search

Knowledge Juicer


Knowledge Juicer is an AI-powered tool that helps you generate FAQ from existing web content, text files or Helpfruit articles.

Helpfruit customers can use Juicer inside the Helpfruit portal. It's available on all Helpfruit plans, including free trials.

Not yet a Helpfruit customer? You can try Juicer here

Juicer benefits

  • Quickly and easily add a new area of content to your chatbot
  • Built in match-detection prevents you from doubling up on content
  • Refresh your chat content when your website or docs are updated
  • Q&A are tagged to the source - so you know where the knowledge has come from

How to use Juicer

Use Juicer in the Helpfruit portal

  1. Log into your Helpfruit account and navigate to Juicer > Sources in the menu
  2. Add knowledge sources. You can add web pages (via URL), text files (eg docx, pdf) or even existing Helpfruit articles
  3. Select up to 10 sources to create a run
  4. Review the generated Q&A, and select the ones you want to add to your knowledge base

Update existing content with Juicer

If you have already extracted knowledge from a webpage, and have since updated the webpage, you can run Juicer again on the same source.

Just select the source/s and click New run. We have duplicate detection built in, so you can review possible matches and exclude those:

Use Juicer outside of Helpfruit

If you don't have a Helpfruit account, it's possible to use Knowledge Juicer outside of Helpfruit. You are limited to one run of up to ten sources, and can export your Q&A results as a CSV or HTML file.

If you subsequently sign up for a trial, you'll have access to your Juicer Q&A inside Helpfruit.

Knowledge sources

Knowledge sources are the web pages and documents that we can extract knowledge from and turn into Q&A. Knowledge Juicer can process three kinds of knowledge source:

  • web pages - add via URL
  • word or text-based PDF files
  • Helpfruit articles

Knowledge extraction - limitations

Juicer can't extract Q&A from image-only PDF files.

If you're not seeing good results from web knowledge sources, it may be because the services we use are struggling to extract readable content from the page. Some workarounds you could try:

  • Copy text from the page to a word document, and add that document as a Juicer source
  • Print the webpage to PDF, and upload that PDF as a Juicer source

Data security 

We use generative AI to optimise the process of chatbot building, but you remain in control of your content. Your knowledge is secure. Source content is encrypted in transit and at rest, and is not used to train AI models. 

Managing AI risks

Before publishing your Q&A or adding them to your bot, you can review the generated content for accuracy, compliance, and brand consistency. Your chatbot will not answer a question with a made-up answer - all chat content is reviewed and published by you before being available to your customers.

Plan availability

Knowledge Juicer is available to all Helpfruit customers with the following limits: 

  • Trial/starter plan - process up to 20 knowledge sources per calendar month
  • Standard plan - process up to 40 knowledge sources per calendar month 
  • Premium plan - process up to 80 knowledge sources per calendar month 

Limits will be reset at the beginning of each calendar month. Credits are not carried over.

 


GPT
Helpful?