Nuix launches AI Chat as A$50m Discover business faces a growth reset after FY26 decline

Nuix Limited (ASX: NXL) is expanding generative artificial intelligence across Nuix Discover with a new AI Chat interface that allows lawyers and investigators to question case documents in natural language and receive synthesized answers linked back to the underlying evidence. The Australian investigative-analytics company has also introduced document summaries, semantic search, similar-document discovery, AI clustering and visualizations across early case assessment and full legal review, with those capabilities becoming generally available on September 4. AI Chat itself remains in an Early Adopter program and is scheduled for general availability in December 2026, giving Nuix several months to test accuracy, workflow integration and legal defensibility before a broader rollout. The timing is strategically important because Nuix reported a strong FY26 result in which group revenue rose 18.8% to A$263.2 million and Nuix Neo ACV surged 179% to A$78.5 million, yet the Discover product line, representing roughly A$50 million of annualised contract value, declined by around A$2.5 million. The new AI capabilities therefore have a measurable commercial assignment: help turn one of Nuix’s weaker product lines back toward growth while competing with legal-tech platforms that already offer conversational access to enormous document collections.

Nuix is positioning auditability rather than unrestricted automation at the center of the product. AI Chat generates answers from case material, supplies links or citations to supporting documents and records user interactions so reviewers can trace how an AI-assisted conclusion was reached. That architecture addresses one of the central problems facing generative AI in legal work, where an answer that sounds plausible but cannot be verified against evidence can create substantially more risk than value. For Nuix, the opportunity is to combine the conversational simplicity of modern AI with the forensic data-processing foundation on which its business was originally built.

Why could asking a case a question change legal review more than another AI document summary?

Document summaries save time, but they largely preserve the traditional structure of review. A lawyer opens a document, asks software to compress its contents and then moves to the next document. Conversational case analysis changes the starting point because a reviewer can begin with the legal or factual question and ask the system to find the relevant evidence across the corpus.

A lawyer investigating an internal dispute could ask which executives discussed a particular decision, when they first discussed it and what changed afterward. An investigator examining hundreds of thousands of communications could ask for evidence connecting particular individuals, transactions or dates without constructing every keyword combination manually. The AI can then synthesize an answer and point the reviewer back toward the underlying documents for verification.

That does not eliminate traditional search or human review. Legal questions can depend on context, privilege, intent, jurisdiction and nuances that a language model may misinterpret, while the relevant document might use terminology different from the wording of the user’s question. Nuix is therefore combining AI Chat with semantic search, clustering, similar-document identification and its established processing technologies rather than presenting chat as a substitute for the complete eDiscovery workflow.

The potential productivity gain comes from changing the order of investigation. Instead of reading large numbers of documents before deciding which factual patterns matter, teams can test hypotheses earlier and then investigate the supporting evidence. That could be particularly valuable during Early Case Assessment, when organizations are deciding whether a matter warrants expensive full-scale review.

Why does the audit trail matter more in legal AI than it does in a normal enterprise chatbot?

A conventional workplace chatbot can produce a useful response even when the user never needs to prove exactly where every sentence originated. Litigation and regulatory investigations operate under a different standard because evidence, discovery decisions and representations to courts may eventually be challenged by opposing counsel, regulators or judges.

Nuix says every AI Chat interaction is logged and that responses are grounded directly in case documents with supporting citations. That allows a lawyer to inspect the underlying material rather than treating the model output as an authoritative answer. The objective is not to eliminate hallucination risk but to make AI-assisted reasoning easier to verify and reconstruct.

The requirement is becoming more important as governments and professional bodies formalize expectations around lawyer use of artificial intelligence. California lawmakers, for example, recently advanced legislation requiring lawyers to take reasonable steps to verify AI-generated information and restricting the use of confidential or non-public information in inappropriate generative AI systems. Similar professional obligations around competence, confidentiality and supervision are already shaping legal AI deployment well beyond California.

Nuix’s ability to deploy Discover on premises, in the cloud or within sovereign environments therefore becomes commercially relevant. Law firms, government agencies and regulated corporations may want generative AI without moving sensitive evidence into an uncontrolled public service. Nuix also promotes Bring Your Own AI capabilities elsewhere within Nuix Neo Legal, allowing organizations to integrate preferred models while retaining greater control over data governance and model choice.

Is Nuix actually ahead of Everlaw and Relativity in conversational legal AI?

No. Conversational interaction with a legal document corpus is already becoming an established competitive capability, which means Nuix cannot rely on AI Chat alone as a durable differentiator.

Everlaw’s Deep Dive allows legal teams to ask natural-language questions across large document populations and receive citation-backed answers grounded in the underlying evidence. The company moved Deep Dive into general availability in late 2025 and has subsequently expanded its use across early case assessment, review, deposition preparation and trial strategy.

Relativity has taken a somewhat different route with aiR for Review, using generative AI to assess documents for responsiveness, identify issues and surface potentially important material. Relativity also provides natural-language explanations for model decisions, allowing reviewers to understand why individual documents were classified in particular ways.

The competitive question is therefore shifting from whether an eDiscovery platform contains generative AI to how deeply that AI is integrated with the platform’s existing data infrastructure. Nuix’s historic strength is its Nuix Engine, which can process large volumes of structured and unstructured information across more than 1,000 file types. If conversational AI can operate directly on the information that Nuix has already collected, processed, indexed and enriched, customers may avoid moving data between separate processing, review and AI tools.

That integration becomes Nuix’s stronger argument. AI Chat is unlikely to win customers simply because lawyers can type a question into a box. It becomes more defensible commercially if the system can interrogate unusually complex evidence collections while retaining the processing accuracy, auditability, deployment flexibility and data-sovereignty controls that customers already associate with Nuix.

Why does the A$2.5m decline in Discover ACV make December’s AI Chat launch a financial test?

Nuix’s FY26 results reveal a sharp contrast between its newer platform and the Discover franchise. Total Annualised Contract Value increased 13.9% to A$260 million, while Nuix Neo ACV surged 179% to A$78.5 million and represented 30% of group ACV at year-end. Nuix Neo had 135 customers compared with 101 at the half year.

Discover did not participate in that momentum. Management indicated that the product line, representing around A$50 million of ACV, declined by approximately A$2.5 million during FY26, partly because of the loss of a significant SaaS contract. Nuix has responded by appointing a dedicated Executive Vice President for Discover, increasing product investment and establishing a separate commercial model intended to restart growth.

The new generative AI release therefore arrives almost exactly where Nuix needs stronger product differentiation. Document summaries, semantic search, clustering and visualizations became generally available on September 4, while AI Chat provides another potential upgrade and cross-selling reason when it reaches full availability in December.

The size of the opportunity is meaningful without being transformational by itself. A product line near A$50 million of ACV represents almost one-fifth of Nuix’s A$260 million group ACV. Returning Discover to sustained growth could therefore contribute several percentage points to group expansion if AI capabilities improve customer acquisition, upselling or retention.

Success would also strengthen Nuix’s platform migration strategy. Management has found that existing customers moving from component products into Nuix Neo typically generate ACV uplift of around 30% to 50%, while a new Neo sale can be two to three times the size of a traditional non-Neo sale. Integrating more legal workflows and AI capability into the broader platform increases the number of reasons customers have to make that transition.

How strong was Nuix’s FY26 result before it committed another A$15m to accelerated R&D?

Nuix enters the legal AI expansion from a materially stronger financial position than a year ago. FY26 revenue increased 18.8% to A$263.2 million from A$221.5 million, while Adjusted Management EBITDA increased 60.4% to A$59.8 million. The corresponding margin expanded from 16.8% to 22.7%, demonstrating substantial operating leverage.

Statutory EBITDA increased 40.5% to A$66.9 million, and Nuix moved from a A$9.2 million statutory loss in FY25 to net profit of approximately A$16.5 million. Underlying cash flow increased 154% to A$51 million, while free cash flow climbed from only A$4 million to A$37.4 million. Nuix ended June with approximately A$49.9 million of net cash even after completing its acquisition of graph-analytics company Linkurious.

Management is not allowing all of that operating leverage to flow through immediately into FY27 profit. Nuix plans an additional A$15 million R&D accelerator investment intended to broaden its technology platform and address a larger market. As a result, management expects FY27 Adjusted Management EBITDA to be broadly similar to FY26 despite targeting further ACV growth.

That investment creates a higher standard for products such as AI Chat. Generative AI features are inexpensive to demonstrate but expensive to develop, secure, test and operate reliably at enterprise scale. If the additional R&D spending produces higher-value platform contracts and revives Discover growth, flat near-term EBITDA can be justified as reinvestment. If product adoption remains limited, investors could instead see the additional spending as postponing the operating leverage that FY26 finally demonstrated.

Why does Nuix Neo’s 179% ACV growth show what Discover needs to achieve next?

Nuix Neo has become the clearest evidence that the company’s transition from individual software components toward a unified platform can change contract economics. Neo ACV increased from around A$28.1 million at FY25 to A$78.5 million at FY26, a 179% expansion in twelve months.

The migration strategy matters because Nuix has a large established customer base that historically bought individual components for processing, investigations, review and other tasks. Moving those users onto a broader platform can expand the amount of technology deployed inside each organization without requiring Nuix to acquire an entirely new customer for every incremental dollar of ACV.

Discover complicates that strategy because legal review itself remains a large and competitive software category. Law firms and corporations may use different tools for collection, processing, hosting and review, while major competitors have built deep customer relationships around their own cloud platforms. Nuix must therefore make the integrated proposition materially better than preserving existing best-of-breed workflows.

AI can help connect those layers. A matter collected and processed through Nuix can be enriched with AI, searched semantically, clustered, reviewed and questioned conversationally without sending the evidence into a separate generative AI environment. The more steps Nuix can perform within one controlled architecture, the stronger the economic argument for consolidating spending on the platform.

Nuix’s FY27 ACV guidance of A$285 million to A$300 million implies growth of approximately 9.6% to 15.4% from the A$260 million FY26 base. Neo migrations will remain central to that target, but Discover returning to growth would make the upper half of the range easier to achieve than continued contraction in a roughly A$50 million product line.

Why does Linkurious make the long-term Nuix AI story broader than document review?

Nuix completed the acquisition of Linkurious in April 2026 for approximately A$27 million, adding graph visualization and analytics to its existing ability to process enormous volumes of unstructured data. Linkurious contributed around A$12 million of ACV at June 30 despite being included for only part of the year.

Graph technology becomes particularly interesting when combined with generative AI because investigations rarely revolve around documents in isolation. Legal and regulatory teams often need to understand relationships among people, companies, accounts, communications, transactions, locations and events. Graph analytics turns those relationships into structures that can be explored rather than forcing investigators to reconstruct networks manually from individual records.

Nuix has not announced that AI Chat currently performs conversational graph analysis across Linkurious data, so that should not be treated as an existing capability. The combination nevertheless illustrates the broader direction of the company’s platform strategy: ingest difficult data, identify entities and relationships, analyze evidence and ultimately allow professionals to interact with those findings through increasingly intuitive interfaces.

That could expand Nuix beyond conventional eDiscovery into financial crime, regulatory investigations, intelligence, cybersecurity and data privacy. The common asset is not the legal document itself but Nuix’s ability to make large and complicated data populations understandable.

The competitive opportunity is substantial, but so is the engineering challenge. Connecting forensic processing, legal review, graph analytics and generative AI requires consistent data models, security controls and user experiences across technologies developed at different times. The financial benefits will depend on whether customers actually buy the combined platform rather than continuing to purchase individual modules.

What does Nuix’s share-price recovery say about investor expectations after its FY26 turnaround?

Nuix closed at A$2.00 on September 4, gaining 5.3% during the final completed trading session of the week. The shares were about 4.3% below their August 28 close of A$2.09 but approximately 46% above the A$1.37 close recorded on August 4.

The stock’s 52-week range stands at approximately A$1.07 to A$3.35, leaving Nuix about 40% below its high despite the recent recovery. At A$2.00, the company had an equity market capitalization of approximately A$678 million based on around 338.8 million shares outstanding.

The most important recent move occurred on August 24, when Nuix released FY26 results and unveiled its expanded Discover AI capabilities on the same day. Shares jumped 34.4% to A$1.915 from A$1.425, with the financial results providing the more obvious fundamental catalyst because they showed higher-than-expected profitability, cash generation and Neo growth. It would therefore be misleading to attribute the rally specifically to AI Chat.

The stock subsequently reached A$2.15 on August 25 before retreating, showing that investors are still determining how much of the improved financial performance can be sustained through FY27. The A$15 million R&D accelerator effectively asks shareholders to accept another year of investment while management pursues a larger platform opportunity.

AI Chat can contribute to that argument, but investors will ultimately require contract evidence. Discover ACV returning to growth, rising Neo customer counts, larger migrations and stronger net dollar retention would show that Nuix’s technology investment is translating into recurring commercial value.

What will prove that Nuix AI Chat is more than another generative AI feature inside legal software?

The first proof point comes in December when AI Chat is scheduled to move from Early Adopter status into general availability. Nuix will need to show that feedback from real legal users produces a system reliable enough for broad deployment across matters where evidence may be sensitive, commercially consequential and eventually scrutinized in court.

Usage will then matter more than availability. The strongest indicator would be customers repeatedly using conversational interrogation across large matters rather than trying the feature only during onboarding. Questions asked per matter, cited documents reviewed after responses and time saved during Early Case Assessment would provide useful operating measures, although Nuix has not committed to publishing those metrics.

Discover ACV provides the clearest financial test. The business declined by roughly A$2.5 million during FY26 while Neo expanded dramatically. If the enhanced AI portfolio contributes to renewed Discover growth during FY27, the release can be viewed as part of an effective commercial reset rather than simply a response to competitive feature pressure.

Accuracy and defensibility remain the final hurdle. Lawyers increasingly have access to general-purpose AI systems capable of summarizing documents and answering questions, so a specialist legal platform needs to justify its cost through better governance, integration and trust. Nuix’s source citations, interaction logging, deployment flexibility and forensic processing provide a credible foundation, but competitors such as Everlaw and Relativity are addressing many of the same requirements.

Nuix therefore enters the next stage of legal AI with an unusual combination of momentum and pressure. The wider company has just produced its strongest financial improvement in years, Neo ACV is expanding rapidly and cash generation has strengthened, yet Discover has not kept pace. AI Chat gives Nuix another chance to turn its deep data-processing heritage into a modern legal workflow that begins not with a keyword or document list, but with a question. December will show whether customers find that capability useful enough to help reverse the A$50 million product line’s recent decline.

What are the key takeaways from Nuix’s generative AI expansion in legal review?

• Nuix has introduced AI Chat for Nuix Discover, allowing legal reviewers and investigators to ask natural-language questions of case data.
• Responses are grounded in underlying documents and include supporting citations, while interactions are logged to support auditability.
• AI Chat is currently available through an Early Adopter program and is scheduled for general availability in December 2026.
• Document Summaries, Semantic Search, Similar Documents, AI Clustering and Visualizations became generally available on September 4.
• Nuix Discover can be deployed in cloud, on-premises or sovereign environments, an important consideration for sensitive legal and regulatory data.
• Nuix is not first to conversational legal AI, with Everlaw Deep Dive and Relativity aiR already providing competing generative AI workflows.
• FY26 Nuix revenue increased 18.8% to A$263.2 million and Adjusted Management EBITDA rose 60.4% to A$59.8 million.
• Nuix Neo ACV surged 179% to A$78.5 million, while the roughly A$50 million Discover product line declined by approximately A$2.5 million.
• Nuix expects FY27 ACV of A$285 million to A$300 million but plans an additional A$15 million R&D accelerator investment that will constrain near-term EBITDA growth.
• Discover returning to ACV growth after AI Chat reaches general availability will provide one of the clearest financial tests of whether Nuix’s legal AI investment is working.