Epiq AI vs Lighthouse: how they compare in 2026

The product is the AI. Epiq AI is a line of applications whose work is done by the models. Epiq AI for Review turns review protocols into prompts and classifies documents for responsiveness, issues, PII and privilege; Epiq AI for Privilege classifies privilege and drafts logs; Epiq Assist answers questions in natural language over a dataset. All run on the Epiq Laer platform, which orchestrates models, agents and reviewers. Without the models there is no product. Epiq's wider eDiscovery platform and services, and resold third party AI such as Leah, are separate products.

Lighthouse now leads with its AI applications, built on an eDiscovery platform and service that runs without them. LighthouseIQ's four applications answer questions in plain language from a matter's documents, draft chronologies and case strategy, code documents for responsiveness and issues, and identify privilege and draft privilege logs, using predictive and generative AI. Spectra runs integrated machine learning and AI services inside its self serve workflows. Underneath are processing, hosting, Relativity review and production, and a large forensics and managed review business, all of which operate without the models.

Epiq AI for Privilege gives reasoning and citations at line level for each call, and Epiq Assist answers with links to supporting documents. Epiq's validation article sets out how results are tested: recall, precision, elusion and richness estimated from random samples with confidence intervals, a control set of about 500 documents kept out of training, and recall and precision tradeoffs for each tag from k fold cross validation. Epiq publishes no accuracy figure on a test set. Its outcome claims, such as reducing lawyer review by up to 90 percent with TAR or CAL, review costs down up to 65 percent and 500,000 documents an hour, are efficiency figures rather than accuracy measures.

IQ Answers draws its answers only from the documents ingested for the matter and handles one question at a time to avoid context drift. Each answer passes through interpretation, layered retrieval combining predictive relevance models, semantic similarity and retrieval that keeps context intact, then synthesis, and each response carries highlighted source text and exemplar documents so every statement can be traced. Lighthouse makes several accuracy claims. IQ Priv's privilege descriptions were 12 percent more accurate than contract attorneys in blind comparisons, and IQ Priv is said to be 6.6 times more accurate than privilege keyword terms. IQ Review reached 76 percent recall at 79 percent precision on 3 million documents and 85 percent recall and precision in another matter. No test set or method accompanies any of them.

Human validation is built in and described. Classifications carry reasoning and citations for reviewers to check, each issue tag comes with guidance for choosing a cutoff, and Epiq recommends sample validation of the responsive and nonresponsive sets before production, with review services pairing the AI with reviewers. Prompts can be edited and classifications rerun at no charge, privilege output gets quality control after export, and Epiq AI Accelerate promises audit trails, workflow approvals and traceability for each AI action. The Software Terms state that generative outputs are preliminary and that the client is responsible for verifying them. What happens when validation fails, and any route for reversing a wrong output once produced, are not published.

Under Lighthouse's defensibility guidance, prompt engineers set AI instructions with counsel, attorneys review results on sample populations to confirm accuracy, and iterative feedback refines the prompts. IQ Review moves work from reviewing each document to targeted validation and quality control, with counsel validating sample outputs before applying a workflow broadly and Lighthouse experts handling setup and monitoring. IQ Priv produces log lines that users approve or edit. Lighthouse states that AI does not replace human reviewers. No confidence threshold, stopping point or route for reversing a call once applied is published, and IQ Review is run by Lighthouse consultants rather than configured by the customer.

Epiq says 130 clients, including global corporations and law firms, adopted Epiq AI within its first year. An Am Law Top 25 firm reports avoiding more than $10 million in review fees across 12 matters and 21 million documents, with use disclosed to and accepted by the SEC, DOJ, FDIC and Federal Reserve, and other anonymized studies report a health insurer's 50 percent savings and a media company saving $1.5 million on a regulatory deadline. No customer is named and no method is given. Epiq AI won the 2026 Legalweek Leaders in Tech Law award for best use of AI in eDiscovery and litigation.

Named customers report results. Microsoft's EJ Bastien says IQ Priv returned just over 24,000 privileged documents against about 190,000 expected, Cleary Gottlieb met a DOJ Second Request deadline while saving more than $2 million, and Kirkland & Ellis, Akin, Baker Botts and Redgrave give testimonials. Unnamed case studies report a top 10 United States firm avoiding more than $420,000 of review on a DOJ civil investigative demand, an Am Law 50 firm reviewing 126,000 documents in 72 hours, and a healthcare provider saving $650,000 across 11 million documents. Lighthouse also says Spectra customers saved $2,951,280.50 in July 2026, calculated as native data kept out of Relativity hosting multiplied by the standard hosting charge per GB. No method accompanies the case study figures.

Confidentiality is written into published agreements, with training limited but not excluded and privilege not named. The Epiq Software Terms (revised February 2026) treat every Discovery Document, including material under protective order, as confidential without the usual exceptions. They limit use of client data to providing the software, require notice before compelled disclosure, and require return or destruction at termination at the client's choice. Section 15, however, lets Epiq use client data to fine tune its generative AI tools for that client's environment. The Epiq AI FAQ says it does not train public models with client data, and the Responsible AI page says the Azure OpenAI service behind Epiq's summarization keeps prompts and outputs from OpenAI and other customers. Neither agreement addresses privilege or work product by name.

The Master Spectra Subscription Agreement (2 April 2026) bars Lighthouse from using client data or metadata to train or develop any AI service, its own or a third party's. The same bar covers the inputs and outputs of its AI services. It bars sending client data to any publicly available or shared AI model without written authorization. Privilege has its own section confirming that privileged communications and work product remain protected and that work is done at counsel's direction. The agreement also commits to notice before compelled disclosure and requires deletion of client data within 30 days of termination. Spectra documents matter creation and role based access so users see only the matters they need. The contract governs Spectra; for LighthouseIQ, Lighthouse states the same position on training and retention outside any published agreement.

A clear contractual position on advice versus tooling is published, without jurisdiction limits. The Epiq Software Terms say that nothing Epiq does is the practice of law, a legal opinion or legal advice, and that Epiq has told the client it is not representing it. The Epiq Global Terms of Service say Epiq is not a law firm. The Software Terms add that generative outputs are preliminary, untested and unverified, and that the client bears responsibility for verifying them. The product is sold to law firms and legal departments. No jurisdictional limits are stated.

The Spectra agreement states that Lighthouse is not engaged in the practice of law, provides no legal advice and creates no attorney client relationship. Its services represent Lighthouse's interpretation as a legal technology provider, and work on privileged material is done at the direction of the client or its counsel. IQ Answers is described as not a substitute for attorney judgment, and Lighthouse's defensibility guidance says attorneys govern and validate AI outputs and confirm they reflect their professional judgment. Optional matter conflict checks are offered. The product is sold to law firms and legal departments, and no jurisdictional limits are stated.

A published governance structure names a body, without testing results. Epiq's Responsible AI page says a Responsible AI Committee oversees the development and deployment of its AI, that bias assessments are carried out to identify and reduce bias, and that practices are reviewed regularly. It adds employee training and a route for staff to raise concerns. Epiq AI Labs, led by the founders of Laer AI with Cornell's Center for Data Science for Enterprise and Society as a founding member, researches the technology. No assessment results, release testing or findings about uneven performance are published.

How the generative workflows are governed is set out by Lighthouse's Senior Director of AI and Analytics in a published question and answer on defensibility. Prompt engineers work with counsel, attorneys validate outputs on sample sets with qualitative and quantitative metrics, and prompts are documented and versioned. It also says predictive workflows follow validated practice for technology assisted review. Lighthouse says LighthouseIQ was tested on live matters over three years. No release testing results, named accountability for the AI applications, or findings on uneven performance across languages or data types are published.

The Software Terms commit to breach notification without undue delay with assistance to the client, confidentiality obligations on staff and agents, and return or destruction of personal data at termination, while section 16(d) lets Epiq delegate work to affiliates or subcontractors, including access to and processing of client data, without client consent. The Global Terms of Service allow 30 days to export data after termination, after which Epiq may delete it. The security page describes identity and access management on Zero Trust principles, 24/7 managed network monitoring and regular penetration tests. No subprocessor list, fixed notification window or retention period for AI inputs and outputs is published, and the privacy statement, last modified 21 September 2026, says nothing about AI.

The Information Security Addendum to the Spectra agreement commits to notify a security incident within 48 hours with a description, likely consequences, mitigation and a contact. It also commits to multi factor access controls, central logging protected against tampering, and annual third party penetration tests. The Data Processing Addendum commits to destroy or return personal data within 60 days of termination, and the main agreement requires client data to be destroyed within 30 days. The Processor Compliance Statement adds encryption at rest and in transit and a rolling 30 day backup retention. The published subprocessor list names every subprocessor, affiliate and review staffing firm with its role and processing location, with a dated change log and notice and objection rights for new ones. IQ Answers keeps a history of users, questions and answers for auditability, without a stated period.

A real published position includes indemnities and caps, with an express disclaimer of AI output. The Software Terms warrant that the software performs substantially as documented, with correction or termination and a refund pro rata as the remedy. They indemnify the client for infringement, material breach, gross negligence, security breaches the client did not contribute to and legal noncompliance, and cap liability at six months of fees for the software concerned. The Global Terms of Service cap liability at 12 months of fees, with exceptions for confidentiality breach and gross negligence. Both disclaim any warranty of accuracy for AI outputs, which the client must verify.

Spectra's subscription agreement warrants that the services conform to their documentation, with refund, repair or replacement as the remedy. It indemnifies the client against third party claims that the Lighthouse IP infringes, and caps each party's liability at fees paid in the prior 12 months. The cap excludes fraud, willful misconduct, gross negligence, death or personal injury and third party claims. The client may terminate with a refund if availability or response levels are missed for three months in a row. Nothing addresses an incorrect AI answer, privilege call or log line.

Epiq AI Accelerators add translation, image analysis, OCR and agentic automation inside Relativity, including Relativity aiR and Relativity Server, and Epiq AI for Review and Epiq Assist are built into Epiq Discover. Epiq MCP servers for Epiq AI, Discover, Narrate and Access connect those products to AI assistants using existing permissions and controls, without moving data out. Access runs through Epiq Access single sign on in the Epiq Service Cloud, with Microsoft, AWS, Relativity, Litera and WorkFusion named as partners. No implementer documentation is published.

Spectra runs matter setup, processing and culling, then promotes data into Relativity Server, RelativityOne or Relativity aiR for review and carries it through production and export. Workflow Manager and Native File Manager control what reaches active hosting. IQ Review delivers its results directly in Relativity, and Lighthouse offers Relativity aiR as a managed, client controlled or self service option, citing 359 Relativity certifications. The platform also covers Microsoft 365 and Purview and workflows for Google Vault linked files. Technical documentation is provided to customers through Spectra rather than published.

Deployment is described in real terms, without the regions. Epiq says it runs Tier 3+ data centers in ten locations worldwide, offers private cloud capabilities in the Epiq Service Cloud and sites specific to regions across 14 countries, and that the Epiq Laer platform uses local models to control throughput and cost. The data center locations, where AI processing happens relative to storage, and residency options for Epiq AI are not published.

Lighthouse hosts on Microsoft Azure in the United States and UK and colocation data centers in Seattle, Phoenix, London, Frankfurt and Dornach, and RelativityOne runs in the region the customer specifies. Its subprocessor list gives a processing location for every subprocessor and affiliate, including support outside office hours from India and review staffing in named countries case by case. Cloud workflows are controlled by region, keeping storage and computing within defined boundaries, and the LighthouseIQ workflows run in a private Azure environment. What changes between Spectra and managed service deployments is not set out.

Epiq's security page shows ISO, Cyber Essentials, TISAX and UKAS marks, says it is aligned with HIPAA, NIST and FISMA, and says third parties assess its program, with Bitsight ratings giving clients visibility of its posture. Which ISO standard is held, and its scope, are not stated, the Privacy Shield mark invalidated in 2020 still appears, and no reports or trust center are offered.

A-LIGN Assurance certifies Lighthouse's ISO 27001 information security management system, and its security page lists ISO 27001, SOC 2 Type II and HIPAA aligned controls, with UK operations holding Cyber Essentials. The Spectra agreement commits to an annual assessment against ISO 27001, the HIPAA Security Rule and SOC 2 Type 2, with a summary provided on written request. Lighthouse offers the ISO certificate and a security overview through a Trust Center, and no report date or scope statement is published.

Two providers are named. The Responsible AI page says Epiq's summarization uses the Microsoft Azure OpenAI Service and sets out that service's data terms, and a 21 July 2026 release says Epiq AI Accelerate embeds Anthropic's Claude, working through MCP with Epiq's own reasoning layer. The Epiq AI FAQ says the Epiq Laer platform orchestrates multiple models and agents and uses local models strategically, and Epiq names AWS and DeepJudge as partners. Which models power each Epiq AI application, their versions, the access route and how model changes are notified are not published.

LighthouseIQ runs on IQ Fabric (processing, cognition and orchestration), which combines predictive models trained on attorney coding with generative models, and IQ Review uses different models and workflows for different review tasks. Lighthouse's security page says LighthouseIQ's generative workflows use foundational large language models delivered through Azure infrastructure, and its subprocessor list names Microsoft Azure for hosting and Daypack Research for AI workflow management tools, with notice before new subprocessors. No model, model vendor or version is named.

The pricing unit and structure are published without a figure. Epiq AI is sold per matter, with unlimited prompts and protocol changes and reclassification at no additional fee, and is also offered to managed services clients at a predictable monthly cost. The Software Terms describe prepaid and monthly service fees with subscriptions of 12 months that renew automatically. No price is published.

A rate card for services is published as part of the Master Spectra Subscription Agreement, titled 2023 rates and applying unless an order form says otherwise: online hosting at $12 per GB a month, Nearline at $7.20, Relativity users at $89 a month, forensic collection at $295 an hour and OCR at $0.02 a page, among some forty line items. Spectra's agreement sets out a free Tier 1 support level and a paid Tier 2 expert level, fees set in an order form, automatic annual renewal and rate increases on 90 days' notice. Spectra comes with a savings calculator and cheaper Nearline storage. No price is published for the Spectra subscription itself or for LighthouseIQ.

Epiq AI is sold to corporate legal departments and law firms for investigations, risk assessments, HSR filings and second requests, merger reviews, post execution contract review, PII and PHI identification, early case assessment, incoming production review, deposition preparation and litigation review, with versions for antitrust, compliance and plaintiff class actions. The Canopy acquisition of September 2026 extends it into data breach review. Epiq says it classifies documents in any language, and its Service Cloud serves more than 2,600 clients and 48,000 users in 17 countries. Where the AI is weaker, by language, data type or matter type, is not stated.

Corporate legal departments and law firms in financial services, technology, pharmaceutical, healthcare, telecommunications, manufacturing and construction are Lighthouse's clients. Its specialties include antitrust and HSR second requests, multidistrict litigation, DSARs and Microsoft 365, and Spectra covers investigations, DSARs and compliance reviews, and routine litigation. IQ Answers works on any data with document text, and LighthouseIQ claims no limits on document types or volume; nothing states where the AI applications perform less well.