Blip Copilot

Index:

1. Introduction
2. How Blip Copilot works
3. Smart Summaries
4. Suggested answers
5. Activation
6. Settings
7. Knowledge base
8. Using Copilot with multiple contexts
9. Advanced Settings
10. Blip Copilot Reports

 

Introduction

Blip copilot is a generative artificial intelligence assistant (LLMs) that helps agents with customer service, whether for customer service or sales operations, delivering context about the consumer and complete and accurate responses based on the context of the conversation.

 

How is the Blip Copilot?

Blip copilot has three functionalities, each with a specific purpose, namely:

• Smart summaries;
• Suggested answers;
• Speech to text (audio to text transcription)

Below we will see how each of them works.

 

Smart Summaries

The intelligent summary functionality was created to provide quick context about the consumer to the attendant.

To deliver these contexts, generative AI analyzes this consumer's conversation history with the intelligent contact (pre-transshipment and human service) and generates a summarized context and delivers it to the attendant as soon as he opens this consumer's ticket.

The summary provides information such as:

• Data provided by the customer to the bot or in previous conversations;
• Reason for current contact;
• Number of calls from this consumer in the last 30 days;
• Feeling of the conversation.

The summary is automatically generated for every ticket opened by the agent that has the functionality activated.

A summary is also generated at the end of the ticket, displaying a summary of this last service and the final sentimentalization of the conversation.

 

Suggested answers

The response suggestion functionality was created to help the agent respond to the consumer in complex contexts and subjects that are unknown or new to them.

To deliver the suggested answers, the AI ​​analyzes the context of the conversation and the content of the knowledge base previously configured with the company's context and then gives the agent who requested the suggested answer two options so that he can choose the one that makes the most sense. with the moment.

The functionality of suggestion of answers delivers responses on demand, that is, for it to generate suggestions it is necessary for the attendant to request an answer by clicking on the button available on your Blip desk.

Before sending the response, the agent can make edits and customizations as deemed necessary.

 

Activation

Release of Blip copilot no contract

As it is a product sold separately, to activate it it is necessary to purchase it. After hiring, fill out the activation form available here. 

This form must be completed by Blip people (KAMs, CSMs, AMs, etc.) responsible for the customer's account.

In this form you will need to fill in the following information:

• Enterprise
• Deal win date
• Company ID on the platform
• Company ID on Hubspot
• Link do Deal no hubspot
• KAM/AM responsible
• Responsible CSM
• Context of use
• Email of the company manager
• Emails from service managers
• Functionality to be released
• Emails from agents who will use the functionality

Smart summaries and suggested answers

In the Smart Summaries and Response Suggested features, the manager or person with access to attendant settings on the Portal can enable or disable attendants whenever necessary, without the need to request Blip.

To do this, follow the steps below:

1. Access your portal;
2. Select the contract containing your chatbot(s); (1 in the image)
3. Select the BLip Copilot Menu (2 in the image)
4. Then choose the chatbot you want to configure; (3 in the image)
5. Then access the “Service” menu
6. Go to the “Attendants” side menu;
7. Select the attendants whose permissions you want to change;
8. Then in the right corner click on the icon of the person with the key;
9. The permissions screen will be displayed, adjust the Answer suggestion and Smart summaries permissions as needed,
10. Then click back on the icon next to the word permissions;
11. And that's it, your attendant settings have been adjusted.

 

Settings

Suggested answers

 

(The video may contain some differences in the interface, as we are always updating the product)

Only the response suggestions functionality requires configuration, and it must be configured per BOT, that is, each BOT must go through the following configuration process:

12. Access your portal;
13. Select the contract containing your chatbot(s);
14. Then choose the chatbot you want to configure;
15. Then access the “Service” menu
16. And then the Blip Copilot menu;
17. On this screen, click on the selector (switch) to enable Blip copilot on this bot;
18. Once enabled, fill in the form data according to the table below;

 

19. After filling in the customer information, click save.

Good practices

The form fields are textual fields, so pay attention to the correct spelling of the answers. In the field “Company Name", Enter only the company name with the initial letters capitalized. In the field “Demand(s) for service at Blip Desk”, the answer needs to be objective and direct.

If it is necessary to add new instructions in the field "Additional guidance", here are some best practices:

• Instructions need to be clear and concise:  Formulate instructions directly, avoiding ambiguities and unnecessary complexities.
• Limit the scope of responses:  If necessary, instruct the copilot that suggested answers need to be restricted to a particular domain, topic, or context to avoid out-of-scope responses.
• Avoid entering long instructions: The copilot tends to work best with short, direct instructions, so avoid including unnecessary information.

After filling in the fields click on the button "Save" for the changes to be saved.

 

Knowledge base

So that the Blip Copilot To be able to suggest accurate answers to agents, it is essential that the knowledge base is well fed with relevant information. This base serves as a source of data so that the copilot can formulate suggested responses so that human attendants can resolve the demands that arrive at the Blip Desk.

Therefore, it is important to include detailed information about products and services that are handled by agents, such as benefits, prices, technical specifications and other relevant information. 

The base can be sent in PDF, XLS, TXT and TSV formats and the document structure must be organized so that the CONTEXT NAME (always in capital letters) | Texts that represent such context are on the same line.

It is important that the assembly and categorization of the base is carried out by someone who knows well the main demands that human attendants receive during transshipment.

Below, we have included two examples of how the base should be structured.

Note that in the ‘Model in Plain Text’ it follows the pattern: CONTEXT NAME (always in capital letters) | And texts that represent such context in the same line. The database in question has 6 lines and each line represents 1 context.

In the 'Spreadsheet Template' it will follow the same writing pattern but separated into two columns, where in the left column the NAME OF THE CONTEXT must be filled in (always in capital letters), and in the right column texts that represent that context in the same line.

Template in plain text

Insert image 01

Access the model in .txt from knowledge base here.

Spreadsheet model

Access the .tsv model from knowledge base here.

 

 

Submitting the knowledge base

After ensuring that the base has all the necessary information, upload the file in the Knowledge Base section. Select the file xls, .txt ou .tsv that contains the knowledge base and then click “Upload”.
Optionally, a PDF file can be used to train the knowledge base. Using a PDF file, the AI ​​will search this file for relevant excerpts of knowledge and will itself create a list of content that will be used to suggest answers. This option requires more testing and calibration of the terms found for better operation.

It is important to note that to add, change or remove any information from the knowledge base, it is necessary to make these changes locally in the file and then upload that file. This new file will be the new knowledge base, as the previous .xls, .txt, or .tsv file will be replaced. Furthermore, the base content must be textual only.

Using Copilot with multiple contexts

In the case of subjects related to multiple products inserted in the knowledge base, it is It is necessary to have the data in the following format: CONTEXT NAME (uppercase) |  CONTEXT TAG | Texts that represent such context are on the same line.

Example: PRODUCT REGISTRATION (capital letter) | SUBCONTEXT (capital letter)

| Texts that represent such context

 

 

Messages welcome

Automatically including the attendant's name

The next step is to configure the initial messages, which are sent by the human attendant to the end consumer after transshipment. By default, two messages are added, but you can change them, or add more, or It is also possible to have no initial message after the transfer so that Copilot's first suggested response is regarding the user's last pre-transshipment message.

When adding an initial message, you can use ${NAME} and ${LASTNAME} in the message text to automatically get the attendant's first and last name in the welcome message after overflow.

Example:

To configure, access the “Initial messages” section

The maximum recommended quantity is up to 3 messages.

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

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

For the changes to be saved, after making adjustments to the initial messages, click "Save".

Observation: Initial messages will be suggested when the Desk conversation only contains messages from the customer.

 

 

 

Advanced Settings

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

In advanced settings you can change the temperature of the generative AI.

The temperature controls the degree of precision of the responses suggested by the copilot. A low value indicates that the answers generated by the model will be more accurate and predictable, and will be 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, and may deviate somewhat from the scope of the information contained in the base.

This value can be set between 0.1 and 0.5.

ATTENTION: The default value is 0.2, change this setting only if you are sure it will not interfere with the quality of the answers suggested by Copilot.

If you've made it this far, you've completed configuring Blip copilot and can now start using it in your operation.

 

Blip Copilot Reports

To access the metrics page, go to the Copilot Reports menu.

This is where you will be able to optimize the use of your Copilot, monitoring the main indicators of use of the tool, such as adoption by agents, tickets that were assisted with Copilot, difference in service times with and without Copilot, in addition to the editing rate of suggested answers. 

 

 

On this page, you can also download Blip Copilot metrics, which include suggested and selected responses, as well as a summary of activities that occurred over a period of time. up to 60 days prior to the current date.

To download usage metrics, simply select a time range (up to a maximum of 60 days prior to the current date) and click “Generate metrics”, shortly afterwards an .xlsx file containing all the information will be downloaded onto your machine.

 

For more information, access the discussion on the subject at our community or the videos on our canal. 😃