Chat onboarding
Overview
With web chat, you can seamlessly integrate your Ada AI Agent with your website or app. It looks like the classic web widget that most people think of when they think of an AI Agent - but the generative responses it creates set it apart.
If your use case requires displaying images to customers, you can include URLs in the AI Agent’s response that link to pages containing the visual content. Alternatively, you can configure Handoffs to route specific inquiries to human agents or support channels that support image sharing.
For every AI Agent, the first thing you have to do is connect your knowledge base to your AI Agent. For more information, see the topics in the Set up your AI Agent’s knowledge and behavior section of the Help Docs. Those topics take you through preparing and uploading your knowledge base content to your AI Agent, then customizing its behavior with more advanced features like Guidance, Processes, and Actions for any situations you need your AI Agent to handle with specific actions or content.
Limitations
The Chat channel has the following constraints:
-
Text-only responses: In the Chat channel, the AI Agent displays text-based responses only. It cannot render images, videos, GIFs, or other media formats directly in the chat widget.
-
No inline images: Even if your knowledge base articles contain images (such as product photos, diagrams, or screenshots), the AI Agent cannot display these images in the chat. Instead, it describes the relevant information in text form.
-
Order status and similar use cases: For scenarios where customers expect visual information—such as order tracking images, product thumbnails, or visual confirmations—the AI Agent provides the relevant details as text. If displaying images is essential to the customer experience, consider including links to web pages where customers can view the visual content, or route the conversation to a human agent.
Make your chat widget your own
You can customize the look and feel of your chat widget, so it fits seamlessly within your organization’s branding guidelines. For more information, see Customize your AI Agent’s branding.
Then, when you’re embedding your chat widget into your website or app, you can adapt the embed script as required to further fine-tune your customer experience.
Implement and customize your chat widget
Once the content in your AI Agent is ready, you can add a chat widget to your website or app with an embed code. You can then modify the embed code to customize elements of the widget’s behavior, like:
- Setting the default language for the chat
- Adding meta information about the customer for the AI Agent to use
For detailed information on different ways you can customize your chat widgets, see the developer documentation for the platform you’re interested in:
Want to double-check which browsers you can use to both manage your AI Agent and use your chat widget with? Check out our topic on browser compatibility.
Control your chat widget’s visibility
To start, make sure chat is enabled as a channel for your AI Agent. You can do this by ensuring that Enable chat under Chat availability on the Launch tab is turned on.
Once chat is enabled, you have control over your chat widget’s visibility. This ensures malicious parties can’t embed your AI Agent on unauthorized websites, and also gives you flexibility to roll out your AI Agent gradually.
Check out these topics for more information:
Browser support
Use the following tables to determine which actions you can perform in which browser. For example, if you’re using Firefox, you can use it to interact with Ada as a customer, but you have to switch to either Chrome or Edge to log into your dashboard.
