Promoted to Review: What Earns a Place in the Workspace

Most litigation workflows follow a familiar pattern. Data arrives from many sources, all of it goes into a review platform, and the workspace grows to match. Every document is staged, hosted, and processed before anyone can set aside what never belonged there.

That pattern has costs that are easy to overlook. The review platform is where eDiscovery gets expensive. Workspaces need setup, administration, and hosting, and all of it scales with volume. The AI assistants built into those platforms face a similar problem: they summarize, classify, and answer questions over whatever corpus they're given, and a corpus full of unrelated material is harder to reason over.

The difference is easiest to see in a real matter that ran the other way.

The situation

A highly regulated energy corporation collected data from about 40 mobile devices and other sources. After processing, the matter held 1,670,964 documents.[1] Under the legal team's previous workflow, all of it would have gone into a review platform before a single search was run.

What Consilio did differently

This time, Consilio ran the matter through Aurora CoreECI, our application for deciding what belongs in review before anything reaches it. Consilio's project management team applied the case team's 104 search terms and phrases across the full collection. They then added the family documents for every responsive result and promoted only that targeted set to review.

The result: 75,193 documents promoted to review, a 95.5% reduction from the original collection.

The corporation kept its existing review platform. CoreECI decided what went into it.

A better job for the review platform

A review platform does its best work on documents that deserve attention. Deciding upstream what earns a place there changes what the platform is asked to do:

• The workspace matches the matter, not the collection. There is less material to stage, host, and administer in the environment where those costs are highest. ‍
• The review platform's AI has a sharper corpus to work with. An assistant working over responsive documents and their families draws on material relevant to the case, not noise from the rest of the collection. ‍
• Reviewers see documents that already meet the case team's criteria. Attorney time goes to judgment calls, not to clearing out what never belonged.

Review as a promotion decision

Picture deciding at the curb which boxes come inside the house, instead of carrying them all in and sorting them in the living room. That's the shift. We stop asking what we can remove from review and start asking what earns a place in it.

We think about the shift in four steps: process broadly, investigate early, promote selectively, and review only what meets the criteria. Because the decision happens upstream, the workflow stays flexible, and new data sources can go through the same evaluation as the matter evolves.

Three questions for you and your team

Before your next large collection finishes processing and lands in your review platform, work through these questions with your case team:

1. Are you using review to discover what matters, or deciding what matters before review? If the full collection routinely enters review first, there may be an opportunity to move investigation upstream. ‍
2. What criteria can you apply before promotion? Search terms, responsive families, and other matter-specific criteria can define what needs to continue downstream. ‍
3. Is your review platform, and the AI inside it, working on the right corpus? If your workspace and your AI results reflect the whole collection rather than the case, narrowing upstream can improve both, without changing the review platform you already use.

A smaller review from an earlier decision

Some matters will always need substantial review. But a large collection doesn't have to become an equally large review population, and your review platform shouldn't have to carry one.

See how this matter ran from start to finish: Read the case study

[1] Results are based on matter data supplied by Consilio's Legal Services team. The figures reflect the matter at the time of analysis and may change as additional data is processed. See Consilio Legal Services Team, Matter Metrics (Sept. 9, 2026) (data on file with Consilio).