Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.sagalegal.io/llms.txt

Use this file to discover all available pages before exploring further.

Grid review is how you analyze a set of documents against the same questions in one structured pass. Define what you want to know — termination clauses, effective dates, governing law, liability caps — and the AI works through every document, filling in a table where each row is a document and each column is a question. It’s built for the kind of work where you’d normally open twenty contracts side by side and make a spreadsheet by hand. Grid review does that in minutes, with citations linking every answer back to the source passage.

​
Where grid reviews live

Grid reviews can live inside a project or on their own. Inside a project, you’ll find them under the Grid review tab alongside chat. The grid automatically has access to that project’s knowledge base, so you can pull documents directly from there. You can also create standalone grids from the Grid reviews page in the sidebar. Standalone grids have their own document set. If you decide later that a grid belongs in a project, you can move it.

​
Creating a grid

Click New grid review to start from scratch, or open the Grid library to begin from a template. Give the grid a name that identifies the task — “Lease Agreement Review” or “Employment Contract Extraction” works well. You can also tag the grid with one or more practice areas so it shows up under the right filter when you come back to it later. From there, you add questions and documents. Questions become columns; documents become rows. The AI starts generating answers as soon as both are in place.

​
Answer modes

Each question has an answer mode that tells the AI what kind of response you expect. Choosing the right mode keeps results consistent and makes the grid easier to scan.
ModeWhat the AI returnsWhen to use it
ExtractVerbatim text from the documentPulling out specific clauses, definitions, or passages
Short reviewA brief summary or analysisQuick assessments, concise observations
Long reviewA detailed, multi-paragraph analysisIn-depth reviews, multi-part questions
True / FalseTrue or FalseBinary checks — “Does the contract contain a non-compete clause?”
DateA dateEffective dates, deadlines, expiration dates
NumberA numeric valueFinancial figures, quantities, limits
CurrencyA monetary amountContract values, fees, penalties
When the AI can’t find what you’re asking for, it tells you — “Requested date not found” or “Cannot determine true or false” — rather than guessing. You can also click Improve prompt when writing a question to let the AI refine your phrasing for more precise results.

​
Custom instructions

If there’s context the AI should keep in mind across all questions — a jurisdiction focus, a specific legal framework, a formatting preference — you can add it as a custom instruction when creating or editing the grid. Something like “Answer in the context of Norwegian employment law” or “Focus on financial obligations and penalties.” Each question can opt in or out of the grid-level instruction, so you have control over which questions benefit from the extra context and which should stand on their own.

​
Adding documents

You can add documents from several places: any knowledge base available to you (not just the current project’s), a file upload, or integrations like SharePoint and Epona365 if your organization has them configured. Select as many as you need — the AI generates answers for every document against every question. Documents can be added or removed at any time. When you add new documents to an existing grid, the AI fills in the answers automatically. The grid isn’t fixed at creation.