Resources can be added to GenAI's knowledge in the Knowledge screen via the web crawler and via files.
Web crawler
You can upload a web page by clicking the + button, choosing resource type 'Add web-crawler', and including your website's URL(s).
- If you want to crawl multiple subpages, you can use a wildcard (*). E.g. https://desk.company.com/solutions* will crawl all subpages underneath the solutions page.
- You can also exclude certain web pages from being crawled using the Web Crawler Blacklist on the Profile Settings screen
- Note that the crawler can crawl up to 48 hours in one go
- The best source for high-quality answers is naturally written text. That’s why product text like “Color: Black” also doesn’t work.
Files
To upload a PDF or JSON file:
- Click the "+" button.
- Select "Upload file" as the resource type.
- Note: File size limit is 25 MB.
- PDFs can be in any format; they don’t need a Question/Answer layout.
- Text, lists, and tables can be read, but images are not supported yet.
- For the best knowledge extraction:
- Use text blocks between 10 to 400 words.
- Separate blocks or topics with two blank lines.
JSON
- JSON files should have the following format:
JSON Format
[
{
"trigger": "…",
"content": "…"
}
]
Please note that while it is possible to upload a file that contains multiple languages, this is not recommended, as it may lead to loss of information.
Video
Below is a video on the Knowledge and Settings sections within GenAI.