You can configure how your chatbot behaves and where new chats with chatbot users start, integrate with chatbot channels so your chatbot users can interact with it. You can test your chatbot and chat widgets without leaving the inGenious AI platform and keep full control over when your content goes live.
You can also leverage generative AI to help you create and refine content and improve your chatbot experience, while keeping full control over what content is sent to your chatbot users.
Settings
The chatbot settings control important aspects of your chatbot experience, including:
- The chat experience, including typing indicators and delays between message bubbles.
- Fallback responses the chatbot should send when it can't match a chatbot user's question, or the chatbot user has done something unexpected.
- How the chatbot hands chatbot users over to your livechat agents, and how it behaves when the chatbot users are handed back.
- Other settings such as the languages your chatbot supports, how long your chat histories are retained, and your organisation's business hours.
Some settings, such as fallback responses, can be further customised for each conversation, so your chatbot can respond differently depending on the context.
Entry points
Entry points define the first passage a chatbot user sees when they start chatting with your chatbot. You can define multiple entry points to start different passages based on which URL the chatbot user came from, and configure actions for each entry point to run scripts or save variable values before the passage is started.
Entry points can help a chatbot user get what they need quickly without having to know how to navigate through chatbot menus, and make interactions feel more natural by taking advantage of the user's current context with variables.
Integrations
You can integrate your chatbot with channels such as Meta Messenger, Sunshine Conversations, Microsoft Teams, Zendesk, Khoros, analytics platforms like Mixpanel, and more. If the channel you want to use isn't available out-of-the-box, you can add a custom channel using webhooks.
Different channels have different features and limitations, but you can customise a passage to send different content for a specific channel.
Tests and logs
It's important to test your chatbot thoroughly before it goes live to your chatbot users. You can test your chatbot without leaving the inGenious AI platform by using TestBot and test pages.
TestBot lets you test your chatbot as a real chatbot user would. It uses the content from your most recent draft, so you can test any changes you've made since you last published a version. You can follow the chatbot's behind-the-scenes activity in the TestBot side panel to see when the chatbot matches questions, stores values to variables, run scripts, and performs actions.
Test pages help you test your chatbot widgets' style and function. A test page hosts a chatbot widget and your custom CSS or bot widget code, so you can see how the widget performs.
Your scripts can also send messages to your chatbot's logs so you can monitor your live chatbot for errors, warnings, or notifications about events.
Versions
Versions allow you to develop and test changes to your chatbot without impacting your chatbot users.
Each version is a snapshot of your chatbot's conversations, passages, logic, scripts, webviews, and notifications, as well as chatbot settings and some configurations like entry points, guardrails, or data source updates. Changes to your chatbot must be reviewed before they can go live so that you can test the changes thoroughly.
When you're ready to make the changes live, publish them as a new chatbot version. You can deploy newer or older versions of your chatbot as you need.
Generative AI
You can leverage powerful generative AI features to improve your chatbot experience and streamline your content creation. By connecting your chatbot to a Large Language Model (LLM), you can create LLM usages and configure LLM Tasks to help your chatbot match difficult questions or generate handover transcript summaries for your livechat agents. Generative AI features can also help you create, refine, and translate content, explore trends in your chat histories, and improve the prompts of your tasks.
The generated content is never directly visible to your chatbot users. You have full control over which generative AI features you enable, which LLM models you use, and the prompts used to generate the results. You can also create guardrails to monitor and block content being sent to and from LLM models.
You can also create a knowledge base to provide your LLM models with grounding information or additional context. You can add multiple data sources, including uploaded files, crawled websites, and even your chatbot's annotated conversations and passages.
Data stores
Data stores are custom lookup tables that you can use in your chatbot scripts to create dynamic content, map legacy product names, check batch or order numbers, provide unique promotional codes, and more. The structure of each data store and the data that it stores is completely up to you. Your chatbot can even add and remove data from a data store via your chatbot scripts, allowing it to automatically maintain the data stores in the background.