Nebula AI Case Explorer™ | Creating a Project
New projects are set up through a guided, 4-step wizard, available to users with the Case Explorer – Admin permission (see Getting Access and Understanding Roles). No Boolean query knowledge is needed. The whole flow is designed to take well under an hour for most matters. You can create more than one project per Nebula repository/matter.
Step 1 — Project
Select the Give your project a name (e.g., "Enron Fraud Investigation - Phase 1"), then provide your case summary one of two ways:
- Upload — You can author the Case Summary in a separate editor, such as Microsoft Word and then upload it to AI Case Explorer. A downloadable case summary template is available directly from this screen. You can drag and drop the file (PDF, DOCX, or TXT) or click to browse and select your file to upload.
- Text Editor — Alternatively, you can choose to use the built-in text editor to craft your case summary. A draft will be periodically saved automatically.
- Add a theme manually
- Edit a theme's name or description
- Deselect or delete a theme that doesn't belong
Tip: See Writing a Strong Case Summary before you start — this is the input that drives everything downstream. You can always reach out to your KLD representative to request help to create and evaluate your case summary’s effectiveness.
Once themes are generated, your case materials lock. Modifying the case summary after this point resets all themes, so make sure it's in good shape before moving on. If you need to change it later, use Modify Case Summary — you'll be warned that this can't be undone and will rewrite your themes.
Step 2 — Review Themes
Based on your case summary, the AI generates a set of key themes for the matter. From here you can:
Tip: We recommend no more than 15 key themes. Clear and distinct key themes will yield better outcomes.
Step 3 — Data Sources
Select the document source for the analysis — one or more collections and/or saved searches from Nebula. Once selected, you'll see a summary of total documents – this may not be a deduplicated count if a document is in more than one selected data source.
Tip: Use a representative sample set of data for your data prior to running an analysis. This can help you identify unexpected issues or gaps that can be addressed by adjusting the case summary used.
Step 4 — Project Summary
Review your full configuration — name, selected data sources, document count, active themes — before launching. You can choose to:
Once you launch, the analysis cannot be manually stopped, so it's worth a last check of your themes and data sources here. Processing time depends on document volume and can take a few hours for large sets — you don't need to stay on the page and monitor it.
There's currently no fine-grained progress indicator (e.g., a percent-complete bar) or notification when the analysis completes — check back on the Project List to see status. See Processing Status and Exceptions.
After analysis completes
You'll land on the Case Overview Dashboard, which summarizes what the AI found. From there you can explore documents, review key people and events, and start using AI Chat.