The AI FAQ Bot is designed to automatically answer customer questions by drawing on information from your knowledge base. While it uses your FAQ articles to craft responses, you can boost its accuracy and knowledge by providing external web pages as additional context. This article shows you how to get set up!
Understanding the AI FAQ Bot's Knowledge Sources
The AI FAQ Bot builds its knowledge base from two key sources: your existing Re:amaze FAQ articles and external URLs you provide. All published articles in your Re:amaze FAQ section are automatically included as a core knowledge base.
For additional context, you can add up to 5 external URLs. The bot will scrape the content directly from these single pages and integrate it into its knowledge base. This is especially useful for incorporating information that lives outside of your Re:amaze FAQ, such as product documentation, specific help guides, or relevant blog posts.
Adding URLs to the AI FAQ Bot
Navigate to your AI settings by going to Settings > Brands > AI Settings.
Scroll down to the "Single page URLs for AI context" section, where you can enter up to 5 accessible URLs. Once you're done, click "Submit."
Benefits of Using External URLs:
- More Accurate Answers: The bot has a wider range of information to draw from, resulting in more precise and relevant responses to customer inquiries.
- More Targeted Responses: With expanded context, the bot is better equipped to provide highly relevant and specific answers, reducing instances where it might otherwise offer general or unhelpful information.
- Dynamic Knowledge Base: You can easily update the bot's knowledge by simply changing the content on the linked URLs. The bot will periodically re-scrape these pages, ensuring its information stays current without needing to manually update Re:amaze FAQ articles
Important Considerations for URLs
- URLs must be public and accessible without a login. The bot cannot access content behind authentication, paywalls, or restricted networks.
- The bot will only scrape the content from the single page URL you provide. It will not crawl through links on the provided URL (no site-wide crawling).