ChatBot
SearchAI ChatBot is designed to converse with your data and maintain context when interacting with users. Leveraging the latest Large Language Models (LLM), the chatbot transforms your existing content into a conversational AI chatbot.
With Chatbot, you'll be up and running quickly, optimizing your workflow and enhancing customer service. Say goodbye to manual organization and hello to efficiency.
SearchAI ChatBot can be accessed from SearchBlox Admin Console> Collections > SearchAI > ChatBot.
Prerequisites
- ChatBot server URL, which is up and running is required. Example :
https://xxx-yyy.com- Stop SearchBlox service, go to SearchBloxServer installation directory, go to C:\SearchBloxServer\webapps\ROOT\WEB-INF\chatbot-config.yml.
- Update the attribute value
chat-bot-server: ""with ChatBot server URL.- Example : chat-bot-server: "
https://xxx-yyy.com"- Restart SearchBlox service.
- If Prerequisites are not followed, the below error can be seen on ChatBot screen.
- Once the URL is updated, you can go to SearchBlox Admin Console > SearchAI > ChatBot.
Chatbot Creation
- To create a new ChatBot, click the
Create a new ChatBotoption or use the "+" icon, as depicted in the screenshot below.
- Provide the ChatBot Name and select the Collections on which the ChatBot needs to be trained.
- You can select upto
3collections. - The documents in the selected Collections are used to train the ChatBot.
- Then click on
Create, you can see the chatbot is training as shown below.
- Once created, you can see the
ChatBotspage as shown below.
| Field | Description |
|---|---|
| Create a New ChatBot | New ChatBot can be created using Create a New ChatBot or “+” icon. |
| Refresh ChatBot | Refreshing can be done by clicking the Refresh icon after performing any actions |
| Chat QnAs | List of all edited responses and their corresponding questions. |
| Retrain ChatBot | Use the Retrain action button to initiate a retraining process for the ChatBot. |
| Delete ChatBot | Delete action button can be used to delete the ChatBot. |
| Last trained | Last trained column shows the last trained/retrained time of the ChatBot |
| Last used | Last used column shows the time when the ChatBot was last interacted with. |
Chat QnAs
It's possible to edit individual responses and revise the answers within the conversations section. Upon saving, these edits can be stored in the chat cache.
The page displays saved conversations, their associated chatbot, and the date they were last edited.
You can swiftly add conversations to the fast cache, delete them individually, or delete them in bulk by selecting multiple questions. Additionally, buttons are available to refresh, filter, and clear all conversations.
Once a ChatBot is created, you have the following tabs:
- Preview
- Settings
- Sources
- Schedule
- Conversations
- Embed
Preview
Previewtab, allows you to preview conversations with the ChatBot.
Settings
SettingsTab allows you to configure few options.- You can change the
User and Bot Label, as well as theInitial Messagethat is displayed by the ChatBot when a conversation begins. - Save detailed logs - When activated, debug information concerning chat responses and indexing activities will be recorded.
- Save the settings by clicking
Savebutton, to see the changes onPreviewTab.
Sources
- In
SourcesTab , the Collections that were initially used to train the ChatBot, can be seen. - You can also find an option to
Retrainthe ChatBot from this tab.
Schedule
Adjust the frequency and specify the start date/time for retraining the ChatBot.
Embed
EmbedTab is used to Integrate ready-to-use ChatBot within your website or application. Place this HTML code between the header and footer section of your HTML page to setup ChatBot.View widgetwill show the Preview of the ChatBot.
Conversations
Conversationstab displays the Conversations and metrics related to conversations.- You can view the Total Conversations, Total Messages, Total UpVotes , Total DownVotes and the conversations history.
Updated 8 days ago