Product and intended use
Define the configuration, people, information and tasks being assessed.
Product identity and current name
Public claim only
The product is carried in the vendor's own documentation at help.relativity.com, read 2026-09-27, under the heading 'Case Strategy (aiR for Case Strategy)' with the page last modified 25 September 2026. The name changed in the vendor's September 2026 rebrand: 'aiR for Case Strategy is now Case Strategy. References to aiR for Case Strategy may still appear in this documentation, in the product, and in other sources such as marketing materials, published content, and commercial or legal agreements. Both names refer to the same product/feature.' The platform documentation lists Case Strategy as one of five aiR systems.
Ask: Which name does our current agreement use, and does the rebrand change the scope of what the Case Strategy contract licenses?
Release stage
Public claim only
Documented as generally available to customers holding the relevant contract, not as a pre-release programme: it is 'available as a secured application from the Application Library', the install steps, object permissions, billing mechanics and ARM compatibility are all published, and a separate documented feature (Assist) depends on a completed Case Strategy job. Nothing read describes a beta, early-access or advanced-access status for it.
Ask: Is Case Strategy generally available to Australian customers today on the same documentation and terms as elsewhere?
What the product does and what it produces
Public claim only
Documentation read 2026-09-27: 'Case Strategy uses Generative AI to support and potentially accelerate the building of your case strategy. With Case Strategy, you can automatically identify and extract potential facts from key documents, generate customizable document summaries to assist with review, and use collections of these facts to help draft witness summaries and deposition outlines to support preparation for interviewing individuals.' The platform page adds that it 'supports robust fact extraction, timeline development ... summarization capabilities, and work product creation' (vendor wording, not adopted here). The documented outputs are draft fact chronologies, document summaries, witness summaries and deposition outlines; Transcript Summary is a related documented job type. The vendor's own hedging is in the nouns and verbs: facts are 'potential facts', chronologies and outlines are 'draft', and the AI 'support[s]' and 'help[s] draft' rather than produces final work product.
Ask: Which of these outputs does your documentation treat as ready for a lawyer to work from, and which are described as drafts requiring rework?
Deployment, installation and where the work happens
Public claim only
Case Strategy is an application installed into a workspace rather than an always-on platform feature: 'Case Strategy is available as a secured application from the Application Library. You must have an active Case Strategy contract to use it in standard workspaces', installed from the Relativity Applications tab. It works on documents already in the workspace and writes Fact, Case Strategy Case Document and Case Strategy Prompt Criteria objects into it. Inference is separate from storage: the platform documentation states aiR systems 'use AI models and conduct AI processing using Microsoft Azure services unless otherwise specified in the terms of your Agreement with Relativity', and that such processing 'may take place outside the Microsoft Azure region where your Relativity aiR instance is hosted'. ARM is compatible with Case Strategy, so its objects can be archived, restored and moved with the workspace. Only a cloud deployment is documented.
Ask: For our instance, in which region is the fact and chronology data stored, in which region does fact extraction run, and where is deposition video hosted?
Jurisdiction and legal content
Public claim only
Case Strategy carries no legal-content library: the corpus is the matter's own documents and transcripts, and the vendor states that facts are extracted from 'the extracted text of documents'. Nothing read claims coverage of Australian legislation, case law or rules of court. Two documented design points are United States-shaped and a firm should map them to local practice: the named outputs are 'witness summaries and deposition outlines', and the vendor recommends that 'an attorney or administrator carefully review the outputs'. Nothing read addresses Australian examination or cross-examination practice, or Australian discovery and production formats.
Ask: Has Case Strategy been tested on Australian litigation material, and do the deposition-outline templates and prompts reflect Australian examination practice?
Vendor content directing how AI describes the product
Not confirmed
A deliberate check was made on 2026-09-27 and nothing of this kind was found. Checked: the two Case Strategy documentation pages, the platform AI products page, the Assist documentation page, the vendor's Artificial Intelligence Principles page, the Calder7 product security and compliance pages, the privacy policy, and a cached 40-URL map of relativity.com. No page carried an 'Instructions for AI Assistants' heading or any equivalent block directing how AI models, assistants or crawlers should describe the product, and the map returned no /llms.txt, /ai-info, /for-ai or /ai.txt. No credit was spent guessing URLs, so the absence of such a file at an unmapped path is not established. The related check for a vendor answering a comparative or superlative question about its own product was also made: nothing read does so, and the only such wording on the pages read is a link to 'best practices' documentation.
Ask: Does Relativity publish any file or embedded block that tells AI assistants or crawlers how to describe Case Strategy or the aiR platform, and if so where?
Data, controls and requirements
Compare the actual configuration with firm, client and jurisdiction requirements, supported by evidence.
Who decides what the AI reads and what it writes into the matter
Not confirmed
Case Strategy does not only read the matter, it writes into it: extracted facts become Fact objects, with scores, rationales and dates, that the rest of the case team will see and may rely on. Access is controlled by object permissions granted per group — Fact and Case Strategy Case Document with View, Edit and Add — so whoever holds those permissions can generate facts across the documents they can see. On a matter holding another party's produced documents, or material under a confidentiality regime or a court order, deciding which documents may be put through fact extraction, and who may then edit a generated fact, is a client-authority and undertakings question rather than a technical one. Nothing read describes a restriction that keeps a defined subset of documents out of a Fact Chronology job.
Ask: Can a Fact Chronology job be confined to a defined subset of a matter so that documents subject to a confidentiality undertaking or a court order are excluded, and who in our firm authorises a run?
Foundation-model provider
Public claim only
Named at platform level, not per system. The platform documentation, last modified 25 September 2026, states: 'Relativity's aiR systems currently leverage a combination of generative AI LLMs including GPT-4o, GPT-4.1 mini, GPT-5 mini, GPT-5.1, and GPT-5.2 through Microsoft Azure services. The specific LLM or combination of LLMs used by an aiR system depends on the aiR system objective, and the capability and availability of an LLM that meets Relativity's minimum product requirements.' The Generative AI considerations page confirms the dependence — 'Generative AI technology, specifically large language models (LLMs) are central to the capabilities of Case Strategy' — without naming a model. No customer model choice is described anywhere read, so none is recorded.
Ask: Which named model performs fact extraction, which performs document summarisation and which drafts witness summaries and deposition outlines for our instance today, and what written notice will we get before any of them changes?
Subprocessors
Not confirmed
No subprocessor register was found. The cached 40-URL map of relativity.com shows no subprocessor page, and neither Calder7 page nor the privacy policy carries a register. Microsoft Azure is the one named infrastructure and model-hosting provider, from the statement that aiR systems 'use AI models and conduct AI processing using Microsoft Azure services'. Nothing read names the provider for the up-to-150 GB of deposition video hosting. The privacy policy addresses third parties only in general terms and expressly excludes the data a firm cares about: 'Data uploaded by our customers into the software is not covered by this policy and you should contact your software administrator if you have questions about how Relativity processes such data.'
Ask: Provide the subprocessor register that applies to Case Strategy, naming each entity, its country and its function, including the model-hosting entity and the provider of deposition video hosting, with our notice and objection rights.
Training on customer content
Public claim only
The vendor publishes one statement on the point, in the platform documentation last modified 25 September 2026: 'Customer data processed using aiR systems is not used to train or improve third party models or services.' Its scope is third-party models and services. Nothing read states whether documents, extracted facts, prompt criteria, generated chronologies, witness summaries or deposition outlines are used to train, fine-tune or evaluate Relativity's own models. The vendor's Artificial Intelligence Principles page, read the same day, refers to models Relativity itself trains and states 'We utilize data minimization in the models we train, using personal data only to the extent necessary to fit the purpose of the model and help users efficiently organize information', without saying whose data those models are trained on. The privacy policy lists 'improving our software, websites and related services' among its purposes but excludes customer-uploaded data.
Ask: Will you warrant in our agreement that no document, extracted fact, prompt criterion, chronology, witness summary or deposition outline from our workspace is used to train, fine-tune or evaluate any model, including Relativity's own models, and not only third-party models?
Feedback and quality monitoring
Not confirmed
Nothing read describes a feedback or rating control inside Case Strategy, or what happens to a correction a lawyer makes to a generated fact, chronology, witness summary or deposition outline. Because facts are stored as editable objects with View, Edit and Add permissions, a reviewer's correction is captured somewhere in the workspace, but no page read states whether corrections are used to improve the system, whether they stay inside the instance, or whether prompt criteria are reused beyond the matter they were written for. The vendor's documentation site carries a helpfulness form, which is about the documentation rather than the product.
Ask: When a lawyer corrects or deletes a generated fact or rewrites a deposition outline, does that content leave our instance, and is it used to improve the product or any model?
Retention of prompts, facts, generated work product and video
Not confirmed
No retention period is published for prompt criteria, extracted facts, document summaries, fact chronologies, witness summaries, deposition outlines or the up-to-150 GB of deposition video. What is established is where they live and that they move with the workspace: facts and case documents are workspace objects, and 'ARM is compatible with Case Strategy', so they can be archived, restored and moved with it. The only published retention sentence found anywhere is in the privacy policy, last updated 22 June 2022 — 'We only keep your Personal Data for as long as it is necessary for the legitimate business purpose for which it was collected' — and that policy states in terms that customer-uploaded data is outside its scope. Nothing read addresses prompt or completion data held at the model endpoint.
Ask: State the retention period for prompt criteria, extracted facts, generated chronologies, witness summaries, deposition outlines and deposition video, and for any prompt or completion data held outside the workspace.
Deletion and certified erasure
Not confirmed
Nothing read establishes what is deleted when a matter closes or a subscription ends, over what period, or whether erasure is certified. Facts and case documents are workspace objects that a user with Edit permission can presumably change, but no deletion behaviour is documented for them, for prompt criteria, for generated work product or for deposition video, and no page read states whether deleting a workspace deletes prompt or completion data held at the model endpoint. The privacy policy offers individuals a deletion right over their personal data but excludes customer-uploaded data from its scope.
Ask: On matter closure and on exit, what is deleted, over what period, does it include prompt criteria, generated facts, chronologies, memos and deposition video wherever stored, and will you certify the deletion in writing?
Vendor human access to content
Public claim only
The Calder7 product security page publishes a control aimed at this: 'Customer Lockbox — Prevent administrators – including those from Relativity – from seeing your data unless you say so', with Security Center able to 'Perform real-time audits of Lockbox to enable and revoke access to individuals as needed'. That is a security page rather than a contractual document; no standing approval process, access-log format or retention of access records is described; and nothing read states whether Lockbox covers generated facts, chronologies, witness summaries, deposition outlines and deposition video as well as the underlying documents.
Ask: Is Customer Lockbox available on our plan, does it cover generated facts, chronologies, memos and deposition video as well as matter documents, and what audit record do we receive of any Relativity access?
Access control, permissions and audit
Public claim only
Case Strategy is permissioned through named workspace objects rather than a single switch: 'Case Strategy Case Document - View, Edit, Add', 'Case Strategy Prompt Criteria - View, Edit, Add' and 'Fact - View, Edit, Add', together with the Case Strategy and Case Home tabs and Document Preview. It is also a 'secured application' installed from the Application Library, so installation is itself an administrative act. Platform-level controls on the security page are 'OpenID Connect and SAML2 protocols, two-factor authentication, and single sign-on', with 'discrete permissions to give you total control'. Nothing read states what is written to the audit trail when a Fact Chronology, Transcript Summary, witness summary or deposition outline job is run, who ran it, over which documents, or with which prompt criteria.
Ask: What does the audit trail record when a Case Strategy job runs — who ran it, over which documents, with which prompt criteria and which model — and can we export that record?
Encryption and customer-managed keys
Public claim only
The Calder7 product security page states 'By default, Relativity encrypts data at rest and in transit' and offers 'Customer-managed keys' which 'add an extra layer of protection, giving you total control over your environment's encryption key'. No algorithm, protocol version or key-management detail is published; the statement sits on a security page rather than in a contract; and nothing read states whether customer-managed keys extend to generated facts and work product, to the deposition video store, or to data in transit to the model endpoint.
Ask: Name the encryption algorithms and key management, and confirm in writing that customer-managed keys cover generated facts, chronologies, memos and deposition video, and that traffic to the model endpoint is encrypted.
Independent security assurance
Public claim only
The Calder7 compliance page publishes a certification list, read 2026-09-27: 'ISO/IEC 27001:2022 certification, ISO/IEC 27018:2019 certification, FedRAMP Class C Certified, HIPAA Compliance, IRAP Assessed at Protected, SOC 2 Type II Report, SOC 3 Report, Cloud Security Alliance's (CSA) CAIQ', with certificates available on request to securityquestions@relativity.com. The IRAP entry is the Australian-relevant one. Four limits: no certificate or report was seen, so none is verified here; the page states no scope, date, assessor or exceptions; the FedRAMP authorisation is attributed elsewhere on the same site to Relativity aiR Government rather than the commercial platform; and nothing read states that Case Strategy, its generative-AI processing or the deposition video store sit inside the assessed scope of any of them.
Ask: Provide the current SOC 2 Type II report, the ISO/IEC 27001:2022 certificate and the IRAP assessment letter with their scope statements, and state in writing whether Case Strategy and its model inference are inside those scopes.
Contract, DPA, entity and liability
Not confirmed
No contractual document was read and none is published on the pages read. There is no subscription agreement, data processing addendum, AI schedule or subprocessor annex on the vendor's public site as mapped, and the Terms of Use link in the documentation footer governs use of the website rather than the platform. The only entity name published on any page read is the site copyright line, '© 2026 Relativity ODA LLC', which is not a statement about who would contract with an Australian customer; the published address on every page read is 231 South LaSalle Street, 20th Floor, Chicago, IL 60604, with a United States telephone number. No governing law, forum or liability position was found. The documentation confirms that the agreement is where the commercial and geographic facts actually live: an 'active Case Strategy contract' is required, video beyond 150 GB is charged 'at your contractual Review rate', and model and processing geography apply 'unless otherwise specified in the terms of your Agreement with Relativity'.
Ask: Provide the subscription agreement and data processing addendum that would cover Case Strategy, name the entity that will contract with us in Australia with its ACN or ABN, and state the governing law, forum and liability cap.
Hosting region and cross-border processing
Public claim only
The vendor publishes a table of aiR instance regions against inference locations, on the platform documentation page last modified 25 September 2026. Australia appears in both columns: instance region Australia, aiR system inference location Australia. The same section sets the general rule and its limits: aiR systems 'use AI models and conduct AI processing using Microsoft Azure services unless otherwise specified in the terms of your Agreement with Relativity'; inference 'may take place outside the Microsoft Azure region where your Relativity aiR instance is hosted'; 'Your data remains stored in the region you selected, only the aiR system AI inferencing may occur elsewhere'; and the table 'reflects aiR System processing locations at the time of publication. These locations may change based on AI model availability and the evolving capabilities of our infrastructure providers.' So an Australian inference location is published, but as documentation the vendor reserves the right to change rather than as a commitment. Nothing read states where the up-to-150 GB of deposition video is hosted. The privacy policy's own transfer statement expressly does not cover customer-uploaded data.
Ask: Will our agreement warrant that fact extraction, summarisation and work-product generation all run in the Australian region for our instance, where is deposition video hosted, and what notice will we get if the published processing-location table changes?
Incident notification and history
Not confirmed
No incident-notification obligation, period or history was found on any page read on 2026-09-27. The Calder7 pages describe detection and prevention — around-the-clock monitoring, external penetration tests 'multiple ... a year', a bug bounty, 'AI-driven threat detection and response' and Security Center alerts to the customer for unusual events — but say nothing about notifying a customer of a breach, the period from discovery, incidents at a subprocessor or model provider, or the Australian Notifiable Data Breaches scheme. The privacy policy contains no breach-notification clause.
Ask: State the contractual notification period from discovery for our instance, confirm it covers an incident at Microsoft Azure or at the model endpoint, and set out how you will support our assessment under the Notifiable Data Breaches scheme.
Change to the product, the model and the output
Not confirmed
The vendor publishes that aiR systems change without a fixed notice period and that outputs change with them: 'Relativity regularly updates aiR systems by modifying Relativity inputs/instructions, AI technology, and other features ... As a result, customers may notice changes in the product experience or differences in outputs after an update, even when the same prompt criteria or workflow is used.' Communication is 'through product release notes, knowledge base articles, and/or official announcements'. Two further facts compound this for Case Strategy. Output is already non-reproducible within a single version: 'Witness summaries and deposition outlines will have differences if run multiple times, even if the input facts list and the prompt criteria are the same.' And throughput may be varied at will — 'Relativity may, in its sole discretion, rate-limit, queue, prioritize, or restrict usage of these features', applied 'independently of any job size limits your administrators configure'. The product was also renamed in September 2026.
Ask: What written notice will we get before the model or the vendor's own instructions behind fact extraction change, and how do we evidence to a court which version produced a chronology already served?
Export and exit
Public claim only
One route is documented and it is the platform's, not the product's: 'ARM is compatible with Case Strategy', so facts, case documents and prompt criteria can be archived, restored and moved with the workspace. Nothing read describes an export of a fact chronology, witness summary or deposition outline to a document format, an export of the Fact object store to a structured file, or the retrieval of up to 150 GB of deposition video on exit. Generated chronologies and outlines may become part of a firm's own file and may be served or relied on, so the absence of a documented export format bears on more than convenience.
Ask: On exit or matter transfer, in what format can we take fact chronologies, witness summaries, deposition outlines, prompt criteria and deposition video, over what window, and at what cost?
Connectors and whether external tools can reach the work product
Not confirmed
At platform level the vendor publishes Relativity MCP, 'an interface that allows AI assistants access to Relativity aiR capabilities through Model Context Protocol, an open standard for connecting AI tools to external data and services through MCP clients', and also Gavel for drafting in Microsoft Word and the Relativity App Hub for third-party applications. Nothing read states whether any of them can read or create Fact objects, chronologies, witness summaries or deposition outlines, who may enable them, or whether such access is audited.
Ask: Which connectors, App Hub applications or MCP clients can read or write Fact objects and generated work product in a workspace where Case Strategy is installed, and who can enable them?
Open-web retrieval
Public claim only
Everything documented points inward, and the vendor is specific about how narrow the input is: facts are extracted from 'the extracted text of documents', 'A single document is sent to the AI for fact extraction at a time', and to build witness summaries and deposition outlines the system 'uses only the main fields for the facts submitted including description, issues, date, etc. and the document summaries for the documents linked to those facts'. No page read states that Case Strategy makes no request to any source outside the workspace, and the model is a third-party large language model reached through Microsoft Azure services, so the absence of a statement is not a statement.
Ask: Does Case Strategy make any request to a web index, a model-provider tool or any source outside our workspace at any step of fact extraction, summarisation or drafting, and can that be disabled and evidenced?
What the product writes into the matter without a human step
Not confirmed
Fact extraction writes into the matter: it 'automatically identif[ies] and extract[s] potential facts from key documents' and stores them as Fact objects carrying a description, a date, issues, a score and a rationale, which the rest of the case team can see. The vendor's own limits mean those stored values can be wrong in ways that look right — a fact may be given 'a low fact score' when it is key, a date 'might differ from what is actually in the document', and a fact 'not present in the document' may be created. Duplicates are expected where documents are near-identical. Nothing read describes a confirmation step before facts are written, a marker distinguishing an AI-generated fact from one a lawyer entered, or a cost or size gate before a job runs beyond the 50,000-document limit on integrated pricing, where 'the job fails to run'.
Ask: Are generated facts marked in the workspace as AI-generated and distinguishable from facts a lawyer entered, and is any approval required before a Fact Chronology job writes into a matter?
Human review before reliance
Public claim only
The vendor states the expectation in terms, on the Generative AI considerations page last modified 25 September 2026: 'Broadly speaking, we expect the users of Case Strategy to always have a human overseeing the work produced by the AI', and 'We strongly recommend that an attorney or administrator carefully review the outputs to verify their accuracy as part of your overall workflow.' The main documentation directs users to that page before use: 'Before using AI, we recommend learning about how it works so that you can leverage it effectively.' These are recommendations, not enforced controls: nothing read describes a setting that requires review before a generated fact becomes visible to the team, records that a reviewer checked a fact against its source document, or marks a generated chronology or deposition outline as AI-generated when it leaves the workspace.
Ask: Can we require and evidence that a named lawyer verified each generated fact against its source document before a chronology is used, and does the system mark generated work product as AI-generated?
