Nebula AI Case Explorer™ | Frequently Asked Questions

AI basics

I'm not that familiar with AI terminology — is there a plain-language explanation somewhere? Yes — see AI 101: A Plain-Language Guide to AI and AI Case Explorer. It covers terms like large language model, hallucination, prompt, token, retrieval-augmented generation, and agentic AI, and ties each one directly to a feature in the product.

Does the AI ever just make things up? All generative AI can occasionally state something confidently that isn't accurate — this is usually called "hallucination," and it's a known limitation of the technology in general, not something unique to this product. It's the reason every relevance score comes with stated reasoning; every AI Chat answer comes with a citation back to a source document, and a persistent disclaimer reminds you to verify before relying on AI output.

General

What is AI Case Explorer? KLDiscovery's generative AI-powered case intelligence platform, built on the Nebula platform. It reads your case summary alongside an imported document set and surfaces relevance scoring, key themes, key people, event timelines and storylines, document clusters, and key documents — before a single search query is written or a single document is manually reviewed. See What Is AI Case Explorer?

What exactly does AI Case Explorer do that I can't do today in Nebula Cull? Cull is built for targeted search — you know what you're looking for and use Boolean queries and facets to find it. AI Case Explorer is a different way in: understanding what's in your data when you don't yet know what to search for. It surfaces themes, key people, timelines, key documents, and clusters automatically, and its output feeds back into Cull as new fields.

How is AI Case Explorer different from Nebula ECAi? ECAi is an AI based document classification feature inside Cull that helps you categorize documents into themes. AI Case Explorer is a purpose-built case intelligence platform with its own interface, generative-AI-driven relevance scoring with reasoning, and a tracked agree/disagree feedback loop. ECAi's capabilities have been absorbed into AI Case Explorer.

Do I need separate login credentials? No. AI Case Explorer is accessed through your existing Nebula login. See Getting Access and Understanding Roles.

Which browsers are supported? Chrome or Firefox, for the best experience.

Where does AI Case Explorer sit in my Nebula workflow? Between the Nebula ECA/cull and Review phases. Its findings can be displayed in Cull as overlay columns and will also be displayed in Nebula Review. See Getting AI Case Explorer Results into Nebula.

We use Relativity for review — can we still use AI Case Explorer? Yes. See Results in Nebula and Relativity for more information.

What does it cost? Pricing details are available from your KLDiscovery account representative.

Access and setup

Who can create a new project? Only users whose Nebula user group has the Admin permission. Everyone else with Case Explorer access can view, explore, and use AI Chat once a project exists. See Getting Access and Understanding Roles.

How long does analysis take? It depends on document volume — a few hours is typical for large sets. You don't need to stay on the page and monitor it. The onboarding process itself (setting up the project, before analysis runs) is designed to take well under an hour.

Do I need to know Boolean search to set up a project? No. The onboarding wizard walks you through project setup, case summary, theme review, and document source selection without requiring query syntax. If you don't already have a written case summary, a guided Case Summary Builder walks you through it field by field.

Why do my results look off? The most common cause is a case summary that's missing detail in a specific area — usually primary focus areas, key people, or relevance guidance. See Writing a Strong Case Summary.

What happens if I need to change my case summary after analysis has already run? Editing it directly rewrites your themes and can shift scoring, so it's treated deliberately. You can manually edit the case summary at any time from the AI Relevancy tab. Alternatively, you can use AI Case Explorer’s AI-assisted revisions to propose targeted changes based on relevance feedback you’ve already given. Every edit and re-run is saved as a version you can always get back to. See AI Relevancy and Refining Your Case Summary, Manually Editing Your Case Summary, and Project List and Version History.

What document types are supported? AI Case Explorer supports plain text versions of any document type Nebula processes and extracts text from — emails and their attachments, Word documents, PDFs, Excel files, transcripts of audio or video files. A configurable 300KB per-document text limit applies; documents over that limit, or excluded for other reasons, are flagged in an exceptions report rather than silently dropped. See Processing Status and Exceptions.

Using the product

What does a relevance score mean? Each document is scored into one of six levels, from least to most likely relevant: None, Highly Unlikely, Unlikely, Possibly, Likely, Highly Likely. See Viewing Documents or AI 101: A Plain-Language Guide to AI and AI Case Explorer.

Can I disagree with an AI relevance score? Yes. Open the document in the Document Viewer and submit relevance feedback — agree or disagree, ideally with a reason. That feedback is what powers AI Relevancy.

Can I edit the Key Themes the AI suggests? Yes, themes are fully editable and also act as a filter throughout the product.

How are Key People organized? Into three tiers: Key Person (who the AI is confident is likely central), Custodian (your data source's custodians), and Other (people the AI noticed but wasn't confident enough to flag as central). Showing "Other" instead of hiding it is intentional.

What's the difference between Key Docs and Saved Documents? Key Docs is a shared collection of documents flagged as potentially significant and worth a deeper look — by the AI or any team member. Saved Documents is your team's hub of documents you've chosen to save for later from the Document Viewer. See Saved Searches, Saved Docs, and AI Chat.

Can my whole team see saved searches and chats? Yes — Saved Searches & Docs and AI Chats are shared and visible to your whole project team, so saving something there means your team can see it.

How accurate is AI Chat, and can I trust its answers? Basic AI Chat provides cited answers pointing back to source documents — check the citations against the underlying documents, especially early on, to build confidence in the tool for your specific matter. A persistent disclaimer reflects this: AI can make mistakes, and outputs should be reviewed with professional judgment before reliance. While the Agentic Chat is also intended to return cited sources and document lists, it is in beta and some functionality may be incomplete or not function as expected.

What's the difference between Basic and Agentic AI Chat? Basic is a more direct retrieval-and-answer mode, good for quick factual lookups. Agentic (currently Beta) can reason through a question in multiple steps and synthesize findings across the whole corpus, showing its progress as it works. The mode locks once a conversation starts, so choose before your first message.

How do I know I can trust the AI's findings generally? The AI shows its work rather than acting as a black box: every relevance score comes with a specific reason, Key People surfaces who the AI wasn't confident about rather than hiding them, and you can track your own agreement rate as you review — a concrete number you can cite ("we reviewed 150 documents and agreed with the AI's relevance 93% of the time"). When you re-run an analysis, you can see exactly what has changed.