Ever wondered why your AI Assistant responded the way it did?
If the answer is wrong, it's usually attributed to so-called “hallucinations." But what we’ve actually seen over the last 4 years of implementing conversational AI into businesses is that often the cause lies elsewhere: in messy, unorganized knowledge bases, which result in generating inaccurate or just plain wrong responses.
To address this and give you more control over your AI, we’ve added a new, widely requested feature:
The Message Sources. 🚀
Message Sources reference the exact article from your Knowledge Base that was used to generate a particular response. With huge, extensive knowledge bases it’s hard to spot the exact bit of information that influenced the AI's response and to understand its impact. Refining this data is a crucial step in improving AI performance, and we hope we've made the process a bit easier for everyone.
To use it, just hover over the lightbulb 💡 icon next to the message you want to investigate in the Inbox. Check out this short video to see how it looks in the interface:
VIDEO
It's the first step in this direction, and we're excited to already set the scene for the next version of Message Sources that's coming soon ! It will let you see the exact paragraph from your Knowledge Base that influenced the AI's response. Stay tuned.
And as always, we look forward to your feedback and ideas. Feel free to contact us using the Widget or by sending an email to contact@quickchat.ai .
.png)
.png)
.jpeg)
