Relativity
Relativity is an enterprise legal data platform. Its cloud product, RelativityOne, combines e-discovery, investigations and contract review with the aiR generative AI suite, delivering document-level citations and reasoning that legal teams can validate and defend.
What is Relativity?
Relativity is a Chicago-based legal technology company whose platform, RelativityOne, is built on Microsoft Azure and sold as software as a service. An installed variant, Relativity Server, remains available in 2024, 2025 and 2026 releases, and a separate RelativityOne Government edition runs on the Azure Government Cloud with FedRAMP Class C certification. The company positions the whole thing as legal data intelligence: organise complex data, find what matters inside it, then act.
The generative AI layer is the aiR suite, developed with Microsoft on Azure OpenAI, and the company reports more than 200 customers using it. Five products sit under that name. aiR for Review predicts relevance, tags issues, identifies key documents and, through Custom Analyses, extracts and classifies data against a matter's own requirements. aiR for Privilege finds, logs and protects privileged material. aiR for Case Strategy extracts facts and builds narrative. aiR for Data Breach Response locates impacted personal information and drafts notification lists. aiR Assist adds conversational insight inside the platform. Alongside them, Relativity claiR, currently in Advanced Access, lets practitioners question a matter in everyday language and receive cited answers grounded in the record.
Data reaches the platform through Collect, which pulls from Microsoft 365, Google Workspace, Slack, Box, iManage, X1 and Cellebrite, and also from ChatGPT Enterprise, Claude Enterprise and Gemini. A Model Context Protocol connector works with Claude.ai, Claude Cowork and Claude Code for the administrative work around a matter.
The consistent argument is defensibility. Every aiR output carries document-level citations, a rationale and its reasoning; subject matter experts write the prompt criteria and validate results; the company states plainly that fully autonomous agents without human oversight do not belong in legal technology. Customer results published on the site include a million documents reviewed in eighteen days, 96% recall, and cost reductions of 85%. Relativity claims 800-plus staff in research and development, 100-plus dedicated to AI, and over $150 million invested in R&D each year.
What it does
- Predict document relevance across millions of files, with a citation and a written rationale behind every call
- Find, log and protect privileged material before it leaves the door
- Collect data straight from enterprise cloud platforms, including conversations held in Claude, ChatGPT and Gemini
- Surface the personal information caught in a data breach and build the resulting notification lists
- Extract facts and assemble a case timeline for deposition and trial preparation
- Question an entire matter in plain language and generate memos and summaries grounded in the evidence
- Provision workspaces, assign access and pull usage reports conversationally through the Model Context Protocol connector
When to use Relativity / When not to
A quick filter to help you decide if Relativity is the right fit.
When to use Relativity
- Litigation teams in law firms that review hundreds of thousands to millions of documents per matter and must defend how those decisions were made
- In-house legal departments running internal investigations, regulatory second requests or antitrust responses under fixed deadlines
- Incident response and privacy teams that need to find impacted personal information fast and build notification lists after a data breach
- Federal, state and local agencies needing FedRAMP-certified tooling and FOIA request handling from intake through release
- Legal service providers and forensic consultancies that run discovery projects for multiple clients and need data residency in a specific country
When not to use Relativity
- Solo practitioners and small firms with no budget visibility: nothing is priced publicly and every path ends at a sales quote
- Anyone hoping to try before buying, since neither a free plan nor a free trial is advertised anywhere on the site
- Lawyers looking for case law research or document drafting: the platform reasons over the evidence in a matter, not over jurisprudence
- Teams working mainly in non-English corpora, as the vendor states aiR has been primarily tested on English-language documents
- Organisations wanting an Android app or self-service onboarding, as mobile is iOS-only and deployment assumes administration and training
How to use Relativity
A typical end-to-end flow, from setup to results.
- Request a demo or contact the sales team, since there is no self-service sign-up and no published price list
- Agree a licence: pay as you go, or commit to one or three years for a lower rate, scoped to your data volumes, workflows and region
- Access your RelativityOne instance through a browser, authenticating via Relativity's centralised login
- Have an administrator open a workspace from a template and set up groups and permissions, or describe the case to the MCP connector and let it do the provisioning
- Collect the evidence: pick the cloud sources, add your custodians, or run a non-custodial SharePoint collection
- Process the collection with its metadata directly into the review workspace, culling duplicates and irrelevant material
- Write the Prompt Criteria for aiR in plain language, ideally in the same language as the documents being analysed
- Run the analysis, then read the citation, rationale and reasoning behind each document decision and validate the output as a subject matter expert
- Measure recall, precision and elusion to demonstrate the process was reliable and consistent
- Continue on the move if needed: the iOS app lets you view and code documents on an iPad or iPhone, offline included, syncing back to the web
Pros & Cons
Pros
- Generative AI is built into the platform rather than bolted on, and aiR for Review and aiR for Privilege are included in standard RelativityOne pricing
- Outputs are explainable by design: document-level citations, rationales and reasoning, with humans validating each step
- An unusually dense compliance base: ISO/IEC 27001:2022, ISO/IEC 27018:2019, SOC 2 Type II, SOC 3, FedRAMP Class C, HIPAA, IRAP and CSA CAIQ
- Data residency across roughly fifteen geographies, with the vendor stating data never leaves the country of the instance
- Customer data analysed by aiR is not retained by Relativity or by Microsoft, and abuse monitoring has been switched off
- A mature ecosystem: App Hub, 80-plus developer partners, certifications, an active community and an academic programme
- Published AI principles and a named in-house security team monitoring around the clock
Cons
- No public pricing whatsoever: budgeting is impossible without a sales conversation
- No free plan and no free trial are advertised, so the platform cannot be evaluated hands-on independently
- No Data Processing Agreement is published and no subprocessor list is disclosed, despite third-party providers being acknowledged
- No Article 27 EU representative is named, and the privacy policy still carries a June 2022 date
- aiR has been primarily tested on English-language documents, which shifts validation effort onto non-English matters
- claiR remains in Advanced Access, with its inclusion in RelativityOne pricing described only as expected
- No published support email address and no Android app: support runs through the Community, mobile through iOS only
Pricing & Plans
No pricing is published. Every product card on the pricing page routes to a Request Pricing form, and quotes are calibrated on data volumes, workflows and region. Licensing is flexible: customers may pay as they go or commit to one or three years for a saving. Neither a permanent free plan nor a free trial is advertised. What is stated clearly is what the subscription contains: aiR for Review, aiR for Privilege and aiR for Case Strategy are presented as included in RelativityOne at no additional cost, and claiR is expected to join integrated pricing when it becomes generally available. The iOS application is free to download but requires a RelativityOne account.
- the main SaaS platform
- priced on quotation
- with aiR for Review
- aiR for Privilege and aiR for Case Strategy included
- FedRAMP Class C edition for agencies
- running on the Azure Government Cloud
- installed deployment
- available as Server 2024
- Server 2025 and Server 2026 (early access)
- Relativity Contracts
- Relativity FOIA
- Relativity Legal Hold and aiR for Data Breach Response
- the pricing page lists products and a request button
- nothing more
Data, GDPR & hosting
A consolidated view of how Relativity handles your data.
GDPR overview
The commitments are concrete rather than decorative. The full privacy policy names Regulation (EU) 2016/679 directly and sets out a dual role: processor for the data customers upload, controller for everything else it collects. Standard Contractual Clauses are implemented for transfers between group companies, vendors and customers. Relativity ODA LLC and its subsidiary Relativity Global LLC are certified under the EU-U.S. Data Privacy Framework, the UK Extension and the Swiss-U.S. Data Privacy Framework, with JAMS as dispute resolution provider and the FTC as enforcement authority. Rights of access, copy, correction, deletion, objection, consent withdrawal and restriction are exercised at privacy.inquiries@relativity.com or through a dedicated online form, with a pointer to local supervisory authorities. Two gaps: no Article 27 EU representative is named, and the privacy policy carries a June 2022 date.
Who owns the data?
The material a customer uploads is treated as Customer Data and stays the customer's. Relativity positions itself as a data processor for it, states that it does not independently access or otherwise handle it, and explicitly places it outside the scope of its own privacy policy. The content and the categories of that data are determined by the customer alone, and questions about how it is handled are routed back to the customer's own software administrator rather than to Relativity. The same split is repeated in the Data Privacy Framework policy, where Relativity describes itself as solely a processor and directs individuals to the relevant controller. RelativityOne is presented throughout as the customer's system of record.
Reuse rights
Two regimes coexist. For Customer Data, Relativity acts on instruction only: data analysed by aiR is never retained by Relativity or by Microsoft, internal access is described as extremely limited and granted just in time, and the company says it has opted out of the abuse and harmful content monitoring Microsoft offers so that no unauthorised party can reach raw inputs or outputs. Data is stated to remain within the Relativity boundary, and aiR runs on Azure OpenAI enterprise models rather than public ones. For the personal data Relativity collects itself through its websites, marketing and support channels, it acts as a controller: that data supports the service, marketing subject to your preferences, targeted advertising through cookies, and sharing with service providers, resellers and advertising partners. Its fourth AI principle commits to data minimisation in the models it trains.
Data retention & training
Hosting summary
RelativityOne runs on Microsoft Azure. The documentation lists hosting geographies covering Asia (East and Southeast), Australia (East), Brazil (South), Canada (Central), Europe (North and West), Germany (West Central), India (Central), Japan (East), Korea (Central), South Africa (North), Switzerland (North), the UAE (North), the UK (South) and the United States (East and Central). The residency commitment is explicit: data for an instance does not leave the borders of its country, though disaster recovery backups may sit in a different region within that same country. RelativityOne Government is separate, running on the Azure Government Cloud with all customer data processed in the continental United States. Customer data is described as staying within the Relativity boundary. Azure is credited with meeting more than 90 international and industry-specific compliance standards, and Relativity's vendor risk team states it reviews Azure's security and privacy posture at least annually.
Things to keep in mind
Risks and trade-offs to weigh before adopting Relativity.
- Cost is unknowable in advance: with no published pricing, the real figure depends on volumes, workflows and region, and only a sales conversation reveals it
- The promise that analysed data is never retained is a commercial statement on a product page, not a contractual clause; get it written into the agreement
- No published DPA and no subprocessor list means the due diligence a privacy team would normally do from public pages has to be done through procurement instead
- European personal data is processed without a named Article 27 representative on the site, which is a gap worth raising before signature
- aiR was primarily tested on English documents, so teams working in other languages carry more of the validation burden themselves
- Automation bias is the human risk: outputs arrive with citations and confident rationales, and a tired reviewer may stop genuinely checking them, which is precisely what the audit trail is meant to prevent
- Adopting a platform this large reshapes how a team works; without training and administration, the tooling outruns the people using it
Setup & Integrations
Technical difficulty
Low for the end user, moderate for IT. RelativityOne is SaaS, so nothing is installed: access is through a browser with centralised authentication. IT work is limited but real, mainly firewall openings on port 443 to region-specific domains, plus workspace, group and permission administration. Using aiR requires no data science background; the useful skill is writing clear, structured prompt criteria. Training, documentation, certifications and more than 80 developer partners are available. Organisations that cannot move to the cloud can deploy Relativity Server instead.
Deployment
Apps stores
Integrations
Supported languages
Behind Relativity
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What exactly does Relativity sell?
What can the aiR products actually do?
How much does it cost?
Is there a free trial or a free plan?
Is my data used to train the models?
Where is the data hosted?
Which certifications does Relativity hold?
Is there an API?
Is there a mobile app?
Does the AI replace the lawyer?
Should you pick Relativity?
Relativity is an enterprise platform, not a tool you sign up for on a Tuesday afternoon. That framing explains both its strengths and its frictions. The generative AI is part of the foundation rather than a separate purchase, and the argument the company makes for it is unusually disciplined: every aiR decision arrives with a citation, a rationale and its reasoning, experts write the instructions and validate the results, and the vendor states outright that fully autonomous agents have no place in legal work. For a profession that has to justify its process to an opponent or a regulator, that emphasis on defensibility matters more than raw speed.
The compliance and hosting picture is similarly serious. ISO 27001 and 27018, SOC 2 Type II, SOC 3, FedRAMP Class C, HIPAA, IRAP and CSA CAIQ is a long list, and data residency across roughly fifteen geographies with a commitment that data does not leave the country of the instance is rare at this level. The statement that customer data analysed by aiR is never retained by Relativity or Microsoft is a strong one, and worth writing into a contract.
The frictions are real too. Pricing is entirely opaque, and there is no trial to fall back on. No Data Processing Agreement is published, no subprocessor list is disclosed, no Article 27 EU representative is named, and the privacy policy has carried a June 2022 date for some time. claiR, the most interesting thing on the roadmap, is still in Advanced Access. And aiR's testing has focused on English documents.
The honest verdict: if your document volumes and your obligation to defend the process justify a platform of this weight, Relativity is a serious candidate. If they do not, the entry cost is the wrong shape entirely.
- Choosing a selection results in a full page refresh.
- Opens in a new window.