Skip to main content

Configuring Questions

The questions you define determine what Review Matrix extracts from your documents. Judicio provides three modes for configuring questions, each suited to different workflows. This guide explains how each mode works, when to use it, and how to get the best results.

Question modes

Smart mode

Smart mode uses AI to analyze a sample of your selected documents and automatically generate a set of relevant questions. This is the fastest way to set up an extraction when you are exploring a new document set or are not sure which questions to ask.

How it works:

  1. Select Smart from the question mode selector.
  2. Judicio reads a sample of your documents and identifies the document type (e.g., lease agreement, share purchase agreement, loan agreement).
  3. Based on the document type, Judicio generates a tailored set of questions covering the most commonly extracted data points.
  4. The generated questions appear in an editable list.

When to use Smart mode:

  • You are reviewing an unfamiliar document type for the first time.
  • You want a quick overview of what is in a document set before defining targeted questions.
  • You are short on time and need a reasonable starting set without manual configuration.

Refining Smart-generated questions:

  • Edit any question by selecting it and modifying the text.
  • Delete questions that are not relevant by selecting the remove icon next to the question.
  • Add additional questions manually using the Add Question button at the bottom of the list.
  • Change the question type using the dropdown next to each question (see Question types below).
Get better Smart results

The more documents you select before entering Smart mode, the better Judicio can identify the document type and generate relevant questions. If your selection contains mixed document types, Smart mode bases its questions on the most common type in the set.

Question types

Every question has a type that tells Judicio what format the answer should take. Choosing the right type improves extraction accuracy and makes results easier to analyze.

TypeDescriptionExample questionExample answer
TextFree-form text answer. Use for open-ended questions.What are the key indemnification provisions?"The Seller shall indemnify the Buyer against all losses arising from breach of warranty, subject to a cap of..."
Yes / NoBinary answer. Use for questions that have a definitive yes or no answer.Does the agreement contain a change-of-control provision?Yes
DateExtracts a specific date.What is the lease commencement date?1 January 2025
NumberExtracts a numeric value.What is the annual base rent?250,000
ListExtracts multiple items as a structured list.What are the tenant's permitted uses of the premises?1. General office use 2. Client meetings 3. Storage of business records
Clause ExtractExtracts the exact text of a clause verbatim.Extract the termination for convenience clause."Either party may terminate this Agreement for convenience by providing not less than ninety (90) days' prior written notice..."
Choosing the right question type

When in doubt, use Text -- it is the most flexible type and works well for most questions. Use specific types like Date, Number, or Yes / No when you need structured data that you plan to sort, filter, or analyze in a spreadsheet after export.

Reordering and organizing questions

  • Drag and drop questions to reorder them. The order determines the column order in Table view and the sequence in exported files.
  • Group related questions together (e.g., all financial terms, then all governance terms) to make the results matrix easier to read.
  • There is no hard limit on the number of questions, but more questions increase credit cost and extraction time.

Previewing before extraction

Before running the extraction, the summary screen shows:

  • The total number of documents selected.
  • The complete list of questions with their types.
  • The estimated credit cost.

Review this summary carefully. It is more efficient to refine your questions now than to re-run the extraction later.

Next steps