Knowledge Base

Overview

• Organize with Folders: Keep your knowledge base structured by creating folders. You can place relevant documents or crawled URLs under each folder for better management.
• Upload Documents: Add PDF, Word, or PowerPoint files. Once uploaded, the file is parsed to extract text-based content.
• Import Website: Provide a URL to crawl. Artiko.ai will automatically fetch and parse the text content from the website.
• Learned Files: After uploading a file or crawling a website, your documents are “learned” by Artiko.ai and can be referenced by your AI assistants.

1. Accessing the Knowledge Base

1. Open Artiko.ai Dashboard: Log in to your Artiko.ai account and navigate to the Knowledge Base section.
2. Create or Select a Folder: On the left sidebar, you can create a new folder or select an existing one. This folder will contain the documents or crawled data you want to add.

2. Uploading Documents

Under the “Create Documents” area, you’ll find an Upload button and a drag-and-drop zone. Supported file types include PDF, Word, and PowerPoint.

1. Select or Drag Your File
   Click on Upload or drag and drop your file into the dashed region. Accepted file types: PDF, DOCX, PPTX.
2. File Parsing
   Once the file is uploaded, Artiko.ai processes and “learns” the content. A progress bar will appear, showing the status. When completed, you’ll see a confirmation, such as File learned!.
3. Manage Uploaded Files
   A list of documents appears below the upload area. You can search for a specific file by name, delete, or rename the file as needed.

3. Importing a Website (Webpage Crawl) BETA

If you prefer to ingest content from a web source, you can use the “Import Website” option to crawl a URL and parse its text content.

1. Click “Import Website”
   Next to the Upload button, select Import Website.
2. Enter the Website URL
   In the input box labeled “Website URL,” type or paste the link you’d like to crawl. For example: https://example.com.
3. Crawl Website
   Click Crawl Website. Artiko.ai retrieves the text content from the page(s) and indexes it. Depending on the complexity and number of pages, this process can take some time.
4. Website Import Completed
   Once crawling is finished, you’ll see the content become available in your folder. It’s now “learned” and part of your knowledge base.

Tip: If a website contains multiple subpages,the system should miss some webpage depending on your link structure and your firewall. Please review accurately the pages imported after the process completes. Eventually you can use a tool like WebPage to Pdf to manually convert missed webpages to Pdf and the upload them manually.

4. Searching and Organizing Your Content

1. Use the Sidebar
   Folders appear in the left sidebar. Click on any folder to see its contents. You can create new folders to segment data by project, department, or topic.
2. Search Files
   A search bar below the upload area lets you quickly locate files by name.
3. Renaming or Deleting
   Each document or crawled data entry typically has options to rename or delete. Use these to keep your knowledge base tidy.

5. Ensuring Your Assistant Uses the Data

Once documents or web content are added, your AI assistants will reference them when generating answers— provided you have:

• Linked Your Assistant to the Knowledge Base: Check your assistant’s settings (User Guide Reference) to ensure it’s allowed to use the knowledge base data assigning the selected folder he has to learn.
• Refined Prompting: If you need more specific queries, adjust your assistant’s prompt or instructions to prioritize knowledge base information.

This step helps ensure the uploaded or crawled data is effectively leveraged in user interactions.

6. Maintenance and Updates

• Delete Outdated Files: Remove or update documents that are no longer valid to prevent your assistant from referencing outdated information.
• Re-crawl Websites: If the website content changes frequently, consider re-crawling the site periodically to keep your AI up-to-date.
• Upload New Versions: If you have revised documents, upload the latest copies so your assistant always references the most current data.

7. Troubleshooting and Support

• Upload Failures: Make sure your file is in an accepted format (PDF, DOCX, PPTX) and does not exceed size limits.
• Crawl Errors: Verify the URL is accessible (no firewalls or password requirements). If the site uses JavaScript-heavy content, partial crawling may occur.
• Indexing Issues: If the file or webpage says “File learned!” but the assistant isn’t using the data, check your knowledge base settings and confirm your assistant has permission to access it.

For additional help, consult the Artiko.ai Support Team or refer to the Artiko.ai Knowledge Base Documentation online for more detailed troubleshooting steps.

Conclusion

The Knowledge Base is the heart of your AI assistant’s domain-specific knowledge. By uploading your documents and crawling relevant web pages, you can keep your AI well-informed, ensuring faster and more accurate responses to user queries. Keep your content organized, updated, and relevant, and your AI assistant will do the rest!