Capability, technology, practice, and security/compliance vocabulary. Independent definitions. No vendor PR.
A quantitative measure of how often an AI system produces correct outputs on a defined test set — critical for evaluating legal AI tools where errors carry professional responsibility risk.
SecurityEU AI Act Article 53 requires general-purpose AI providers to publish training data summaries, copyright policies, and technical documentation for EU market access.
Tech / ModelAI bias in legal contexts refers to systematic errors or disparate outcomes in AI model outputs caused by imbalances in training data, model design, or task framing — potentially producing results that disadvantage certain parties, jurisdictions, or case types.
CapabilityAI-generated chronological reconstruction of case facts from documents, emails, transcripts, and filings; must be verified against source documents before reliance.
SecurityA lawyer's working knowledge of AI tools sufficient to use them effectively, supervise outputs, and meet the professional duty of technological competence.
CapabilityAI tools that assist contract negotiation by suggesting redlines, explaining counterparty language risks, or drafting counter-proposals based on the firm's playbook.
SecurityPrinciples guiding fair, transparent, and accountable use of AI in legal practice, including bias prevention, explainability, and professional responsibility.
SecurityFrameworks, policies, and oversight mechanisms that law firms and legal departments use to manage AI adoption responsibly.
EU RegulationA structured set of policies, roles, controls, and accountability mechanisms governing how an organization develops, procures, and deploys AI — required under the EU AI Act for high-risk use cases.
Tech / ModelAI hallucination in legal research is when a generative AI system produces case citations, statutes, or holdings that appear authoritative but are factually false or entirely fabricated.
SecurityA documented plan for detecting, containing, and remediating failures of AI systems — including legal AI tools — covering output errors, data breaches, and model misbehavior affecting client matters.
CapabilityAI systems that rank incoming potential client inquiries by conversion probability, case value, or fit criteria based on intake responses, source, and case type.
CapabilityAI legal billing uses machine learning to automate time capture, invoice review, billing guideline enforcement, and spend analytics for law firms and in-house legal departments.
EU RegulationThe legal framework — including the EU AI Liability Directive — governing who bears responsibility when AI systems cause harm, defects, or errors in commercial or legal contexts.
SecurityThe foundational ability to understand how AI systems work, evaluate their outputs critically, and engage intelligently with AI-related legal and policy issues.
Tech / ModelA standardized documentation artifact describing an AI model's intended use, performance characteristics, limitations, and training data — essential for legal AI vendor due diligence.
Tech / ModelAnchoring AI-generated text in specific retrieved source documents, reducing hallucination; a grounded response cites the specific passage supporting its claim.
CapabilityThe process of confirming AI-generated legal content — citations, summaries, fact characterizations — is accurate before use; a professional responsibility obligation that does not shift to the AI.
SecurityAdversarial testing of a legal AI system by deliberately attempting to induce failures — hallucination, bias, data leakage, prompt injection — to identify vulnerabilities before deployment.
CapabilityAI-condensed summaries of legal documents that preserve legally material facts; used on depositions, contracts, and case opinions, with lawyer verification required.
EU RegulationThe EU AI Act's requirement that providers of certain AI systems disclose their AI nature to users, enabling informed interaction and supporting accountability in legal AI deployments.
Legal PracticeThe use of AI tools to prepare deposition outlines, analyze prior testimony for inconsistencies, and generate real-time transcript summaries during depositions.
CapabilityUsing AI to generate or complete legal text — contracts, motions, briefs, correspondence — based on lawyer prompts or templates; lawyer reviews and edits before use.
CapabilityAI tools that analyze time entries, suggest billing codes, flag write-off risks, or draft narrative descriptions to reduce billing write-offs and improve invoice compliance.
CapabilityAI-powered legal intake automates first contact with prospective clients — qualifying leads, gathering case facts, checking conflicts, and routing to attorneys using natural language understanding.
CapabilityAn iterative ML approach in eDiscovery where the model continuously updates relevance predictions as reviewers code documents, prioritizing the most uncertain documents for review.
CapabilityAgentic AI in legal refers to AI systems that execute multi-step legal tasks autonomously — drafting, reviewing, routing, escalating — without requiring a prompt at each step.
Tech / ModelAn AI-driven multi-step legal process — such as intake to routing to drafting — that executes autonomously across defined stages without per-step human prompting.
SecurityAttorney-client privilege is the legal doctrine that protects confidential communications between a lawyer and client made for the purpose of seeking or providing legal advice, shielding those communications from compelled disclosure in legal proceedings.
SecurityHow attorney-client privilege applies when AI tools process confidential legal communications, and risks of inadvertent waiver through AI vendor data handling.
SecurityAn audit log is a chronological, tamper-evident record of system activities — including user logins, document accesses, queries, and configuration changes — that enables security monitoring, compliance verification, and investigation of incidents in legal AI environments.
SecurityA tamper-evident record of AI system activity—queries, outputs, user actions, and access events—used to support oversight, accountability, and compliance documentation.
Legal PracticeA contract provision that automatically extends a contract term unless a party gives timely notice of non-renewal; tracked by CLM tools to prevent unintended renewals.
Bates numbering assigns a unique sequential identifier to every page of every document produced in litigation, enabling parties, witnesses, and courts to cite exhibits unambiguously.
Legal PracticeBates stamping is the process of applying sequential identification numbers to each page of documents produced in litigation, investigation, or transactional due diligence, enabling parties to precisely reference and track specific pages throughout the proceeding.
SecurityTime or fees a firm removes from a client invoice, increasingly scrutinized as AI reduces task duration and raises questions about value-based billing.
SecurityThe legal and regulatory obligation to notify affected individuals, supervisory authorities, and sometimes the public when a security incident exposes personal or privileged legal data.
CapabilityA brief analyzer is an AI tool that reads a legal brief and automatically extracts arguments, identifies cited authorities, assesses argument structure, and surfaces potential weaknesses or gaps in the legal reasoning.
Legal PracticeBrief writing is the process of preparing written legal arguments submitted to a court or tribunal, requiring integration of factual record evidence, relevant legal authority, and persuasive narrative to support a party's requested legal outcome.
AI modeling of the likely outcome of litigation based on case facts, jurisdiction, judge history, and analogous precedents to inform settlement or trial strategy.
CapabilityCitation validation in legal AI verifies that every case, statute, or regulation cited by an AI system actually exists, is accurately quoted, and still stands as good law — the essential check against hallucination.
Legal PracticeA legal research tool that tracks the subsequent history and treatment of a case or statute, enabling lawyers to confirm whether authority remains valid and binding.
CapabilityAI identification of contract clauses deviating from a firm's standard position, flagging for review; requires a configured playbook defining what 'standard' is.
CapabilityA clause library is a structured repository of pre-approved, standardized contract clauses that lawyers and legal teams can access when drafting, negotiating, or reviewing agreements, often integrated with AI tools for automated clause selection and insertion.
Legal PracticeA contract provision requiring a party to return previously paid compensation or consideration upon occurrence of specified events; tracked in AI-assisted contract management.
Legal PracticeA systematic review comparing an organization's current practices against applicable regulatory requirements to identify deficiencies and prioritize remediation.
Legal PracticeCompliance monitoring is the ongoing process of tracking regulatory requirements, legal obligations, and internal policies to ensure an organization's operations remain within applicable legal standards — often supported by AI tools that flag changes in regulations and potential violations.
Tech / ModelHardware-level encryption using Trusted Execution Environments that protects data even during AI processing, so cloud providers cannot access client data while the model runs.
SecurityIn the legal AI context, confidentiality refers to the obligation of lawyers and legal AI vendors to protect client information from unauthorized disclosure, and to the technical and contractual measures that implement that protection when client data is processed by AI systems.
CapabilityConflict check AI is software that automates the identification of potential conflicts of interest by searching a firm's client and matter database against new prospective client or adverse party information.
SecurityUsing AI and automated database search to screen new clients and matters against existing relationships, identifying potential conflicts of interest before representation begins.
SecurityA conflict of interest in legal practice arises when a lawyer's representation of one client is materially limited by responsibilities to another client, a former client, a third person, or the lawyer's own interests — requiring disclosure, consent, or withdrawal from the conflicted representation.
EU RegulationThe EU AI Act's mandatory pre-deployment verification process confirming a high-risk AI system meets safety, transparency, and accuracy requirements before market placement.
Tech / ModelThe context window is the maximum amount of text — measured in tokens — that a large language model can process at one time, determining how much document content, conversation history, and instructions the model can consider when generating a response.
Tech / ModelAn eDiscovery review method where the AI updates its relevance predictions after every reviewer decision, continuously prioritizing the most likely-relevant documents.
Legal PracticeExtracting key data points from contract text into structured fields — parties, term, governing law, renewal dates, payment obligations, liability caps; AI compresses this from minutes to seconds per contract.
CapabilityAutomated routing of contracts through defined approval steps — legal, finance, executive sign-off — with an audit trail of approvals, reducing bottlenecks in high-volume contracting environments.
Legal PracticeEnd-to-end management of contracts from initiation through execution, performance, renewal, and termination; AI-enhanced CLM automates drafting, routing, negotiation, execution, and obligation monitoring.
Legal PracticeStructured data describing a contract — parties, effective date, expiration, governing law, contract value, renewal type — stored separately from full text; AI extracts metadata at scale to enable portfolio analytics.
Legal PracticeA legal team's documented negotiation positions, approved fallback language, and escalation rules that guide AI-assisted contract review and redlining.
Legal PracticeA centralized system for storing, organizing, and retrieving executed contracts, enabling search, reporting, and obligation tracking across a contract portfolio.
Legal PracticeContract review is the legal process of analyzing a contract's terms, conditions, and obligations to identify risks, ensure compliance with applicable law, assess alignment with the client's interests, and negotiate or recommend changes before execution.
CapabilityContract Review AI is software that uses natural language processing to automatically identify, extract, and flag clauses, risks, and deviations from standard terms in legal contracts.
CapabilityAI-generated numeric or categorical risk scores assigned to contracts based on clause-level analysis and deviation from standard positions, helping prioritize contracts needing lawyer review.
Legal PracticeAI analysis of the opposing party's contract positions, negotiation patterns, and risk profile to inform legal strategy during commercial contract negotiations.
EU RegulationThe transmission of personal data from one jurisdiction to another, subject to GDPR transfer mechanisms such as Standard Contractual Clauses or adequacy decisions when EU data is involved.
A contract required by GDPR between a data controller and processor, governing how personal data may be handled, secured, and returned or deleted.
SecurityWhere a legal AI vendor physically stores and processes client data — a compliance requirement under GDPR, data sovereignty laws, and attorney confidentiality obligations.
Tech / ModelA subset of machine learning using multi-layered neural networks that powers contract clause extraction, semantic search, and LLMs; modern legal AI tools are predominantly deep learning systems.
Legal PracticeDeposition preparation is the process of organizing case facts, reviewing documents, developing questioning strategies, and preparing witnesses before a deposition to ensure effective examination or cross-examination of the deponent.
CapabilityAI tools that process deposition transcripts to surface inconsistencies, generate chronologies, and create cross-examination outlines, reducing prep time significantly.
CapabilityDiscovery AI is software that applies machine learning and natural language processing to litigation discovery, automating document review, relevance classification, and issue identification across large document collections.
Tech / ModelSplitting legal documents into smaller segments for AI processing within finite context windows; chunk size and overlap strategy affect retrieval quality and contract review accuracy.
CapabilityDocument Drafting AI is software that uses large language models to generate, edit, or refine legal documents — including contracts, briefs, letters, and pleadings — based on lawyer-provided instructions or templates.
Legal PracticeDocument production is the process of delivering to opposing parties in litigation or investigation the set of documents that are responsive to discovery requests, non-privileged, and within the scope of the applicable discovery order or agreement.
Legal PracticeDue diligence is the systematic investigation of a company, transaction, or legal matter to identify material risks, liabilities, and issues before a deal closes, an investment is made, or a legal decision is taken.
CapabilityAI-powered review of large document sets in M&A, financing, or real estate transactions to identify risks, obligations, and anomalies; AI flags issues, lawyers assess materiality.
Electronic submission, review, and approval of legal invoices — typically following LEDES billing standards — enabling AI-assisted auditing of time entries for compliance with outside counsel guidelines.
CapabilityE-discovery (electronic discovery) is the process of identifying, preserving, collecting, reviewing, and producing electronically stored information in response to litigation, investigations, or regulatory demands.
SecurityRegulation 2024/1689, the world's first comprehensive AI law, classifying AI systems into four risk tiers with obligations applying to providers and deployers including law firms.
EU RegulationThe EU's comprehensive AI regulation, in force August 2024, imposing risk-tiered obligations on AI developers and deployers — with legal sector compliance requirements escalating through 2026–2027.
EU RegulationThe requirement that personal data of EU residents be stored and processed within EU borders, affecting cloud-based legal AI deployments under GDPR and national data sovereignty laws.
Legal PracticeAn M&A deal term making part of the purchase price contingent on the target's post-closing performance; a complex obligation tracked by AI-assisted contract and deal tools.
Tech / ModelAn embedding is a numerical vector representation of text — such as a word, sentence, or document — produced by a machine learning model, enabling AI systems to measure semantic similarity between texts and retrieve relevant information.
Legal PracticeAI-assisted drafting and review of employment contracts, including offer letters, non-compete clauses, IP assignment provisions, and severance terms.
SecurityEncryption at rest refers to the protection of stored data through cryptographic encoding, so that files, databases, and backups on storage media are unreadable without the appropriate decryption key — a baseline security control required for legal AI tools handling confidential client information.
SecurityAn engagement letter is a written agreement between a lawyer and client that defines the scope of the legal representation, fee arrangements, billing practices, and terms governing the attorney-client relationship — and increasingly, the terms under which AI tools may be used in the representation.
Legal PracticeThe process of identifying, preserving, collecting, processing, reviewing, and producing electronically stored information in litigation or regulatory investigations under FRCP and equivalent rules.
Alternative contract language pre-approved by legal for use when a counterparty rejects preferred terms, codified in a playbook for AI-guided negotiation.
Tech / ModelA model's ability to adapt to a new legal task from 2-10 examples provided in the prompt; more accurate than zero-shot for novel tasks, less expensive than fine-tuning.
Tech / ModelFine-tuning is the process of further training a pre-trained large language model on a domain-specific dataset to improve its performance on tasks in that domain, such as legal document analysis, contract drafting, or jurisdiction-specific research.
SecurityA fixed-price billing model where AI efficiency gains are absorbed into predictable project fees rather than passed through as reduced hourly billings.
Legal PracticeA contract provision excusing performance when extraordinary events beyond a party's control prevent fulfillment; a common focus in AI-assisted contract risk review.
EU Regulation 2016/679 governing personal data collection, processing, and transfer for EU residents — directly applicable to law firms using AI tools on EU client matters.
SecurityThe General Data Protection Regulation (GDPR) is the European Union's comprehensive data protection law, establishing requirements for how personal data of EU residents must be collected, processed, stored, and transferred — directly affecting how legal AI tools handle client and matter data.
SecurityGDPR Article 22 gives individuals the right not to be subject to purely automated decisions that produce legal or similarly significant effects.
SecurityUsing AI tools to identify, manage, and document compliance obligations under the EU General Data Protection Regulation across organizational data practices.
How the Health Insurance Portability and Accountability Act applies when AI tools process protected health information in healthcare legal matters.
SecurityThe Health Insurance Portability and Accountability Act (HIPAA) establishes federal standards for protecting individually identifiable health information, creating compliance obligations for healthcare lawyers and legal AI tools that process protected health information (PHI) in connection with healthcare matters.
Tech / ModelHallucination in legal AI refers to instances where an AI model generates factually incorrect, fabricated, or unsupported output — such as nonexistent case citations, invented statutes, or inaccurate summaries of legal holdings — presented with apparent confidence.
SecurityHarmonised standards are voluntary EU technical specifications that, when followed, create a legal presumption that an AI system complies with the EU AI Act's requirements.
EU RegulationAn AI system classified under Annex III of the EU AI Act as posing significant risk to health, safety, or fundamental rights, subject to conformity assessment before deployment.
EU RegulationThe EU AI Act's mandate that high-risk AI systems be designed to allow human monitoring, intervention, and override — directly applicable to legal AI tools used in client-facing or adjudicative contexts.
Tech / ModelCombines on-premise and cloud AI processing — sensitive client data stays on firm infrastructure while non-sensitive processing uses cloud AI — addressing data residency concerns with added architectural complexity.
IP filing refers to the formal submission of applications to protect intellectual property rights — including patents, trademarks, and copyrights — with relevant government authorities, requiring precise documentation, adherence to procedural requirements, and accurate legal description of the protected subject matter.
Legal PracticeAI-assisted drafting, review, and management of intellectual property license agreements, including royalty structures, field-of-use restrictions, and term obligations.
SecurityISO/IEC 27001 is an internationally recognised standard requiring organisations to establish and maintain a certified Information Security Management System (ISMS).
SecurityThe international information security management standard whose certification signals that a legal AI vendor has implemented systematic controls over data confidentiality, integrity, and availability.
EU RegulationThe first international standard for AI management systems, providing a framework for responsible AI development and deployment — increasingly referenced in legal sector AI governance.
Legal PracticeA contract provision obligating one party to compensate another for specified losses or liabilities; among the highest-risk clauses flagged in AI contract review.
Tech / ModelIn AI, inference is the process of running a trained model to generate outputs from new inputs — as distinct from training, which creates the model. Every time a lawyer submits a query to a legal AI tool, inference occurs.
Tech / ModelAI systems that capture, organize, and surface a legal team's historical matter knowledge — past positions, precedents, and playbook decisions — to inform current work.
A large language model (LLM) is an AI system trained on large volumes of text data to predict and generate human-like text; it serves as the core engine underlying most legal AI tools for research, drafting, and document analysis.
CapabilityLegal AI refers to software systems that apply machine learning and natural language processing to automate or assist with legal tasks such as contract review, research, drafting, and compliance monitoring.
Legal PracticeThe process by which law firms and legal departments evaluate, implement, and integrate AI tools into legal workflows, including change management, training, and professional responsibility compliance.
CapabilityA configured AI system that autonomously executes multi-step legal workflows — research, summarize, draft, cite-check — without per-step prompting.
Tech / ModelA standardized test evaluating AI model performance on defined legal tasks — bar exam questions, clause extraction, citation accuracy; notable benchmarks include LegalBench and vendor hallucination rate studies.
Tech / ModelA standardized evaluation measuring an AI system's accuracy, reliability, or performance on defined legal tasks — used to compare tools and validate fitness for professional use.
Tech / ModelSystematic AI model outputs that disadvantage certain groups due to training data patterns; documented examples include eDiscovery tools underperforming on non-English documents and risk score racial disparities.
SecurityA formal credential verifying that a lawyer or legal professional has demonstrated defined competencies in using AI tools in legal practice.
SecurityStructured curricula offered by law schools, bar associations, and legal tech organizations that train and credential legal professionals in AI tool use and governance.
SecurityA firm or department's written rules governing which AI tools are approved, how they may be used, and who is responsible for oversight and compliance.
SecurityThe process law firms and legal departments use to evaluate, select, contract, and onboard AI vendors while managing security, compliance, and ethical risks.
CapabilityAn isolated testing environment where lawyers evaluate AI tools against representative tasks without exposing live client data, used in procurement due diligence and pre-deployment benchmarking.
CapabilityLegal citation check is the process of verifying that cited cases exist, that quoted language accurately reflects the decision, and that cited authority remains valid and has not been overruled or significantly limited by subsequent decisions.
CapabilityStructured storage, version control, access management, and retrieval of legal documents organized around the client matter — not a generic folder hierarchy.
CapabilityA legal hold (also called a litigation hold or preservation notice) is a formal directive issued to individuals within an organization requiring them to preserve all potentially relevant documents and data when litigation or investigation is reasonably anticipated.
SecurityUsing AI to identify, notify custodians, and track preservation obligations when litigation or investigation triggers a duty to preserve electronically stored information.
SecurityThe systematic capture, organization, and retrieval of a legal organization's institutional knowledge—precedents, playbooks, and expertise—increasingly AI-assisted.
Legal PracticeLegal operations is the business management function within corporate legal departments responsible for technology, vendor management, financial oversight, and process improvement.
SecurityQuantitative metrics used by legal operations teams to measure departmental performance, cost efficiency, matter cycle times, and vendor management effectiveness.
CapabilityThe operational platform of a law firm — centralizing matter tracking, time, billing, document management, client communication, trust accounting, and reporting in one system.
CapabilityLegal Research AI is software that uses natural language processing and large language models to retrieve, summarize, and analyze case law, statutes, and secondary sources in response to natural language queries.
CapabilityAI-powered analysis of legal department expenditure by matter type, outside firm, and practice area, identifying cost drivers, billing guideline violations, and budget anomalies.
Legal PracticeThe systematic tracking, analysis, and control of an organization's legal expenditures — including outside counsel fees, e-billing, and budget forecasting — increasingly AI-assisted in corporate legal departments.
CapabilityCentralized tracking and AI-powered review of outside counsel invoices and legal expenditures to enforce billing guidelines and control costs.
CapabilityAI-driven automation of repeatable legal processes — document routing, approval chains, deadline tracking — reducing manual steps; ROI clearest in high-volume transactional environments.
Legal PracticeA contract clause capping the maximum damages one party can recover from another; routinely flagged and benchmarked by AI contract review tools against standard thresholds.
CapabilityLitigation analytics applies statistical and machine learning methods to court data — judicial rulings, motion outcomes, and case results — to inform litigation strategy and risk assessment.
Legal PracticeStructured analysis of the probability, cost, and exposure of litigation using AI-generated insights from case law, damages data, and opposing counsel history.
Legal PracticeLitigation support encompasses the services, tools, and processes that assist lawyers in preparing and managing cases — including document management, e-discovery, evidence organization, trial preparation, and case analysis.
A provision allowing a buyer to exit an M&A deal if the target experiences a material adverse change between signing and closing; central to AI-assisted deal review.
CapabilityAI-assisted review of master service agreements flagging indemnification scope, IP ownership issues, liability cap deviations, and data processing obligations across complex, interdependent clauses.
Tech / ModelAlgorithms that learn patterns from labeled legal data — relevance decisions, risk labels, outcome records — to make predictions on new documents or cases; TAR is the most established application.
Legal PracticeMata v. Avianca is the 2023 SDNY case in which attorneys were sanctioned $5,000 for submitting a brief citing six non-existent cases fabricated by ChatGPT.
CapabilityAI-powered tools automating new client and matter intake — smart forms, conflict screening, case value estimation, and routing — reducing intake time and improving data completeness.
Legal PracticeThe complete arc of a legal matter — from intake and engagement through active work, billing, and closure — increasingly managed and analyzed using AI-assisted practice management tools.
CapabilityCentralized tracking of all information tied to a single legal case or transaction — documents, deadlines, parties, tasks, time entries, and communications.
SecurityUsing AI to track, organize, and surface insights across legal matters—from intake through closure—integrating documents, deadlines, budgets, and communications.
Tech / ModelA structured disclosure document that describes an AI model's intended uses, performance metrics, training data, and known limitations for informed evaluation.
Tech / ModelAI legal tools trained or configured to apply and reason across multiple legal jurisdictions simultaneously, enabling analysis of cross-border contracts or regulatory compliance.
Tech / ModelAI that processes multiple input types — text, images, tables, scanned PDFs — in a unified model; legal applications include scanned document review, exhibit analysis, and financial disclosure extraction.
AI-accelerated review of NDAs identifying non-standard confidentiality scope, structural issues, duration problems, and definition gaps; the most widely used AI contract review application.
Tech / ModelThe AI discipline enabling computers to interpret, analyze, and generate human language text; powers contract clause extraction, legal research, document classification, and entity recognition.
Tech / ModelRecommends or triggers the next workflow step for a matter based on current status, deadlines, and pattern recognition from similar past matters, reducing dropped balls and improving matter velocity.
Systematic monitoring of contractual commitments after signing — deliverables, payment dates, renewal windows, notice periods; AI extracts obligations and creates calendar triggers and alerts.
SecurityAI models deployed on infrastructure owned or controlled by the law firm or legal department, keeping all data and computation within the organization's own environment.
SecurityOn-premise deployment of legal AI means running the AI software and models on the law firm's or organization's own servers and infrastructure, rather than using cloud-based vendor services, keeping all data processing within the firm's controlled environment.
Legal PracticeA corporate client's written requirements governing how outside law firms must handle matters, bill for work, manage data, and use technology — including AI tool restrictions and disclosure obligations.
CapabilityAI tools that help in-house legal departments manage law firm relationships through automated invoice review, performance benchmarking, matter allocation, and rate management.
In legal AI, a playbook is a configured set of rules, preferred positions, and fallback language that guides how an AI system reviews, negotiates, or drafts contracts — encoding the legal team's standard negotiating positions for automated application.
CapabilityAutomated comparison of incoming contract drafts against a firm's approved positions and language, with systematic flagging of deviations and suggested fallback language.
Legal PracticeThe application of statistical and machine learning models to legal data — case outcomes, judge rulings, settlement patterns — to inform legal strategy and risk assessment.
CapabilityA TAR technique where the system learns from attorney-coded seed documents to predict relevance across the full document set; court acceptance depends on validation methodology.
Tech / ModelAn LLM deployed exclusively for one organization with no data sharing with other customers or the model provider for training; provides stronger confidentiality guarantees at higher infrastructure cost.
Legal PracticeA privilege log is a document produced in discovery that identifies each document withheld from production on grounds of privilege, describing the document without disclosing privileged content, enabling the opposing party to assess the validity of the privilege claim.
CapabilityPrivilege review is the process of examining documents in an e-discovery collection to identify and withhold materials protected by attorney-client privilege, work product doctrine, or other applicable privileges before production to opposing parties.
SecurityThe intentional or inadvertent disclosure of privileged communications to a third party, potentially destroying attorney-client or work product protection.
Tech / ModelPrompt engineering is the practice of designing and structuring the text instructions given to a large language model to produce more accurate, relevant, and usable outputs for specific tasks.
SecurityAdversarial instructions embedded in user input or external documents that manipulate an AI system to override its intended behavior or bypass safety constraints.
Retrieval-Augmented Generation (RAG) is an AI architecture that combines a retrieval system — which fetches relevant documents from a specified corpus — with a generative language model that produces answers grounded in those retrieved documents, rather than relying solely on the model's training data.
Legal PracticeRedaction is the process of permanently obscuring specific text or content within a document before it is produced or disclosed, to protect privileged information, confidential data, personally identifiable information, or other material that should not be visible to the recipient.
Legal PracticeMarking up a contract draft with proposed changes — deletions in strikethrough, additions underlined — during negotiation; AI tools now generate suggested redlines based on a firm's playbook.
Legal PracticeAutomated AI surveillance of legislative, regulatory, and enforcement developments affecting a client's industry, triggering alerts when material compliance obligations change.
Legal PracticeAutomated AI surveillance of regulatory updates, rulemaking, and enforcement actions relevant to a client's industry or jurisdiction to flag compliance obligations.
Legal PracticeRegulatory research is the process of identifying, analyzing, and applying the rules, statutes, agency guidance, and enforcement standards that govern a specific industry or activity — often across multiple jurisdictions with overlapping or conflicting requirements.
Legal PracticeAI-assisted review of representations and warranties in M&A and commercial contracts to identify inaccuracies, gaps, and negotiation risk before signing.
Tech / ModelDesign, deployment, and governance practices ensuring legal AI systems are safe, fair, transparent, and accountable; encompasses hallucination mitigation, bias testing, auditability, and professional responsibility alignment.
Tech / ModelAn AI architecture combining a language model with a retrieval system that fetches relevant documents at query time, grounding responses in authoritative source material to reduce hallucination.
SOC 2 (Service Organization Control 2) is an independent audit framework that evaluates a service provider's security, availability, processing integrity, confidentiality, and privacy controls — commonly cited by legal AI vendors as evidence of their data security practices.
SecurityAn independent CPA audit confirming a vendor's security controls operated effectively over 6–12 months against AICPA Trust Service Criteria.
SecurityThe 2020 CJEU ruling that invalidated the EU-US Privacy Shield and imposed conditions on Standard Contractual Clauses for transfers of EU personal data to the United States.
Tech / ModelRetrieves documents based on meaning rather than keyword matching, using embeddings and vector search; significantly improves recall in legal research compared to Boolean search.
Legal PracticeAI-assisted estimation of the likely settlement value or probability in litigation based on case characteristics, jurisdiction patterns, and historical outcomes.
Legal PracticeThe process of verifying a case's current validity using Shepard's Citations — LexisNexis's citator system — to confirm the case has not been overruled or negatively treated.
Legal PracticeEU-approved model contract clauses for transferring personal data to countries outside the EEA; required for GDPR-compliant cross-border data transfers.
Tech / ModelThe first independent large-scale accuracy benchmark for commercial legal AI tools, finding mistake rates from 17% to 88% depending on the platform tested.
Legal PracticeThe use of AI to analyze statutory text, legislative history, and regulatory guidance to identify the meaning and application of law to a specific legal question.
Two AI-assisted document review approaches in eDiscovery: TAR 1.0 uses a frozen trained model; CAL continuously updates as reviewers code documents.
Tech / ModelAn AI-assisted document review format that extracts and presents contract or regulatory data in structured table form, enabling rapid comparison across multiple documents or data points.
CapabilityA court-accepted eDiscovery methodology using machine learning to rank documents by relevance, reducing manual review volume; also called CAL or CAR.
Tech / ModelIn the context of large language models, a token is the basic unit of text the model processes — roughly a word fragment, word, or punctuation mark — used to measure both input length and output length, with practical limits imposed by the model's context window.
CapabilityAI monitoring of trademark registries and marketplace platforms to detect potentially infringing marks or unauthorized brand use, reducing manual watching costs for large portfolios.
Tech / ModelTraining data is the corpus of text and examples used to train a large language model, establishing its capabilities, knowledge, and limitations; the quality, recency, and composition of training data directly affects the model's reliability for legal tasks.
Tech / ModelThe neural network architecture underlying modern LLMs (GPT, Claude, etc.) that enables contextual understanding across long documents; has dominated legal AI since approximately 2020.
Vector search is a retrieval method that finds documents semantically similar to a query by comparing numerical vector representations (embeddings) rather than exact keyword matches, enabling natural language queries to surface conceptually relevant results.
SecurityThe risk that a law firm or legal department becomes operationally dependent on a single AI vendor's proprietary formats, models, or infrastructure, limiting future flexibility or migration options.
SecurityWhether an AI legal tool uses client-submitted content — contracts, queries, briefs — to train or improve its models, with direct implications for attorney-client confidentiality.
Legal PracticeA secure digital repository for sharing confidential deal documents in M&A transactions, enhanced by AI to automate document categorization, redaction, and Q&A.
An AI vendor commitment that customer inputs and outputs are not stored beyond the immediate processing session — the strongest available privacy assurance for sensitive legal queries.
SecurityZero retention is a data handling policy under which an AI tool vendor does not store or retain any client-submitted content after the active processing session ends, ensuring that confidential information is not persisted on the vendor's servers.
SecurityAn AI vendor policy under which user inputs and outputs are not stored after the session ends, leaving no persistent record of the interaction on vendor infrastructure.
Tech / ModelA model's ability to perform a legal task it was not explicitly trained on, relying on general language understanding; lower performance than purpose-trained models on specialized tasks.