Query

Overview

Query is the search workspace inside the Knowledge Graph. Use it when you need to find documents, clauses, or extracted values by combining natural-language text with structured filters.

Query is broader than a document list search: it can search across document metadata, clause text, taxonomy classes, attribute values, parties, matters, document types, and custom fields.

What you can query

Common query patterns include:

• Documents with a specific party, counterparty, matter, or document type
• Clauses that match a taxonomy class, such as indemnification or non-solicit
• Attributes such as governing law, forum, term, notice period, cap amount, dates, and locations
• Custom fields created in the Knowledge Graph
• Text phrases or semantic concepts across the corpus

Use document scope when you want one row per document. Use clause scope when you want the matching provisions themselves.

Filters

Filters narrow the result set before you build reports or inspect source documents:

• Document type limits results to a contract category
• Matter limits results to a deal, project, or workspace grouping
• Party and counterparty filters by extracted entities
• Clause class filters by the taxonomy path
• Field filters compare built-in, mapped, LLM, or computed field values
• Confidence limits results to higher-confidence extracted values

From query to report

After a query returns the right population, turn it into a report by choosing columns and row scope in Reports. The saved report keeps the query definition, so it can be rerun as documents and field values change.

Agent integration

The Agent can run similar corpus queries from natural language. Use the Knowledge Graph directly when you want to inspect or save the exact query; use the Agent when you want it to infer the query, explain assumptions, or generate a report from a prompt.