Knowledge Base

The Knowledge Base serves as the central intelligence hub for your AI ecosystem. It is the foundational layer where raw organizational data is transformed into structured, accessible information that powers your AI Agents. By bridging the gap between static documents and conversational interfaces, the Knowledge Base enables Retrieval-Augmented Generation (RAG) ensuring that every user interaction is grounded in your specific data.

Key Capabilities

• RAG-Driven Intelligence. Leverage Retrieval-Augmented Generation to provide AI Agents with the exact context needed from your private documents to answer complex queries.
• Data Integration. Seamlessly ingest information from diverse sources, including document uploads, web crawling, and third-party integrations.
• Contextual Retrieval. Uses advanced indexing to ensure AI Agents retrieve the most relevant "knowledge chunks" during active conversations.
• Knowledge Maintenance. Manage the information lifecycle through automated sync schedules, manual overrides, and version control.

In this guide, you’ll learn how to access the Knowledge Base, how to add and manage data sources, and how to effectively configure the Knowledge Base to improve your AI Agents ability to handle user queries.

Accessing the Knowledge Base

To access the Knowledge Base, select the desired AI Agent and solution and from the NLU menu, click Knowledge Base.

When you access the Knowledge Base for the first time, the page is empty. To ingest your data, add as many data sources as you want, extract the data and train the KB.

Adding data sources

Druid allows you to aggregate knowledge from multiple enterprise silos:

• Unstructured Files: PDF, DOCX, and TXT files (Ideal for manuals and policies).
• Structured Data: Excel and CSV files (Best for tabular or FAQ-style data).
• Web Crawling: Direct URL ingestion for public-facing help centers or blogs.
• Connectors: Native integration with SharePoint, OneDrive, and Google Drive.

Use the Knowledge Base

By default, if the NLP model cannot match the user's input to any existing flow, it will trigger the Intent not found flow (if configured on your AI Agent).

To enhance your AI Agent ability to handle such cases, Druid provides a variety of Knowledge Base solution templates, including solutions powered by generative AI. These templates help deliver more complete and natural answers when flow-based logic isn’t enough. For more details, explore the available KB solutions in the Solution Library.

To set up the Knowledge Base behavior, follow these steps:

1. Import your preferred KB solution from the Solution Library and configure it.
2. Navigate to Knowledge Base Settings.



3. Tap on Use Knowledge Base to provide your AI Agent with the capability to search through the Knowledge Base when the user input does not match with any of the existing flows.
4. Select the Knowledge Base response flow (used when an answer is found).
5. If you want to rephrase / improve user intent by using an external service (e.g., GPT) to a user intent that is optimal for your AI Agent model, select the Intent rephrase flow specifically designed for rephrasing user intent.

After the user sends the intent, the AI Agent first executes the "Intent rephrase flow" that rephrases the utterance and then it uses the result (the rephrased intent stored in [[Intent]].Text) to predict in the model.

IMPORTANT! The "Intent rephrase flow" is executed only when the user input is sent while the conversation is in idle mode. Once in a flow, the flow no longer executes (the user intent is not altered).
Info: This is particularly useful for Knowledge Base with ChatPT.
6. Configure any additional general or advanced KB settings as needed.

This setup helps your AI Agent handle a wider range of user queries while maintaining a high-quality conversational experience.

Knowledge Base General Settings

To access the Knowledge Base settings, in the Knowledge base page, click the Knowledge Base Settings tab.

The General Settings appear by default.

The table below provides the description of the Knowledge Base general settings.

Setting	Description.
Embeddings Provider	There are three providers available: DRUID (the default provider). DRUID uses a non-generative Large Language Model (LLM) for embeddings. If you select DRUID as embeddings provider, you can choose the Embeddings Model (either English or Multilingual). Open AI. To use OpenAI’s text embeddings, select OpenAI as embeddings provider and enter the password of your OpenAI account. Microsoft Azure. To use the Azure OpenAI embeddings, select Microsoft Azure as embeddings provider and enter the Azure OpenAI embeddings endpoint (https://{your-resource-name}.openai.azure.com/openai/deployments/{deployment-id}/embeddings?api-version={api-version}) and the password of your Azure OpenAI account.
Embeddings Model	An embeddings model is a machine learning model that transforms data, such as text or images, into a vector of numbers (an embedding). This vector representation captures the semantic meaning or relationships within the data, allowing for more efficient comparisons, searches, and analysis. The following embedding models are available in Druid: English: Creates embeddings optimized for English text, capturing semantic meaning in English language inputs. English – score correction: Generates English embeddings with additional score correction, improving the accuracy and relevance of similarity scores. HigherEducation.v1 (English): This specialized model is trained on a rich dataset of higher education content in English only, including university-specific information, leading to improved accuracy when dealing with queries related to this domain. MultiAspect: An enhanced version of our existing Multilingual model, MultiAspect offers improved performance across multiple languages and incorporates score correction for more accurate results. Multilingual: Produces embeddings for multiple languages, enabling cross-lingual semantic understanding. Multilingual – score correction: Offers multilingual embeddings with score correction, enhancing the precision of similarity scores across languages. Paraphrase: Specializes in identifying semantically similar phrases, even if they differ in wording, useful for paraphrase detection tasks. Info: The Paraphrase embeddings model processes up to 125 tokens per paragraph and is ideal for short sentences, while other models support up to 512 tokens per paragraph.
Set results threshold	The Results Threshold settings determine how matching utterances are evaluated against the Knowledge Base (KB). These settings vary depending on whether the AI Agent is new or existing. For new AI Agents, Use AI Agent NLU Thresholds is enabled by default. The KB uses the NLU thresholds configured on the AI Agent (NLU menu > Configurations > Intents tab > Thresholds and Parameters section). For existing AI Agents, the behavior varies based on the NLU thresholds: Both 'Min Match Score' and 'Target Match Score' are null: 'Use AI Agent NLU Thresholds' is enabled, and the slider shows the NLU thresholds. The slider is disabled. Either 'Min Match Score' or 'Target Match Score' has a value: 'Use AI AgentNLU Thresholds' is disabled. The slider becomes editable, letting you adjust the threshold for the parameter with a value. The empty parameter uses the NLU thresholds set on the AI Agent. Both 'Min Match Score' and 'Target Match Score' have values: 'Use AI Agent NLU Thresholds' is disabled, and you can set values for both parameters using the slider. To control how the KB evaluates and matches user input, disable Use AI Agent NLU Thresholds and adjust the 'Min Match Score' and 'Target Match Score' using the slider, ensuring that it aligns with your desired performance thresholds. NOTE: If you enable Use AI Agent NLU Thresholds, the threshold values set on the slider will be lost.
Search balance	By default, the search within the knowledge base is performed using a mix of the following two algorithms: the keyword (Text) search algorithm and the semantic (Vector) search algorithm. Additionally, you can use the reranker to perform further analysis and enhance the result quality. If you don't use the reranker, the recommend value for the search balance is Vector 80% , Text 20% , meaning that the search will use 80% the semantics search algorithm (the search uses the semantic search algorithm that returns more accurate results) and 20% the text search algorithm (the search uses the keyword /text algorithm, which might return a lot of noise). Move the slider to set the search balance based on your needs. Info: The values you set for the Search balance slider and the Score Calculator strategy in Advanced Settings are synchronized. Any changes made to one will be reflected in the other.
Search inside answers	Tap on if you want the user says to be matched against both the question and answer pairs available in structured data sources. If this option is off, the user says is matched only against the questions available in the structured data sources.
Use Knowledge Base	Tap on to provide your AI Agent with the capability to search through the Knowledge Base when the user input does not match with any of the existing flows. From the Knowledge Base response flow drop-down, select Knowledge-Base-response-flow.
Intent rephrase flow	If you want to rephrase / improve user intent by using an external service (e.g., GPT) to a user intent that is optimal for your AI Agent model, select the flow you specifically designed for rephrasing user intent. After the user sends the intent, the AI Agent first executes the "Intent rephrase flow" that rephrases the utterance and then the AI Agent uses the result (the rephrased intent stored in [[Intent]].Text) to predict in the model. IMPORTANT! The "Intent rephrase flow" is executed only when the user input is sent while the conversation is in idle mode. Once in a flow, the flow no longer executes (the user intent is not altered). Info: This is particularly useful for Knowledge Base with LLM.

Save the settings.

Training or retraining the Knowledge Base

When managing your Knowledge Base, you can choose between Train and Retrain actions, depending on your needs:

• Train - This option trains only the data sources that require training (e.g., newly added or updated sources that haven’t been trained yet). It's the default option when you're adding new content or making updates that trigger a training need.
• Retrain - Use this option to retrain all already-trained data sources and train once any that require training. This is especially useful after making changes to the Knowledge Base’s general settings (e.g., Embeddings model, Embeddings provider, or Search inside answers).

At the data source level, you can retrain an individual source after modifying its Trainable Elements, even if its training status doesn’t change. You also have the option to Retrain all data sources directly from a data source.

Testing the Knowledge Base performance

Testing the performance of your Knowledge Base is important because it ensures that the Knowledge Base is delivering accurate and relevant responses, helping to identify and address any issues to improve overall performance.

To test the performance of your knowledge base, on the Knowledge Base page, enter a question in the User Says area and select the language. The model will search across all the data sources in the Knowledge Base and list the articles with a matching score higher than 0.5, along with the data source where each article was found. If you have changed the threshold ([[Intent]].KBQnAItems[0].Score) in the solution configuration, only the articles meeting that threshold will be listed.

If you selected a score calculator strategy from the KB settings, for each result, you will see the total matching score and the weights of the algorithms used. You can view the graphical representation of these weights by clicking the Info icon.

Search within the KB

The Knowledge Base (KB) search feature helps you quickly find information across your entire KB or within specific data sources using exact match searches. Search results include all matching data source elements (nodes and leafs) and articles containing your specified keywords.

When you search at the KB level, up to 30 matching records (if available) are displayed under each corresponding data source name. When you search within a specific data source, up to 30 results are shown if available.

Advanced Search Options

Available in Druid version 9.9 and higher, advanced search options let you choose which article elements to search within:

• URL – Search only within article URLs.
• Title & Content – Search within article titles and paragraph text.
• All – Search across all article elements (default).

This targeted search capability helps you locate information more efficiently based on where it appears in your articles.

Filtering Search Results

You can refine your Knowledge Base search results using filters for:

• Data source type
• Document type
• Exclude options – Exclude specific data sources or elements from your results.

These filters help you perform more precise and efficient searches, giving you greater control over what is displayed.