Uploading Documents to Data Lake

The Documents for Copilot page is designed to facilitate uploading documents to a data lake and interacting with these documents using the Copilot chat feature. The primary objective of this page is to allow users to manage documents that are to be indexed and made available for conversational queries through Copilot.

Copilot allows context-aware assistance during analysis. It uses uploaded documents, like technical manuals and maintenance guidelines, to generate relevant insights and recommendations. This enhances the efficiency of any process within IFS Cloud by automating data retrieval, making analysis more focused and accurate.

Once documents are uploaded to the data lake, Copilot can reference these sources to suggest solutions and strategies to any queries. This minimizes the need for manual input and guarantees that recommendations are based on authoritative data. Additionally, documents uploaded to the data lake can be removed if necessary, providing flexibility in document management.

 Limitation

• There is an upload limit of 200 documents.
•  Currently, no specific access controls are applied to the documents uploaded to the data lake. This means that once a document is uploaded, it becomes accessible to anyone using the Copilot chat feature.
•  Only PDFs are supported for upload to ensure compatibility with the data lake indexing and Copilot's interaction capabilities.
• This page is intended exclusively for uploading documents for Copilot interaction. It is not supported for document content searching through Document Management.

Note: The documents uploaded are accessible to all the users in the organization who have access to Copilot. Therefore, users should avoid uploading any sensitive or restricted content.

Tips and Tricks for Optimizing Document Interaction

To ensure successful interaction and indexing of your documents in Copilot, follow these recommendations:

• Copilot cannot extract information from images within PDFs. Ensure key data is in text form.
• If your document includes tables, ensure they are created as text-based tables rather than images. PDFs don’t have an inherent table structure like HTML or Markdown, so tables must be handled carefully. Image-based tables will not be indexed.
• Simple, well-structured documents with clear headings and readable fonts will improve Copilot’s ability to understand and interact with the content.

Pages

Documents for Copilot

Activity Diagrams

Upload Documents to Data Lake