From 14,442 Participant Names to 39 Key Players

In complex discovery, identifying the important people should be straightforward. It rarely is. A large collection may contain thousands of participant names, document authors, display-name variations, and references to individuals within the text. The same person may appear differently across emails and documents, while unrelated individuals share similar names. The question for counsel was not simply which names appeared in the collection — it was which documents reflected direct communications involving the 39 key players, and which documents merely mentioned them without including them as a participant or author.

BlueStar began with the participant and author information already available in Relativity. The collection contained 14,442 normalized conversation participant names, 360 unique document authors, and 39 key players relevant to the matter. Next-Gen analyzed and mapped the full participant and author data to those 39 individuals — reducing thousands of name records into a consistent, matter-specific identification structure the legal team could search, filter, and organize throughout the review.

BlueStar applied two complementary workflows. Participant Identification focused on the people directly connected to each communication: email senders, recipients, and document authors. This allowed counsel to isolate actual communications involving a key player without retrieving every document that merely mentioned the person's name. Mention Identification expanded the analysis to names referenced within the document text — capturing situations where a key player was discussed in an email, report, or memorandum even though the person did not send, receive, or author it. Participants and mentions remained separately coded throughout, giving counsel two precise search options: find communications in which a key player directly participated, or find documents in which the key player was discussed or referenced.

BlueStar validated both workflows against human-review coding using a sample of 1,597 documents. Next-Gen Participant Identification produced 2,673 name identifications compared with 2,009 through human review — approximately 33% more. Mention Identification found an additional 724 references beyond what Participant Identification captured, producing 3,397 total name identifications, approximately 69% more than human review. Most importantly, Mention Identification captured 100% of the names identified by human reviewers in the validation sample. The validation also surfaced a quality-control issue in the human-review results: reviewers had coded 61 names across 44 documents even though those names could not be located in the document content — a type of unsupported coding that Next-Gen's traceable process made visible.

With the Next-Gen results loaded into Relativity, counsel could search for communications directly involving any of the 39 key players, separately identify documents that discussed those individuals, and respond more accurately to discovery requests involving communications between specific people. Instead of asking reviewers to recognize and code thousands of name variations across the collection, the legal team began with a validated map of the people who mattered and the documents connecting them to the case.