Findsight

Findsight's advanced search filters offer powerful tools for content discovery and research. These features build upon traditional search methods while introducing AI-powered capabilities that enhance precision and flexibility. The mention filter provides basic search functionality, while more specialized tools like references, state, and answer filters enable sophisticated content exploration. Together, these features represent a sophisticated approach to information retrieval, combining familiar search principles with innovative AI-powered enhancements.

Basic and AI-Powered Filters

Mention Filter

The mention filter operates similarly to traditional search engines, identifying sources that contain specific literal text. While functional like Google, it may lack the precision required for highly specialized searches. Users begin by entering their query into the standard search input field. The platform then presents results based on the exact phrasing provided, making it particularly useful for locating direct quotes or specific statements. Users may enhance their query by incorporating prefixes and alternatives using OR logic, such as "dus_" OR "wash_" to find both "dusting" and "washrooms."

References Filter

For more structured searches, Findsight AI's references filter employs a more refined approach. Users initiate their search by typing a keyword, followed by the "+" operator (e.g., +running). The system then presents a range of suggested entities, with options marked "W" linking directly to Wikipedia entries. These suggestions represent named entities across various categories. To refine their search further, users can add multiple entities, specifying whether a source must contain one or all selected terms using a dropdown menu. The filtering process continues by selecting the desired conditions and submitting the query.

State Filter

The state filter represents an AI-powered enhancement, combining multiple search functionality features into a single tool. This versatile filter supports several operators, including +answer for querying specific information, +discuss for exploring topics found within source materials, +mention for searching directly for terms, +reference for entity-based queries, and +state for creating custom claims. When initiating a state filter search, users can toggle between various display options found in the search bar's right-hand menu, allowing them to adjust the format of results to their preference.

The Mention Filter

The mention filter operates in a manner similar to traditional search engines, with users entering their query into the standard search input field. The platform then returns results based on the exact phrasing provided, making it particularly useful for locating direct quotes or specific statements. Users may enhance their query by incorporating prefixes and alternatives using OR logic, such as "dus_" OR "wash_" to find both "dusting" and "washrooms."

One limitation of the mention filter is its lack of precision compared to more specialized search tools. For example, when searching for a specific term, users may need to employ creative use of prefixes and alternatives to refine their results effectively. The filter's basic functionality is designed to mirror that of Google, making it familiar for users accustomed to traditional web search methods. However, its effectiveness can be improved by combining it with other filtering options, such as the references or discuss filters, which offer more targeted approaches to content discovery.

The Reference Filter

The references filter in Findsight AI allows users to search for content based on named entities, with results categorized by entity type. When initiating a search, users type a keyword, followed by the "+" operator. The system then presents a range of suggested entities, with options marked "W" linking directly to Wikipedia entries. These suggestions represent named entities across various categories, enabling users to search for specific organizations, people, or concepts.

To refine their search further, users can add multiple entities, specifying whether a source must contain one or all selected terms using a dropdown menu. For example, a user searching for information on climate change might enter "+climate change +Wikipedia" to find relevant Wikipedia entries, or "+climate +change +Wikipedia" to include only sources containing all three terms. This flexible approach allows users to build complex search queries while maintaining clear categorization of their results by entity type.

The references filter also incorporates the use of asterisk prefixes for creative querying. By adding an asterisk before a term, users can search for variations or misspellings. For instance, entering "dus*" in the search field would return results for "dusting" and "washrooms," demonstrating the filter's flexibility in handling alternative phrasing. This feature, combined with the ability to specify exact term matches, makes the references filter versatile for both casual exploration and detailed research.

The Discuss Filter

The discuss filter operates similarly to the references filter but focuses on specific topics or tags found within source materials. When activated, this filter generates an expandable list of related discussions beneath each search result, providing users with additional context and related claims.

Like its counterparts, the discuss filter employs flexible querying through prefix and OR operators. Users can initiate a search by typing a topic term and appending it with the appropriate operator, such as "+climate change" or "+Wikipedia." The system then generates a list of suggested entities, with options marked "W" leading to Wikipedia entries and others categorized as named entities.

Users can broaden their search criteria by adding multiple topics, specifying whether a source must contain one or all selected terms using a dropdown menu. For example, a search for "+climate change +extreme weather +Wikipedia" would retrieve sources containing all three specified terms, with results categorized by entity type.

The filter's functionality is identical to that of the mention and reference filters in terms of prefix usage. By adding an asterisk before a term (e.g., "dus*"), users can search for variations or misspellings, demonstrating the filter's flexibility in handling alternative phrasing. This feature, combined with the ability to specify exact term matches, enables users to conduct comprehensive topic-based research while maintaining clear categorization of their results.

AI-Powered State and Answer Filters

The AI-powered state and answer filters represent an advancement in content discovery, allowing users to create custom claims or formulate questions and receive tailored results supported by relevant data.

State Filter

The state filter operates similarly to the platform's Random button and explore arrow feature, enabling users to enter custom claims and discover related claims. This functionality proves particularly useful for citation purposes, providing users with relevant content while facilitating deeper exploration of their chosen topic. Users can construct their search queries using the operators +answer, +discuss, +mention, +reference, or +state, offering flexibility in how they approach their research.

Answer Filter

For users seeking specific information, the answer filter offers targeted results by addressing questions entered by the user. This feature provides context and supporting data to help answer the posed question, making it a valuable tool for research and learning. The filter employs the same flexible querying methods as its counterparts, allowing users to refine their search through prefix and OR operators, such as "dus_" OR "wash_."

Both filters display their results in a manner consistent with the platform's other filtering options, categorizing entities by type when appropriate. For example, searches conducted through the references filter return results marked with "W" for Wikipedia entries or categorized as named entities, while the answer filter presents responses that address the user's specific question. The platform maintains a user-friendly workflow throughout the filtering process, allowing for easy refinement of search criteria through the use of dropdown menus and clear display options in the search bar's right-hand menu.