Fields

Overview

Fields are reusable data definitions in the Knowledge Graph. They let you standardize how ContractRabbit names, extracts, maps, computes, and displays contract data across query, reports, dashboards, and Agent workflows.

Field types

Fields can be created from several sources:

• Built-in fields from document metadata and system extraction
• Mapped fields that normalize one or more existing attribute paths into a reusable concept
• LLM fields that extract a value with a custom prompt
• Computed fields that derive a value from other fields

Use mapped fields when the data already exists but appears under multiple paths. Use LLM fields when you need a new extraction. Use computed fields when the answer can be derived from existing values.

Scope

Fields can apply to:

• Document scope for values such as counterparty, effective date, expiration date, governing law, or contract value
• Clause scope for values tied to a provision, such as liability cap amount, non-solicit duration, forum, renewal notice period, or indemnity qualifier
• Both when a field can be useful at either level

Clause-level fields keep citations back to the source provision so reports and Agent answers remain traceable.

Output types

Fields can produce text, numbers, money, dates, percentages, yes/no values, tags, or bulleted lists. Choose the narrowest output type that fits the data so reports can sort, filter, and aggregate it correctly.

Using fields

Use fields to:

• Filter Knowledge Graph queries
• Add columns to reports
• Reuse extraction logic across matters and document types
• Build dashboard widgets from stable definitions
• Give the Agent consistent names for recurring data points

Changes to field definitions can make existing values stale. Rerun or materialize the field to refresh values against the latest definition.