Knowledge Base Tool

Enhance data reasoning by defining business rules that guide AI in interpreting and analyzing queries based on business-specific logic and domain knowledge. You can specify these rules in natural language.

An example business rule:

Consider a business rule for identifying a significant drop in sales.

The definition of “significant” may vary across organizations. Here, the Knowledge Base tool can be equipped with following business rule:

A sales drop is considered significant if it meets any of the following conditions:

1. Sales decline by 10% or more within a week.

2. Sales decline by 20% or more within a month.

By configuring this rule in the Knowledge Base, the system ensures that queries related to sales performance consistently apply the user-defined threshold, leading to more relevant insights.

Similar to this, you can specify many such rules to govern the behavior of data analysis.

Apart from business rules knowledge base can also be leveraged to specify SOPs, best practices etc. To clearly define “when and how” to use the knowledge base during data analysis, leverage Usage Instructions section.

Usage Instructions #

Define when and how this Knowledge Base should be used for data analysis.

For Example: Use business rules when generating financial reports to ensure correct tax calculations.

Once configured, the Knowledge Base will be integrated into your Data Intelligence application to improve query accuracy and contextual reasoning.

There are many ways to configure the Knowledge Base:

• Manual Entry

• File Upload

• Vector DB

Manual Entry #

Enter business rules inline in the provided text box.

File Upload #

Upload a .txt file (maximum 100 KB) containing business rules.

This option allows for bulk rule additions without manual input.

Vector DB #

In case your business rules, SOPs or best practices guides are available in a Vector database. You can point it from here by choosing Vector Database option. With this, during data analysis when needed your business rules will be picked from the specified vector database.

Configure the Vector DB parameters for the Knowledge Base.

Choose Vector Database Type #

Use VectorDB details from Connection step, or choose a Vector Database.

Supported Vector DB for Data Intelligence:

• Pinecone

Provide Vector Database Connection #

Choose an existing connection or create a new one to link with the selected Vector Database. Test the connection to ensure its availability.

• To add a connection for Pinecone Vector DB, please refer to Pinecone Connection documentation.

Vector Database Index #

Select the index where the business rules are stored under the specified namespace and metadata key.

Namespace #

Enter the namespace for the selected index.

Metadata Key #

Specify the metadata key containing business rules that guide query reasoning.

For example: Content

Configure the embedding parameters for the Knowledge Base.

Embeddings Provider for Knowledge Base #

An Embeddings Provider is required to process and search user queries effectively within the Vector DB.

Choose Embeddings Provider #

Use Embeddings Provider details from Connection step, or choose an embeddings provider.

Supported Embeddings Provider & Models for Data Intelligence:

Embeddings Provider’s Connection & Model #

Choose an existing connection or create a new one to link with the selected Embeddings provider. Test the connection to ensure its availability.

• To add a connection for OpenAI Embedding provider, please refer to OpenAI Connection documentation.

• To add a connection for Voyage AI Embedding provider, please refer to Voyage AI Connection documentation.

Dimension for LLM Models #

Specify the embedding dimension that matches the model used for generating vector representations. This ensures accurate retrieval of business rules and metadata from the Vector Database.

For Example: 1024 for voyage-3-large Voyage AI model. Or 1536 for text-embedding-3-small and text-embedding-3-large OpenAI models.