Reviewers and investigators can now put plain-language questions to a matter's document set. The answers come back drawn from the documents themselves, with citations to each source and a log of every interaction.
The capability called AI Chat is the headline addition to a wider generative AI release for Nuix Discover, the company's legal review and eDiscovery platform. The audit trail is the point of difference Nuix is pressing, with every question and answer recorded for court use.
"AI Chat lets teams simply ask questions and get answers grounded in their case data; answers that are defensible and auditable," said Ilona Meyer, executive vice president, Nuix Discover.
Four further capabilities ship alongside it. Document Summaries generates a plain-language summary of each document as it is ingested. Nuix aims it at long contracts, dense attachments and lengthy email threads, so reviewers can orient to a matter faster.
Similar Documents surfaces material that matches in meaning, based on the summary of the document under review. It is designed to catch related material that keyword matching would miss.
Semantic Search locates responsive material by meaning rather than by exact terms. Nuix says this reduces the risk of missing documents that use different phrasing for the same concept.
Clustering and Visualisations groups documents by meaning and maps them visually. That gives review teams an early read on the composition of a collection. The map covers key issues, hot topics, concentrations, outliers and the relationships between clusters.
Nuix positions that last capability as a planning tool rather than a review tool. The stated purpose is to inform review strategy, batching and prioritisation from the first day of a matter.
Built into both stages
All four capabilities are built into early case assessment and into full review. They do not sit only at the triage stage.
Nuix argues this lets teams work from a single foundation of document understanding across the matter lifecycle. That runs from early triage through to full review.
"These GenAI innovations change how review teams engage with case data," Meyer said. "Document Summaries, Semantic Search, Similar Documents, and AI Clustering are built into both ECA and Review, so teams work faster at every stage."
"In ECA, they triage matters with far greater speed and accuracy than traditional approaches, giving teams the confidence to decide what to escalate for full review," she said. "Every insight is grounded in the source material, so the results are defensible."
Nuix Discover can be deployed on-premises, in the cloud or in a sovereign environment. That distinction matters where matter data cannot leave a jurisdiction. The platform runs on the Nuix Engine, which the company positions around processing accuracy and completeness of findings.
Document Summaries, Semantic Search, Similar Documents, and AI Clustering and Visualisations reach general availability on 4 September 2026.
AI Chat is available now through an early adopter program. General availability is set for December 2026.