By Neil Cameron, LITI lead analyst
Legalweek finally has a home worthy of its ambitions. The move to the North Javits Center – the gleaming new expansion of the Jacob K. Javits Convention Center that opened in 2021 on Manhattan’s far west side – gave this year’s event a physical presence that matched the scale of its agenda. The soaring glass halls, the natural light, the extra elbow room between booths: after years of navigating the cramped corridors of the original building, the upgrade felt long overdue. The industry has grown; the venue has caught up. There is one key caveat to the glory through, there is no bar in the venue – and no nice bar or restaurant closer than two blocks. Let us hope that good old American capitalist ingenuity will fix at least the latter before next year.
What greeted visitors on the show floor was a legal technology market in a confident, sometimes restless mood. Agentic AI was everywhere – in vendor pitches, session titles, and casual corridor conversation. But confidence has its own undertow, and not every claim survived close examination. What follows is a category-by-category review of the vendors we visited, followed by a session that stayed with us long after the lights came down.
Legal Research & AI Platforms
LexisNexis
LexisNexis arrived at Legalweek with the most significant product story of any research vendor on the floor: the integration of Anthropic’s Legal Plugin into Lexis+ with Protégé, its consolidated AI workflow platform launched into general US availability in February. The pitch – going from a single prompt to a validated, formatted, client-ready deliverable without ever leaving the Protégé environment – reflects a deliberate strategic choice to compete on integration rather than feature count. Where rivals are racing to add capabilities, LexisNexis is betting on coherence: one workspace, one governance layer, one citation authority in Shepard’s.
The Anthropic partnership, which deepens a collaboration that dates back to 2023 fine-tuning work on AWS Bedrock, adds hundreds of agentic AI workflow capabilities to what was already a library of more than 300 pre-built workflows. The framing from the company’s CTO Greg Dickason was pointed: the future of legal AI is not better chat, but coordinated agentic workflows that execute structured steps, orchestrate tools, and produce finished work product. The platform’s multi-model architecture – drawing on Anthropic, OpenAI and others – positions LexisNexis as a workflow destination rather than a model vendor. Global rollout across Canada, the UK, Europe and Asia Pacific continues through 2026.
The specific examples shown at the booth were instructive. A civil litigation workflow that analyses facts, extracts timelines, surfaces issues and drafts a case-assessment memo in a single end-to-end run. A judicial agentic workflow that drafts bench memos and opinions calibrated to an individual judge’s voice. These are not marginal improvements to existing research workflows; they are attempts to claim a new category. Whether the claims survive production-scale deployment – and whether the citation grounding that Shepard’s provides is sufficient to make the outputs genuinely defensible – will be the test.
The demonstrator was also kind enough to indulge an old man’s whim, and find a old UK case that was the bedrock of Judiciary Lexis (as it then was) training in the UK in 1980 – I know because I devised the UK training sessions at the time: R v Collins [1973] QB 100 is well worth a read, and so astonishingly odd that it has its own entry in Wikipedia.
Thomson Reuters / Westlaw
Thomson Reuters occupied a large booth at the show, with the next generation of CoCounsel Legal as its centrepiece – currently in beta and being demonstrated for the first time to a broader audience. They also had the largest PC touch screens at the show, which allowed a good presentation of the latest AI Deep Research capabilities.
The headline claim was striking: a fully autonomous legal assistant capable of producing what CTO Joel Hron has described as human-level work product. The demonstration involved analysing a complex antitrust matter, with CoCounsel independently designing its own research plan, querying Westlaw and Practical Law, cross-referencing uploaded documents, and producing a structured, citation-backed output without prompting at each step.
The architectural claim matters here. Thomson Reuters is at pains to distinguish this from a general-purpose LLM wrapped in legal branding. The agent, it says, is not following a predefined workflow or script; it is making dozens of independent decisions at each step, guided by Thomson Reuters’ own editorial knowledge infused into the product. The company is also developing a proprietary LLM specifically for legal, tax and compliance use cases – a move that would reduce dependence on external model providers. That ambition, backed by more than $200 million a year in AI investment, is credible. The milestone of one million CoCounsel users across 107 countries, announced just before Legalweek, gave the company strong momentum on the show floor.
Westlaw itself has continued to evolve as the research backbone of the CoCounsel stack. AI Deep Research – now extending to Practical Law as well as primary law, available in the UK and coming to the US in the first half of 2026 – provides the kind of multi-step, multi-source synthesis that genuinely changes research economics. Tabular Analysis, which allows users to upload up to 10,000 documents and query them with up to 100 questions simultaneously, is another product that does not fit easily into previous categories. The Thomson Reuters story at Legalweek 2026 was one of scale and integration – a company that has moved well past the point of needing to explain why legal AI matters.
Trellis
Trellis runs at a different altitude from the research giants, but its booth at Legalweek was consistently busy, and the reason is straightforward: it does something the others do not. Its database aggregates trial court data from more than 2,600 counties across 45 states – hundreds of millions of motions, briefs, and judicial rulings – making the historically fragmented US state court system searchable through a single interface. That data moat is what makes Trellis AI, its generative tools layer, genuinely differentiated rather than another wrapper around a foundation model.
The live training sessions on the show floor were well-attended, walking through specific tactical applications: assessing motion viability based on how similar motions have actually fared in a given jurisdiction; analysing opposing counsel’s prior behaviour; understanding a judge’s tendencies on specific issues. These are litigation-preparation tasks that have traditionally relied on institutional memory and manual research. Trellis compresses that work significantly. The AI layer adds case assessments, complaint analysis, document breakdown, and citation extraction – outputs that are generated at the click of a button rather than through prompt engineering, which the company has deliberately positioned as a user-experience principle rather than an afterthought.
The data quality question is the one to watch. Trellis is only as good as its court data, and coverage across counties and states is uneven. The company’s ability to maintain freshness and accuracy at scale will determine whether it can hold the competitive position it has built. But for litigators who live in the state trial court system, there is currently nothing quite like it.
DeepJudge
DeepJudge draws a different crowd from the research platforms – its customers are typically the knowledge management and IT leads at larger firms, rather than the attorneys themselves. The company’s core product is unique – an AI-powered search and retrieval engine that indexes a firm’s entire document corpus – DMS, SharePoint, HighQ, email archives, network drives – and makes it queryable through intent-based natural language search. The pitch is that generative AI is only as good as its retrieval, and that most firms’ internal knowledge is inaccessible – not because it doesn’t exist but because it can’t be found.
The numbers the company cites are striking: customers report more than 65 hours saved per user annually on search, and adoption rates exceeding 85 percent within two months of deployment. Scepticism about self-reported metrics is appropriate, but the underlying problem DeepJudge addresses is real and well-documented. The announcement at Legalweek of a formal integration with Thomson Reuters CoCounsel Legal – allowing attorneys to surface firm knowledge and Practical Law guidance in a single workflow – represents a meaningful upgrade. DeepJudge finds the relevant deal documents or prior briefs; CoCounsel layers in market standards and current law. The combination addresses the fragmentation problem that has made knowledge reuse so difficult in large firms.
The company recently closed a $41.2 million Series A led by Felicis, with continued backing from Coatue, which gives it the runway to execute on the partnership strategy. Founded by ex-Google researchers Paulina Grnarova, Kevin Roth and Yannic Kilcher, it has the technical credibility to match the fundraising. The Swiss origins are becoming less of an oddity in a market increasingly comfortable with European-headquartered legal AI companies.
eDiscovery & Document Review
Consilio
Consilio came to Legalweek as the self-proclaimed “world’s largest legal data AI technology provider” – a designation it has worked to earn through acquisition. The most significant of those was TrueLaw, a New York-based AI research lab founded by veterans of Google, Apple and Harvard Law School, acquired last summer. TrueLaw’s Expert Legal Model (ELM
) and narrative generation technology are now integrated into Consilio’s Aurora platform, adding automated timeline generation, contradiction detection, and firm-specific language model fine-tuning to what was already a substantial AI portfolio.
The Aurora platform itself continues to evolve as Consilio’s investigative intelligence layer, sitting above the eDiscovery infrastructure and translating large document sets – emails, contracts, financial records – into structured, navigable insights. The Guided AI PrivDetect product, which uses AI to streamline privilege review, was also on show. These are products aimed at the complex, high-stakes end of the market: global investigations, regulatory inquiries, antitrust matters. Consilio’s scale – offices and data centres across Europe, Asia and North America, ISO/IEC 27001:2022 certified – matters to the global in-house legal departments and BigLaw firms that make up its core client base.
The strategic direction is clear: Consilio is consolidating the eDiscovery services layer with genuine AI research capability, rather than simply integrating third-party models. The TrueLaw team is expected to drive continued development of the Aurora platform and bespoke client tools. Whether the research lab culture survives assimilation into a large managed services operation is the perennial question in all such acquisitions.
Knovos
Knovos positioned itself at Legalweek around two innovations that, taken together, reflect a broader shift in how enterprise legal technology buyers are thinking about AI governance: Bring Your Cloud (BYC) and Bring Your AI Model (BYAIM). BYC allows customers to deploy the Knovos platform in their own preferred cloud environment, satisfying data residency requirements and internal infrastructure policies. BYAIM goes further, allowing customers to bring approved AI models into Knovos workflows rather than being locked into whatever model the vendor has selected.
The governance framing is deliberate. Dharmesh Shingala, CEO, put it plainly: the future of legal technology is not just faster, but controllable, governed and defensible. For large law firms and regulated corporate legal departments under increasing pressure to demonstrate AI oversight to clients and regulators, this is a commercially astute message. The platform covers eDiscovery workflows (Knovos Discovery), legal operations and matter management (Knovos Manage), secure file collaboration with DRM controls (Knovos Rooms), arbitration workflow management (Knovos Arbitrate), and governance and compliance readiness (Knovos GRC). The breadth is a genuine differentiator for buyers seeking to consolidate legal operations infrastructure.
New at Legalweek was hybrid vectorised search – combining lexical precision with vector search to surface responsive content faster across large, complex datasets – and enhanced AI assistance for document summarisation and information retrieval. The BYAIM capability in particular speaks to a tension that is becoming more visible in the eDiscovery market: clients increasingly have AI preferences and governance requirements of their own, and they are less willing to accept vendor-selected models as a fait accompli.
eDiscovery AI (now part of HaystackID)
eDiscovery AI arrived at Legalweek having just been acquired by HaystackID, a managed services and technology provider focused on complex data and workflow challenges. The acquisition, announced in late February 2026, formalises a partnership that had already produced client-facing integrations. The combined entity exhibited separately – HaystackID at booth 513, eDiscovery AI at booth 403 – signalling either that the deal was done too late to cancel a booth (call me cynical); or that the brand and product identity are being preserved through the integration process, at least for now (call me more cynical).
eDiscovery AI’s focus is on helping legal teams move rapidly from data overload to actionable clarity, with custom GenAI workflows designed for legal, compliance, regulatory and cybersecurity use cases. The company describes its mission as advancing legal technology through innovative solutions that maximise efficiency and ensure privacy compliance. The HaystackID acquisition gives the company access to a broader managed services delivery capability and an established enterprise client base. The CoreFlex announcement from HaystackID – positioning it as a unified legal workflow spanning Slack, Microsoft Purview, structured chat, forensic collections and AI-enabled discovery – suggests the combined company is building toward a comprehensive incident-to-production workflow.
cimplifi
cimplifi occupies a particular niche in the eDiscovery market: a technology-forward managed services provider with a specific focus on RelativityOne-based delivery. The company has been active in guiding firms through the migration from Relativity Server ahead of the 2028 deprecation deadline, and at Legalweek continued to position itself as a specialist partner for that transition. Its service model combines project management, review management, and analytics expertise with the RelativityOne platform, targeting the mid-market and larger corporate legal departments that need hands-on support rather than pure software licensing.
The AI story at cimplifi is tied to the Relativity ecosystem – including Relativity aiR for Review, which provides AI-assisted coding with citation rationales designed to withstand judicial scrutiny. The company’s value proposition is in translating these platform capabilities into defensible, auditable workflows that clients can stand behind in litigation. In a market where the gap between AI capability and AI governance is widening, cimplifi’s operational focus is its differentiating credential.
Practice Management & Legal Operations
Filevine
Filevine‘s presence at Legalweek centred on its AI-driven contract review capabilities, specifically the integration of AI-assisted contract analysis directly into Microsoft Word for high-volume contract review workflows. The company, whose core platform serves plaintiff-side litigation firms and corporate legal departments with case management, matter management, client communications and document handling, has been pushing into contract lifecycle management as a natural extension of its workflow automation roots.
The AI contract review tools are aimed squarely at legal teams handling large volumes of agreements who need to flag issues, extract key terms and compare against playbooks without manual clause-by-clause review. The Word integration is a practical concession to where legal work actually happens – lawyers will not abandon their drafting environment for a separate application, and Filevine is sensible enough to meet them where they are. The company’s LEX Summit user conference, moving to San Diego in 2026, continues to build the kind of community loyalty that practice management vendors depend on for retention.
Litify
Litify is a Salesforce-native legal practice management platform that serves plaintiff law firms, insurance companies and legal departments with an integrated suite covering intake, matter management, workflow automation, document generation, reporting and client communication. Its architecture – built natively on Salesforce rather than integrating with it – gives it both the flexibility of the Salesforce ecosystem and the depth of a purpose-built legal layer on top of it.
At Legalweek, Litify continued to make the case for AI-enhanced intake optimisation and case strategy analytics as the practical frontiers for its client segment. The plaintiff firm market moves differently from BigLaw – volume, speed of intake, and predictive case value analytics matter more than sophisticated research tooling. Litify’s recent conference appearances, including with leadership at LitiQuest Plaintiff, have emphasised the AI adoption culture shift: not just deploying tools but embedding AI thinking into firm operations. The integration with Salesforce world-class App community and AI capabilities continues to be a practical advantage that more general practice management platforms cannot replicate – if you can afford it.
Tabs3
Tabs3 is a different proposition from the enterprise platforms on either side of it: a billing, practice management and financial management suite built specifically for small to mid-sized law firms, with a reputation for reliability, user-friendliness and strong customer support that has made it a staple of the US small-firm market for decades. The company was present at Legalweek and at TechShow 2026, its traditional hunting ground, maintaining the bar association affiliations and consultant network that underpin its distribution model.
The Tabs3 story at this point in the market is less about breakthrough innovation than about credible integration of AI into an established, trusted workflow. The platform covers time tracking, billing, trust accounting, general ledger and matter management through its PracticeMaster module, and its integrations with other legal software – including document management and payment platforms – give small firms a coherent operational core without enterprise-level complexity or cost. In a market increasingly dominated by well-funded cloud-native platforms, Tabs3’s durability is itself a competitive statement.
Knowledge Management & Specialist AI
airia
airia focuses on AI governance and model management for enterprise legal and professional services environments – a category that has moved rapidly from theoretical interest to purchasing priority as firms confront the practical reality of deploying multiple AI tools across their operations. The core product provides a governance layer that sits above AI model usage, enabling firms to manage which models are deployed, monitor how they are being used, enforce policy compliance, and maintain the audit trails that risk and compliance functions increasingly require.
The positioning resonates with a specific anxiety visible across the Legalweek floor: the recognition that deploying AI without governance infrastructure creates liability exposure that is difficult to quantify and harder to defend. airia’s argument is that firms need to think about AI governance before they can think clearly about AI capability – that the order of operations matters. In a market where vendors are competing on feature velocity, the governance-first pitch is a deliberate counter-programme.
Xperate
Xperate operates in the legal operations and matter management space, offering workflow automation and process orchestration tools for corporate legal departments and law firms seeking to bring more operational discipline to how legal work is initiated, tracked and completed. The platform’s focus is on the operational layer that sits between the client relationship and the substantive legal work – the intake, triage, routing, tracking and reporting functions that determine how efficiently a legal department actually functions.
At Legalweek, the conversation around Xperate centred on AI-enhanced triage and routing – using machine learning to intelligently categorise incoming legal requests and direct them to the appropriate resource, internal or external, based on matter type, complexity and cost parameters. For corporate legal departments under pressure to demonstrate value and manage outside counsel spend, this kind of operational intelligence is increasingly a board-level concern rather than a legal operations curiosity.
IPSA
IPSA brings a specialist lens to legal technology: its focus is on the investigation and due diligence market, providing AI-enhanced background research, risk screening and intelligence tools for law firms, financial institutions and corporations conducting pre-transaction or pre-engagement due diligence. In an environment where sanctions screening, beneficial ownership complexity and reputational risk have all escalated simultaneously, the manual processes that previously governed this work are under acute strain.
The centrepiece of the IPSA pitch at Legalweek, however, was not its investigative methodology but its deployment architecture. IPSA operates as a fully private, on-premise AI environment: the LLM runs inside the client’s own infrastructure, and no data – no queries, no documents, no outputs – ever leaves it. The company’s own positioning is blunt: “AI that never exposes your client data.” In the due diligence and investigation context, where the subjects of enquiry are often sensitive and the confidentiality obligations acute, this is not a marginal differentiator. It is the whole argument.
The AI layer that IPSA has developed augments human analyst judgment rather than replacing it – identifying patterns, surfacing adverse media, cross-referencing data sources and flagging risk indicators for review. At Legalweek, the emphasis was on the defensibility of the process as much as its speed: clients in regulated industries need to demonstrate that due diligence was conducted rigorously, and IPSA’s audit trails and provenance documentation are designed to support that demonstration. In a market where the conversation around AI data governance has become increasingly anxious – and where clients in banking, healthcare and life sciences are beginning to write AI data controls into outside counsel guidelines – the combination of investigative capability and sovereign deployment is a commercially precise proposition.
The CEO also had the discretion, and class, to sponsor and introduce the Judge’s session (see below) without any reference to his product.
A final note: Safeguarding justice and the rule of law – A reminder of why we’re here
The judges panel on Wednesday morning opened with a subject that few legal technology conferences would normally venture near: the physical safety of the people who run the courts. US District Judge Esther Salas, who moderated, spoke with composed authority about what happened to her family in July 2020, when a gunman posing as a delivery driver attacked her home in New Jersey, killing her twenty-year-old son Daniel Anderl and critically wounding her husband. She has since made judicial security her cause, and her account of the tactics now being used to intimidate judges – including coordinated pizza deliveries to judges’ homes, ordered in their deceased relatives’ names as a signal that their addresses are known – was both specific and chilling.
Judges Kenly Kiya Kato of the Central District of California, Karoline Mehalchick of the Middle District of Pennsylvania, and Mia Roberts Perez of the Eastern District of Pennsylvania joined her to discuss what it means to exercise judicial independence under conditions that have changed fundamentally in a decade. The conversation was candid, non-partisan, moving and quietly urgent. It was the best session at Legalweek 2026. The panel was not about legal technology in any direct sense. But it was, in another sense, about exactly what wider purposes legal technology and its conferences exist to serve: the administration of justice, the rule of law in a democracy and the institutional conditions that make it possible.
See also:
Comment: Legalweek’s GenAI mock courtroom may be the warning nobody heeded
The post Legalweek 2026: The floor report and a sobering reminder of why we’re here appeared first on Legal IT Insider.