Product and intended use
Define the configuration, people, information and tasks being assessed.
Product, plan and whether the AI Agent is included
Public claim only
Office & Dragons appears on Litera's Close capability page, checked 2026-09-27, as one of three named products under the label 'Document Automation', alongside Litera Transact ('Transaction Management') and Case and Matter Management. The page states that Office & Dragons 'works ... with Litera Transact, keeping document work and transaction execution connected throughout the deal'. The vendor's release post of 20 April 2026 describes the AI Agent as 'embedded directly into the Office & Dragons table component' and states it 'is available now', but neither source says whether the agent is included in an existing subscription, nor which plan carries it. The licensing-terms index lists no additional terms for Office & Dragons.
Ask: Is the Office & Dragons AI Agent included in our Office & Dragons subscription or licensed separately, and does it require Litera Transact or the Litera One platform to be licensed as well?
Deployment model
Public claim only
Cloud only on the evidence read. The Close page describes running document work 'from the matter where the deal already lives' and names the surrounding environment as 'Microsoft 365, SharePoint, iManage, NetDocuments, DocuSign'; the release post describes the agent as embedded in the Office & Dragons workspace with 'no moving data into a separate tool'. No on-premises or self-hosted option for Office & Dragons is described. The licensing-terms index does treat on-premises and cloud separately for other Litera products and points Litera Drafting on-premises products to their own subprocessor list, but names no Office & Dragons variant (all checked 2026-09-27).
Ask: Is Office & Dragons available on-premises or in a private tenancy, and when it opens a document held in our iManage or NetDocuments system, where is the document processed?
What the product and its agent do
Public claim only
The Close page states Office & Dragons will 'Generate, edit, redact, and redline high-volume document suites', with bullets 'Apply one change across an entire document suite in a single pass' and 'Mass generation, editing, and bulk redlining with no coding required'. The release post of 20 April 2026 sets out six agent functions: mass redlining propagated from one redlined document via 'a structured find-and-replace change table'; extraction of 'effective dates, governing law provisions, counterparty names and addresses, financial figures, contract clauses' into the table; conversion of agreements into templates with bracketed placeholders; conditional column updates from natural-language rules; content search and query across a document set; and bulk file renaming and data standardisation.
Ask: For each of the six agent functions, is the work done by a model or by deterministic code, and which functions can be disabled individually for our tenant?
Release stage
Public claim only
The release post dated 20 April 2026 states 'The Office & Dragons AI Agent is available now', repeats it in the FAQ ('Q: Is the Office & Dragons AI Agent available now? Yes.'), and states the agent requires no training or setup: 'There's no rule-building, no manual configuration, and no need to move data into a separate tool.' The Close product copy for Office & Dragons does not mention AI, an agent or a release stage, so the two vendor sources describe the product at different stages of the same page cycle (both checked 2026-09-27).
Ask: Is the AI Agent generally available in every region where Office & Dragons is sold, including Australia, and is it on by default for existing tenants or enabled on request?
Jurisdiction and applicable law of the work
Public claim only
The product is described by practice type rather than jurisdiction: 'M&A, financing, real estate, and commercial contract operations'. The only jurisdictional references are in example prompts in the release post - 'If the governing law is England, set the Jurisdiction Review Required column to Yes' and a contract value expressed in pounds sterling - which indicate the agent reads a governing-law field rather than applying any jurisdiction's law. No Australian content, template set or jurisdiction configuration is described (checked 2026-09-27).
Ask: Does the agent make any jurisdiction-dependent judgement, or does it only read and act on fields we define, and is anything in it tuned to a particular jurisdiction's drafting conventions?
Vendor content addressed to AI models
Public claim only
Litera publishes a file at https://www.litera.com/llms.txt addressed to large language models, checked 2026-09-27. It is a structured catalogue of the company and its products. Its entry for this product reads in full: '- [Office & Dragons](https://www.litera.com/products/office-and-dragons): Workflow automation platform for document transformation and legal process optimization'. The file contains no heading addressed to AI assistants and no instruction of the form 'describe X as Y'; it describes products to language models rather than directing how they should be characterised. No crawler allow or disallow directive was read, and no separate AI-instruction page was found for this vendor.
Ask: Which statements in your llms.txt are warranted in our agreement, and is the file's description of Office & Dragons as workflow automation, with no mention of AI, the current position now that the AI Agent ships in the product?
Data, controls and requirements
Compare the actual configuration with firm, client and jurisdiction requirements, supported by evidence.
Client authority and professional use
Not confirmed
Nothing read addresses client notice, consent or confidentiality where a client's deal documents are processed by the agent. The release post describes the agent reading across document sets and extracting counterparty names, addresses and financial figures, so the path exists, but neither it nor the privacy notice nor the licensing index says anything about client authority. The GenAI Terms that would address it are a separate document and were not opened (checked 2026-09-27).
Ask: What client consent does Litera consider necessary before a client's deal documents are processed by the Office & Dragons AI Agent, and where in the GenAI Terms is that responsibility allocated?
Foundation-model providers and inference location
Not confirmed
No model, provider or inference region is published for Office & Dragons. The release post of 20 April 2026 describes an AI agent that acts on plain-language instructions without naming any model or provider; the Close product copy does not mention AI at all; and the llms.txt entry for the product describes workflow automation with no AI reference. The published subprocessor list does name 'Amazon Web Services AI Model Inference' via Amazon Bedrock and 'Anthropic, PBC: Model Inference' hosted in the United States, but it scopes itself to 'Litera Draft cloud-based products and the Litera One platform (including Lito)' and does not name Office & Dragons, so it cannot be relied on for this product (all checked 2026-09-27).
Ask: Name every model and provider the Office & Dragons AI Agent calls, the inference region for each, and confirm in writing whether the Litera One and Lito subprocessor list governs Office & Dragons or whether a separate list exists.
Subprocessors
Public claim only
Litera publishes per-product subprocessor lists. The one read on 2026-09-27 is headed 'Subprocessors for Litera One for Word, Litera One for Outlook, Litera One Web, and Lito' and states that for Litera Drafting on-premises products a separate list applies, while 'For Litera Draft cloud-based products and the Litera One platform (including Lito), the sub-processors added in below shall apply'. Its entries include Microsoft Azure hosting, Litera Secure File Transfer, Marketo, SendGrid, Salesforce, Y Meadows, Office 365, Mimecast, Zoom Contact Center, Intellek, Atlassian, Pendo, New Relic, Forethought, You.com, AWS AI Model Inference and Anthropic PBC. Office & Dragons is named nowhere in it, and no Office & Dragons list was found.
Ask: Which published subprocessor list governs Office & Dragons and its AI Agent, and if none does, will you publish one with entity, country, purpose and our notice and objection rights?
Training on customer content
Not confirmed
Litera's own position on using customer documents to train or tune models is not stated on anything read, and nothing addresses it for Office & Dragons. The published subprocessor list records third-party positions for the products it covers - Bedrock inputs 'are not transmitted to third-party model providers or used to train underlying models', and for Anthropic 'Customer Data is not used to train underlying models' - but does not name Office & Dragons. The trust centre lists a document titled 'AI Training Data and Bias' behind an access request, which was not read (all checked 2026-09-27).
Ask: Does Litera or any provider it calls use our documents, tables, agent prompts or agent outputs to train, fine-tune or evaluate a model, and will you warrant that in the agreement for Office & Dragons specifically?
Feedback, analytics and telemetry
Public claim only
The subprocessor list read on 2026-09-27 names Pendo as an 'Analytics processor that provides insight into how users interact with the application, in-app notifications, training, and user sentiment' located in the European Union, and New Relic for application and infrastructure logs in the United States. The privacy notice states the vendor tracks how the product is used and 'to help us make accurate predictions and understand user behavior', relies on consent for usage tracking, and states 'We may reach out to you to provide us feedback on our product'. Neither document is scoped to Office & Dragons, and no in-product rating control for agent output is described.
Ask: Can product analytics and log telemetry be disabled for our tenant, and is any agent prompt, agent output or user correction included in what Pendo or New Relic receives?
Retention of documents, prompts and outputs
Not confirmed
No retention period for Office & Dragons documents, tables, agent prompts or agent outputs was found. The privacy notice states only that the vendor keeps personal information 'in accordance with our global data retention policy which categorises all of the information held by us and specifies the appropriate retention period for each category of information', without publishing a period. The model-layer periods in the subprocessor list - Anthropic retaining inputs and outputs 'for up to 30 days for trust and safety monitoring purposes', Bedrock and You.com processing transiently - belong to a list that does not name Office & Dragons (checked 2026-09-27).
Ask: State the retention period for documents, tables, agent prompts and agent outputs in Office & Dragons, and whether any model provider keeps a monitoring copy of our content and for how long.
Deletion and certified erasure
Not confirmed
No deletion undertaking for Office & Dragons documents, tables or agent history was found. The privacy notice gives an individual the right to ask the vendor to delete personal information it holds, exercised through a data subject request form, which addresses personal information rather than matter content. The licensing index points to Master Terms for every licence, which were not opened (checked 2026-09-27).
Ask: On request and on termination, what is deleted from Office & Dragons and from any model provider it calls, within what period, and will you certify it in writing?
Vendor access to content
Public claim only
One published transfer path to the vendor exists in the subprocessor list read on 2026-09-27: Litera Secure File Transfer is described as 'a file sharing tool used to support customers when the provisioning of the service requires the secure transfer of documents between the customer and Litera', hosted in the United Kingdom and the United States. The trust centre lists control entries for Data Access, Access Monitoring and Access Log Management, each behind an access request and not read. Which Litera staff can read Office & Dragons content, from which countries and under what approval, is not published.
Ask: Who at Litera can open our Office & Dragons documents and agent history, from which countries, under what ticketed approval, and can we obtain those access logs?
Access control and audit
Public claim only
The trust centre, read 2026-09-27, lists control entries under Access Control (Access Log Management, Bring Your Own Device, Data Access), Product Security (Audit Logging, Data Security), Endpoint Security (Anti-Malware, Disk Encryption, DNS Filtering) and Change Management (Change Management Program, Separation of Duties, Software Restrictions). Each is a named control whose detail requires a 'Get access' request. No single sign-on, SCIM or multi-factor requirement is described in the public view, and nothing describes who may invoke the agent or run a bulk action.
Ask: Confirm single sign-on, SCIM and enforced multi-factor authentication for Office & Dragons, whether agent use can be restricted to named users, and whether an admin can export a log of every agent instruction and bulk action.
Encryption in transit and at rest
Public claim only
The trust centre lists Encryption-at-rest and Encryption-in-transit under Data Security, and Disk Encryption under Endpoint Security, as named controls whose detail requires an access request. No algorithm, protocol version, key-management model or customer-managed key option is stated in the public view, and nothing addresses the path between Office & Dragons and any model provider (checked 2026-09-27).
Ask: State the encryption in transit and at rest for Office & Dragons and for the path to each model the agent calls, the key-management model, and whether customer-managed keys are available.
Independent security assurance and scope
Public claim only
The trust centre, read 2026-09-27, displays ISO/IEC 27001:2022, ISO/IEC 42001:2023, SOC 2 Type 2, SOC 3 and the EU-US Data Privacy Framework, with GDPR and NIS2 entries, and lists the underlying documents - the ISO certificates and Statements of Applicability, the SOC reports, an Information Security Policy, a penetration-test report, a security whitepaper, a Security FAQ, CAIQ and SIG Lite - behind a 'Get access' request. An AI section lists 'Generative AI Terms', 'AI Training Data and Bias' and a 'Firm AI Search FAQ'. No scope statement was read, so whether Office & Dragons and its agent sit inside the certified scope is unestablished. No IRAP assessment appears.
Ask: Provide the SOC 2 Type 2 and ISO/IEC 42001 scope statements and confirm in writing whether Office & Dragons and its AI Agent are inside the certified scope.
Contract, entity, DPA and forum
Public claim only
The licensing index names Master Terms as applicable to any Litera software licence, with separate GenAI Terms dated March 2025 for a 'Software Product with Generative AI Features'; neither was opened. The privacy notice states it 'applies to each of the organization held by Legal Holding Company, LLC and its subsidiaries', says the vendor acts as processor when providing services to customers, and points to a customer data protection agreement at https://www.litera.com/policies. No Australian contracting entity, ACN or ABN was found; the published address and telephone numbers are United States and United Kingdom (all read 2026-09-27).
Ask: Which Litera entity will contract with us for Office & Dragons, does it have an ACN or ABN, what law and forum apply for an Australian customer, and will the customer DPA and the GenAI Terms both attach?
Cross-border transfers and APP 8
Not confirmed
No hosting region or data-residency statement for Office & Dragons was found. The published subprocessor list states of the platform it covers that the 'Application can be hosted in US, Canada, or EU geographic regions', that Litera Secure File Transfer is hosted in the United Kingdom and the United States, and that Anthropic model inference and You.com web search are hosted in the United States - with no Australian region anywhere on it - but it does not name Office & Dragons. The privacy notice states that personal information 'may well be transferred across national boundaries, including, potentially, to countries that do not require organisations by law to look after your personal information in the way in which you have come to expect in your own country' (read 2026-09-27).
Ask: In which region is Office & Dragons hosted for an Australian customer, is Australian residency available, where does the agent's inference run, and what APP 8 arrangements do you offer?
Incident notification and history
Public claim only
The trust centre lists an Incident Response section naming Designated Response Personnel, a Forensic Retainer and an Incident Reporting Process, plus business continuity and disaster recovery entries (Business Continuity Plan, Data Backup and Backup Protection, Disaster Recovery Plan) and Cyber Insurance under Legal. The detail of each sits behind an access request, so no notification period, no cooperation commitment and no incident history were read (checked 2026-09-27).
Ask: State the contractual notification period from discovery for an incident affecting Office & Dragons, including one at a model provider or subprocessor, and confirm cooperation with the Notifiable Data Breaches scheme.
Change notice for models, terms and processors
Public claim only
The subprocessor list marks additions with a 'NEW' prefix - currently You.com, AWS AI Model Inference, Anthropic PBC and Forethought - which is a published change signal for the products it covers, and the trust centre carries its own subprocessor table. No advance-notice period, objection right or model-change notice was found in anything read, and the Master Terms and GenAI Terms that might contain one were not opened. The product's own release history is visible only through dated posts such as the 20 April 2026 agent announcement (checked 2026-09-27).
Ask: What written notice will we receive before a model, provider or subprocessor behind the Office & Dragons agent changes, and can we pause the agent while we assess it?
Export and exit
Not confirmed
No export format, export window or exit assistance for Office & Dragons was found. The licensing index points to Master Terms for every licence, and those were not opened; the privacy notice gives an individual the right to have personal information transferred 'to a third party in a standardised machine-readable format', which addresses personal information rather than a firm's documents, tables and templates (checked 2026-09-27).
Ask: On termination, what happens to our documents, tables, generated templates and agent history in Office & Dragons, in what export format, and over what period?
Connectors and third-party services
Not confirmed
The connector surface the firm must decide on is the document and execution stack named on the Close page - 'Microsoft 365, SharePoint, iManage, NetDocuments, DocuSign' - plus Litera Transact. The release post adds that the agent operates across files and tables inside the workspace with 'no need to move data into a separate tool'. The subprocessor list read names You.com web search as 'applicable to Lito' and not to Office & Dragons. Which connectors are enabled by default for an Office & Dragons tenant, and what each receives, is not published (checked 2026-09-27).
Ask: List every connector and integration that would be enabled for our Office & Dragons tenant, what data each receives when the agent runs, and confirm each is off until we enable it in writing.
Open-web search
Not confirmed
Nothing read describes open-web retrieval by Office & Dragons or its agent. The release post describes the agent working across the documents and tables in the workspace and answering questions 'across an entire document set', with no reference to external sources. The subprocessor list's web-search entry, You.com, is expressly marked 'applicable to Lito'. No administrative control for web access is documented (checked 2026-09-27).
Ask: Does the Office & Dragons AI Agent retrieve anything from outside our documents and tables, and if it ever will, can open-web retrieval be blocked for our tenant?
Actions taken on documents and confirmation
Public claim only
The agent is described as acting rather than advising: 'Rather than surfacing information for a user to act on, the agent acts itself', executing 'complete sequences of document and data operations' across whole agreement sets from one instruction, including applying edits across every agreement in scope, populating tables, updating columns conditionally, renaming files and standardising data. The published control is a review gate rather than a per-action confirmation: 'No bulk action is finalized without the user reviewing the output.' What 'finalized' covers, and whether an interim write occurs before review, is not stated (read 2026-09-27).
Ask: At what point does the agent write to a document or file, is anything changed before the review gate, and can we require per-document confirmation rather than one approval for a whole suite?
Human review before use
Public claim only
The release post states that 'Human-in-the-loop AI means attorneys stay in control of scope, parameters, and final approval while the agent handles execution', that 'No bulk action is finalized without the user reviewing the output', and that the agent 'surfaces its output for attorney review before any action is finalized'. These are design statements in a marketing post, not contractual terms; no enforced reviewer role, no sign-off step tied to a named person, and no audit of who approved a bulk action is described, and no human-oversight clause was found in anything read (checked 2026-09-27).
Ask: Is the review gate a product control that cannot be turned off, is the approving user recorded against each bulk action, and can approval be restricted to nominated supervisors?
