Case Context

Modified on Wed, 11 Feb at 2:44 PM

Need help with anything in this article or have other questions? Contact us at support@noticiasolutions.com

This article explains what Case Context is, when to use it, and how to use it effectively with Noticia Go.

What is Case Context?

Case Context is a concise, shared summary of a single legal matter that lives inside Noticia Go for a given case + organization.
It is designed to help your AI prompts “understand” the matter without re-explaining it every time, and to keep your whole team aligned on:
  • Who is involved (parties, people, counsel)
  • What the case is about (facts, timeline, issues)
  • What counts as relevant or important for review
  • Where privilege is likely to arise
  • Key collection details and additional notes
Case Context is:
  • Per case + organization: each (Case, Org) pair has its own Case Context.
  • Shared across users: anyone in that case/org can see and edit the same Case Context (subject to locking rules).
  • Structured: it’s stored as structured JSON and rendered as sections in the UI.
  • Used by the AI: when enabled in the Prompt Generator, the AI reads Case Context to write better prompts for your review workflows.

When should I use Case Context?

Use Case Context when:
  • You are working on a litigation or regulatory matter and want the AI to stay grounded in the actual pleadings and key facts.
  • Multiple lawyers or reviewers need a consistent understanding of:
  • Parties, relationships, and roles
  • The procedural and factual timeline
  • Relevance guidelines and “hot doc” indicators
  • Privilege rules and sensitive issues
  • You want the AI to write prompt templates that are tailored to this specific matter (rather than generic litigation).
Case Context is not:
  • A full memo or factum
  • A replacement for your internal case strategy documents
  • A place for privileged internal commentary about strategy (it should stay factual and review-focused)

How Case Context is structured

The Case Context editor is organized into the following fields:
  1. Parties & People
  • Who the parties are (plaintiff/defendant/applicant/respondent/regulator, etc.).
  • Key individuals (officers, employees, regulators, witnesses, etc.).
  • Counsel and firms, where known.
  • If it’s not obvious which party your firm represents, add a short note or mark it as unknown.
  1. Description & Timeline
  • A concise narrative of what the dispute or investigation is about.
  • Key factual and procedural events, in rough chronological order, with dates or date ranges where available.
  • Focus on pleadings and core background documents.
  1. Relevance & Issues
  • The main legal and factual issues in dispute.
  • What kinds of documents, communications, or events are typically responsive or non‑responsive.
  • Indicators of “hot” or especially important documents.
  1. Privilege Indicators
  • Who is (or may be) counsel, in‑house or external.
  • Typical privilege scenarios in this matter (e.g., internal investigation, board materials, regulator correspondence).
  • Any special privilege rules or caveats (e.g., joint defense, common interest).
  1. Collection Details
  • Custodians, sources, and date ranges that are actually in this Nuix case, if known.
  • Any gaps or limitations you know about (e.g., missing mailboxes, limited backup tapes).
  • This field is often blank early on; that’s okay.
  1. Additional Notes
  • Any other contextual information useful for reviewers that doesn’t clearly fit into the other sections (e.g., jurisdictional quirks, reputational sensitivities, known weak spots).
  • Keep it factual and review‑focused, not argument.
The AI also maintains an internal field for Open Questions and Last Updated, but these are not editable fields:
  • Open Questions are shown in the yellow follow‑up box after generating Case Context from documents. These are what you should provide to expand case context further. 
  • Last Updated is automatically set when you save structured context; it’s used by the agent but not edited directly.

How to open and view Case Context

  1. In the Noticia Go sidebar for your Nuix Discover case, look for the Case Context icon (stacked layers).
  2. Click the icon to open the Case Context modal.
  3. At the top you’ll see:
  • A short description of what Case Context is and how to use it.
  • A status indicator (Empty / Draft / Needs Input / Ready / Error).
  • Badges showing the number of selected documents and Context docs.
If Case Context is empty, the sections will be blank and status will show Empty.

Editing Case Context manually

You can always edit Case Context manually, regardless of whether you have already added documents.
  1. Open the Case Context modal.
  2. Ensure you have the editing lock:
  • If another user is editing, you’ll see a read‑only notice.
  • Try again later if you cannot obtain the lock.
  1. Fill in or update the text areas for:
  • Parties & People
  • Description & Timeline
  • Relevance & Issues
  • Privilege Indicators
  • Collection Details
  • Additional Notes
  1. Click Save in the bottom button row.
What happens when you click Save:
  • The frontend builds a structured JSON object (contextJson) from your edited fields.
  • The backend:
  • Stores that structured JSON in the Case Context table.
  • Builds a plain‑text version (contextText) in a fixed heading order so it can be used in other flows.
  • Marks the status as Ready (unless otherwise overridden).
Your changes are persisted to the database and will re‑appear when any user opens Case Context for this case.

Building Case Context from documents

Instead of (or in addition to) typing everything manually, you can ask the AI to build or refine Case Context from selected documents:
  1. In Nuix Discoverselect documents in the case.
  2. In Noticia Go, open the Case Context modal.
  3. Make sure:
  • You are connected to Nuix Discover (Noticia Go must be running inside the case).
  • You hold the Case Context lock.
  1. Use the “Add to context button (Layers icon) in the bottom button row:
  • This sends the selected docs’ extracted text plus current context to the Case Context agent.
  • The agent returns an updated contextJsoncontextText, and any follow‑up questions.
After a successful update:
  • The structured fields are refreshed with the AI’s updated summary.
  • The yellow follow‑up box appears if the agent has questions (for example“Which party do you represent?” or “Can you provide the pleadings?”).
  • The Source documents list shows which Main IDs/doc IDs are currently linked to Case Context, with their titles.

Viewing source documents linked to Case Context

In the modal you’ll see a “Source documents linked to Case Context section when there are any linked docs:
  • Each item shows:
  • Title (from the case, if available), or mainId if no title could be retrieved.
  • Below that, in small grey italics, the doc ID.
These are the documents that have been used to build or refine the current Case Context.

Follow‑up questions from the agent

After generating Case Context from documents, you may see a yellow box labeled:
> Follow‑up questions from Case Context agent
  • This box contains any clarifying questions the agent has about the matter (each on its own line).
  • These are not editable fields; they are prompts to help you improve your Case Context.
  • Once you’ve answered them by updating the relevant sections and/or adding better documents, you can:
  • Click Dismiss questions if they’re no longer useful, or
  • Regenerate Case Context from new documents, which may update the follow‑ups.

Locking and concurrency

To prevent conflicting edits, Case Context uses a simple lock:
  • When you open the modal and acquire the lock, you see the editor in editable mode.
  • If someone else is editing:
  • You’ll see a read‑only notice and cannot make changes.
  • Locks have a timeout; if a user abandons editing, others can acquire the lock later.
If you see an error indicating Case Context is locked, wait or coordinate with your team.

Using Case Context with the Prompt Generator

Case Context is currently integrated only with the Prompt Generator, not with the main prompt editor.
  • In the Prompt Generator modal, you’ll see a checkbox:
  • Include Case Context
  • When enabled, the generator uses the current Case Context to draft prompts tailored to your matter (e.g., review prompts that know who the parties are and what issues matter).
Best practice:
  • Keep Case Context reasonably up‑to‑date before using the generator for production prompts.
  • When revising strategy significantly (new pleadings, new key rulings), update Case Context and then re‑generate your prompt templates.

Best practices

1. Start from pleadings
  • Always prioritize pleadings (statements of claim/defense, applications, motions) and core orders.
  • Use the Add to context button with those documents first.
2. Keep it concise and focused on review
  • Write for experienced Canadian litigators, not lay readers.
  • Avoid narrative repetition; keep each section short and punchy.
  • Focus on what helps reviewers decide relevanceissue linkage, and privilege quickly.
3. Avoid speculative facts
  • Only include what is clearly supported by pleadings and key documents.
  • If something is uncertain or contested, note that explicitly in Additional Notes, not as a confirmed fact.
4. Keep privilege rules separate and clear
  • Use the Privilege Indicators section for clear rules and patterns.
  • Don’t mix privilege notes into general relevance guidance where they might be overlooked.
5. Use Additional Notes sparingly
  • This is for high‑value context (e.g., reputational risk, public‑facing facts) that helps reviewers understand why a document matters.
  • If a note clearly belongs in another section (e.g., a key event with a date), put it there instead.
6. Treat Open Questions as a to‑do list
  • When the agent asks questions in the yellow box, either:
  • Answer them in the structured fields, or
  • Provide the missing documents (e.g., pleadings) and regenerate.
  • Dismiss follow‑ups only once they’re handled or clearly not needed.

Troubleshooting

  • I save, but the fields revert
  • Ensure you’re saving without backend errors and that you hold the lock.
  • The Save button now sends structured fields (contextJson) to the backend and they should persist; if they revert, confirm the backend is reachable and the status shows success.
  • I see an error about documentTitle in GraphQL
  • The Case Context feature now uses the title field on the GraphQL Document type as documented in Docs/API-docs.md.
  • If you still see a documentTitle error, it likely means you’re on an older build or connecting to an environment with a different schema; contact support with the exact error.

Summary

Case Context is your shared, structured case synopsis that teaches the AI how to think about your matter:
  • Built from pleadings and a small set of key documents.
  • Edited via a structured, section‑based modal.
  • Persisted per (case, organization) and used by the Prompt Generator for smarter, matter‑aware prompts.
Used well, it reduces repeated explanation, keeps review teams aligned, and makes your AI outputs significantly more precise and useful.

Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article