To download usage metrics, select a time interval (up to a maximum of 60 days prior to the current date) and click "Generate Metrics." Afterward, a .xlsx file containing all the information will be downloaded to your machine.

Enabling the Blip Copilot Feature for Agents:

After completing the entire configuration process and uploading the knowledge base, following the instructions above, send an email with a list of the names and email addresses of the agents who will be using Blip Copilot. Once this list is sent, we will enable the copilot feature for these individuals.

Usage examples

Once properly configured and the feature is enabled for agents, Blip Copilot will be ready for use. The "Response Suggestion" button will appear in the chat of the agents. As soon as the button becomes available, agents can simply click on it to get response suggestions in Blip Desk.

Below is a screenshot of how this button will appear in Blip Desk.

Note: If the response suggestion button does not appear, please contact the responsible team to request the activation of Blip Copilot, or if there is no responsible team, send an email.

Initial messages

The initial messages, initially configured, will be suggested after the transfer to human support.

Response suggestions

When clicking on the "Suggest Response" button, two response options will appear for the support agent. Simply verify if any of them answer the customer's message and select the appropriate option. If the support agent needs to edit any information in the suggested response or only wants to use a part of the text, they can select the respective response and then edit it.

Below are some screenshots that illustrate how response suggestions can be used, as well as a link to a demonstration gif.

Example: Response suggestions

Example: Selected and edited response

For example:

Support

If you have any questions or encounter any issues with the extension, please contact us at support.

For more information, visit the discussion on the subject in our community or the videos on our channel.

Related articles