# References

# FAQs & Troubleshooting

1. I have added files to the EKG but cannot see it in the Agent building journey?
In order for an Enterprise Knowledge Garden (EKG) to be available and visible within the Agent building journey, it must be explicitly published.
After uploading files and completing your configurations (chunking, embedding, metadata, etc.), you must click the “Publish” button in the EKG interface. Only published EKGs are accessible to agents such as RAG or ReAct for knowledge grounding.

2. Some files in my EKG are still in processing (chunking/embedding), while others are ready for retrieval. What happens if I publish the EKG now?
If you publish the EKG while some files are still being processed:

• Only the files that are ready will be immediately available to RAG agents.
• The remaining files will continue processing in the background.
• Once these files are marked as "Ready for retrieval", they will be automatically included in the EKG used by agents, no need to republish.

3. What happens if I change the chunking strategy or embedding configuration after files are marked as "Ready for retrieval"?
If you modify the chunking strategy or embedding configuration after files have already been processed:

• The system will automatically delete the previously generated chunks and embeddings.
• The affected files will be reprocessed using the updated configuration.
• To ensure the updated content is available to RAG agents, you will need to republish the EKG if it was already published earlier.

4. Can I update a Knowledge Garden after publishing?
Yes. You can import new files, reprocess content or reconfigure settings. However, remember to republish after changes for them to take effect.

5. Can I remove a file or chunk from a published EKG?
Yes. Files or individual chunks can be deleted or excluded. Remember to republish the EKG after changes.

6. If I change the visibility of an EKG from Public to Private, will the RAG agent using it continue to work?
Yes, the RAG agent will continue to function even if the EKG's visibility is changed from Public to Private. Changing the visibility does not retroactively remove access for agents that have already been configured to use the EKG. To revoke access: If you want to prevent a RAG agent from using a specific EKG, you must first remove the EKG from the agent configuration. Only then should you change the EKG visibility to Private to fully restrict future access.

7. My file is not uploading. What could be wrong?
Check the following:

• File format is supported : PDF, DOC, DOCX, PNG, JPEG, TIFF, JPG.
• File size is under 20MB and Image size is under 5MB
• File is not encrypted, zipped, or password protected

8. I uploaded a 20 MB file, but it’s not uploading. What could be the issue?
While your file may appear to be 20 MB on your local system, its actual size might be slightly more, for example, 20.1 MB which exceeds the upload limit.
Operating systems often round file sizes for display, so what shows as 20 MB might in reality exceed the system’s maximum allowed limit of 20.0 MB.

9. Chunks are not showing after upload. Why?
Ensure:

• Chunking and embedding configuration was properly set
• File status is “Ready for retrieval"
• Refresh the Chunk Editor tab

10. Retrieval results are not relevant. How can I improve accuracy
Try the following:

• Add or refine metadata/tags
• Use a more specific chunking strategy
• Enable the Reranker model
• Increase Top K for broader chunk selection

11. Why is my EKG not appearing in the Agent configuration?
An EKG may not be appearing in your Agent configuration if it has not been published and made accessible within your workspace.To resolve this issue, you should take the following steps:

• Verify the status of your EKG: Check to see if the EKG is currently in a "draft" or "unpublished" state.
• Publish the EKG: If it is not already published, follow the necessary steps within your specific platform to publish it. This action will make it officially available within your workspace.
• Confirm accessibility: After publishing, ensure that the EKG is accessible from the agent builder. This may involve checking permissions or visibility settings within your workspace.

12. Why is my RAG agent still missing chunks after I upload files?
The RAG agent is missing chunks because not all of your uploaded files have been fully processed yet.
For a RAG agent to access the information within your files, they must first be processed and indexed for retrieval. Only files that have completed this process and are marked as "Ready" will be available to the agent.
Any files still undergoing this procedure will be added to the agent's knowledge garden automatically once their processing is complete.

13. How can I query relational databases using natural language (NL2SQL)?
The platform enables Natural Language to SQL (NL2SQL) functionality by connecting agents to relational databases via specialized connectors/tools. This process is orchestrated by the agent:

• The agent receives the natural language query from the user (e.g., a non-technical subject matter expert)
• It automatically translates the natural language into a valid SQL query from the system instructions (or a similar database query)
• It executes the query against the database
• It returns the final answer back to the user in natural language

14. What tools are available to refine and optimize RAG (Retrieval-Augmented Generation) performance?
The platform offers multiple avenues for RAG refinement:
EKG Features: Create an effective knowledge garden using appropriate metadata, utilize different chunking methodologies, and select specialized embedding models. You can also test retrieval quality before agents use these knowledge.
Agentic RAG (EDE) Features: Advanced controls include self-query and multi-query augmentation, setting specific top N and top K values for retrieval, and selecting reranker models to further improve the final output.

15. Where can I find an inventory of all active Enterprise Knowledge Gardens (EKGs)?
The Enterprise Knowledge Garden module provides a transparent inventory of all the created Knowledge. This allows users to view and manage the data sources being utilized by agents, which is crucial for tracking data lineage and ensuring content governance.

16. How does the platform support faceted search capabilities?
Faceted search is enabled through metadata-driven advanced RAG. The EKG enriches all ingested data with structured metadata and categories, allowing users or agents to filter and refine search results dynamically. By utilizing agentic RAG and self-query techniques, searches can apply these facets to deliver highly precise, context-aware responses, moving beyond simple keyword matching.

17. Can the platform search across multiple, diverse data sources simultaneously?
Yes, Federated Search allows the platform to search across diverse data sources simultaneously without centralizing the data. This capability works with multiple vector databases, as well as structured data sources like databases and enterprise applications, all within a single multi-agent use-case, enabling a unified search experience.

18. Can the platform support metadata-driven search capabilities across data sources?
Yes, the platform facilitates metadata-driven and faceted search across diverse data sources through an advanced RAG. The EKG enriches all ingested data with structured metadata and categories, allowing users or agents to filter and refine search results dynamically. By utilizing agentic RAG and self-query techniques, searches can apply these facets to deliver highly precise, context-aware responses, moving beyond simple keyword matching.

19. Can the platform search across multiple, diverse data sources simultaneously?
Yes, the platform supports federated search that allows searching across diverse data sources simultaneously without centralizing the data. This capability works with multiple vector databases, as well as structured data sources like databases and enterprise applications, all within a single multi-agent use-case, enabling a unified search experience.

# Limitations

The following points detail the current limitations associated with the Enterprise Knowledge Garden (EKG) and its capabilities

# File Upload

• File Size: Individual files must be under 20MB. Images must be under 5MB
• File Security: Files that are encrypted, zipped, or password-protected are not supported
• File Format: Supported file formats are PDF, PNG, JPEG, TIFF, DOCX
• PDF and TIFF Files: These files have a memory limit of 500 MB and a page limit of 3,000 pages
• PDF Dimensions: The maximum height and width is 40 inches and 2880 points

# Language

• Supported Language: The EKG currently supports files only in English

# Data Ingestion

• External Sources: You cannot get data directly from external sources like Amazon S3 through the EKG interface. To ingest data from these sources, use the Data Library or build a flow in the Flow Designer
• Connectors: Some connectors are designated as read-only, which means they are designed exclusively to retrieve or "read" data from a source. They do not have the functionality to modify, create, or update records in the original system

For example, the current ServiceNow and IMAP connectors only support fetching information; you cannot use them to create new records or update existing ones directly from the platform.

• HTML: Web crawlers are currently the only way to ingest HTML files

# Data Retrieval

• Chunking: For large files, chunking may result in context loss while being processed by LLM model which can reduce overall retrieval accuracy and performance of the agent. It is recommended to add Metadata to all chunks to improve overall performance
• Metadata: There is no automatic way to add metadata to chunks within the EKG. You must do so manually or by constructing a custom process using a flow designer
• Top-K Retrieval: The agent is limited to retrieving a maximum of 100 chunks per query, which may cause it to miss relevant information in large datasets
• Top-N Re-ranking: The re-ranking process is limited to refining only the top 50 of those retrieved chunks, potentially excluding other valuable context from the final response
• Indexing Delay: After embedding is completed, RAG retrieval may not immediately return results. We recommend waiting a few minutes for the data to be fully indexed in the database, after which it will be available for retrieval