CassidyAI offers a powerful search function that prioritizes semantic understanding and relevance over simple keyword matching. This function is designed to provide users with high-quality, meaningful results that best match the context and intent of their queries. In this article, we will cover how Cassidy's search function works, its ideal use cases, and provide a simple example to help you understand its capabilities.

CassidyAI employs the Retrieval-Augmented Generation (RAG) system to analyze the context and intent behind user queries. This means that Cassidy prioritizes relevance and context over simple keyword matches. As a result, Cassidy may return a smaller subset of documents that discuss the searched topic in a more meaningful and significant manner.

While this approach ensures high-quality results, it may not always align with user expectations for comprehensive document retrieval based on specific keywords. To address this, users are encouraged to refine their queries with more context or utilize specific features like the "#" tag to better align with their search needs.

CassidyAI's search function is perfect for:

Let's say you need to find information about renewal provisions across your company's documents. Instead of searching for the keyword "renewal," you can frame your query around the context you seek, such as:

By using more descriptive language and focusing on context, CassidyAI will return a smaller, more relevant subset of documents, allowing you to find the information you need with ease.

In conclusion, CassidyAI's search function offers users an effective and efficient way to navigate through large volumes of data. By understanding the importance of semantic search and adapting your queries accordingly, you can unlock the full potential of CassidyAI's capabilities.

While CassidyAI's search function offers numerous benefits, there are certain limitations when it comes to being exhaustive in its search results and handling queries based on numeric data. In this section, we will discuss these limitations and suggest alternative approaches for such queries.

CassidyAI's focus on semantic understanding and context may hinder its ability to provide exhaustive search results based on specific keywords. Since it concentrates on the relevance and meaningful context, CassidyAI might not return every instance of a specific word or phrase present in the documents.

For example, when searching for "contract renewals," CassidyAI might not return every document that contains the term "renewal" if it does not deem the context to be significant or relevant to the query.

Alternative Approach:

To obtain a more comprehensive list of documents containing specific terms, consider using other search tools that specialize in keyword matching or exhaustive document retrieval. These tools will enable you to find every document containing specific terms, although the results may lack the contextual relevance provided by CassidyAI.

CassidyAI might struggle with queries that focus on numeric data or require a precise time-based search. For example, if you asked, "Give me all of the contract renewals we've done since 2016," CassidyAI might not provide satisfactory results as it prioritizes semantic context over numeric information.

Alternative Approach:

For queries involving specific numbers or timeframes, consider using software or tools dedicated to time-based filtering and sorting. These tools can quickly and efficiently filter documents or data based on the specific numeric criteria, allowing you to find the desired information. You can then utilize CassidyAI for more context-based analysis and understanding of the retrieved documents.

In summary, while CassidyAI is an invaluable tool for retrieving contextually relevant and meaningful information, it has limitations when it comes to exhaustive search results and handling numeric data. By acknowledging these limitations and using alternative approaches for such queries, you can ensure a well-rounded and efficient search experience.