Modern cloud technology with no installation requirements (any modern browser will do). All communication takes place through real-time secured communication (Websockets). No page reloading is ever necessary.
Clause9 offers one integrated environment where clauses can be drafted, templates can be constructed and questionnaires can be built & tested. No more back-and-forth between your browser and MS Word.
Documents and questionnaires are interactively recalculated — typically within a few milliseconds — so any change will immediately result in visible changes on the screen.
The software’s end-user interface is completely translated in each of the 28 supported languages, so end-users can be welcomed in their local language.
To maximise their reuse, clauses will auto-adapt their numbering and styling to the level they are positioned at. Accordingly, the same clause can be used both at the highest level, or as a sub(sub)(sub)(...)clause. Fighting with MS Word’s numbering system is finally over!
If desired, sole paragraphs (e.g., clause 14.1) can be set to hide their numbering. The clause number will re-appear automatically re-appear once sibling clauses would get inserted.
TOCs can be inserted with a mouse click, with a configurable position (start of document, after title page, end of document) and a configurable number of levels.
Any clause can have an optional title that can be easily toggled on/off. Together with the automatic numbering, this allows the same clause to be inserted both at a top-level position (typically accompanied by its title) or at a sublevel.
Clauses can include the contents of other clauses, allowing authors to isolate common parts of different clauses into small “snippets” of text that can be centrally updated. For example, when different IP clauses would refer to the same Copyright Act of a certain date, the full name and reference of that Act could be stored as a snippet that is inserted in multiple clauses. If the Act ever needs to be updated, it should be updated only once.
Verbs, adjectives and pronouns can be associated with defined terms, and will automatically update when those terms would get changed, so that — depending on the language — the gender (male/female/neutral), number (singular/plural), defined or undefined article and grammatical case (nominative/genitive/accusative/...) will be automatically adapted. For most languages, Clause9 also hosts a built-in conjugation dictionary.
For each supported language, grammatical knowledge is built-in — from the way numbers, months, names of days and currencies are translated, to whether defined/undefined articles are present, which bullet suffixes (and/or) are to be used.
Clauses can be stacked together into clause hierarchies of any level deep, that can be centrally updated and inserted as a whole. Clause hierarchies can contain other clause hierarchies, any level of deep. This, for example, allows clause authors to build predefined combinations of boiler plate clauses, or even entire chapters and schedules of standard clauses that can inserted with one click of a mouse.
Clause contents can be automatically machine-translated with the click of a button. All the translations of a clause are always kept together and can be simultaneously updated.
Any clause can have an optional title and alternative representation (e.g., plain language for use in an executive summary) that can be easily toggled on/off.
Clauses can be made subject to one or more conditions, any level deep, to define under which circumstances it (and its descendant clauses) should be hidden. For example, in a share purchase agreement, it could be specified that a paragraph should only be shown when a condition precedent is met; while within that paragraph, a certain sentence should only be shown when a second condition precedent is met; and within that sentence, a certain word should be swapped for another word when a third condition precedent is / is not met.
Any (sub)clause can be set to repeat a predefined or dynamically calculated number of times, for example to create bullet lists with party names or legal requirements.
Bullets and enumerations (item lists) can automatically include language-aware suffixes, such as “and” / “or” / “and/or” in English. When the enumeration gets changed — e.g., because the condition for a certain item is no longer met — the suffix will automatically reposition.
Over 200 functions are available to deal with special requirements in legal texts — from mathematical calculations (e.g., sum, maximum number, average), date and duration calculations, text manipulations (e.g., uppercase), grammatical features (e.g., inserting the right article, taking into account the gender, number and case), clause intelligence (e.g., checking whether a definition exists for a certain term), special conditions (e.g., showing part of a clause only to certain user profiles), text processing (e.g., sorting item lists), etc.
Different versions of the same clause can be stored, to account for changes in internal playbooks or external legislation. While the latest version of a clause will be inserted by default, end-users can easily swap that latest version for some prior version.
By default, when any part of a clause (or definition) gets updated, all templates that make use of that clause will be automatically updated to reflect that change. This way, templates will not have to be manually opened to update clauses.
Clauses and definitions can be stored in folders or sub-folders, any level deep, allowing for flexibility in the organisation of files. For example, the corporate clause library could contain a folder warranties, with two subfolders purchaser warranties and seller warranties, each containing further subfolders relating to different types of warranties, and so on.
Folders can store links (shortcuts) to other folders, to allow users to create legal connections between taxonomies. For example, a folder with specific confidentiality clauses for employment law can point to a folder with generic confidentiality clauses.
All legal contents can be translated into any of the 28 supported languages — from the body of a clause to its title and file name, from references to legal doctrine, to memos, questions in Q&As, predefined answers, and so on.
To ensure consistency across documents and offer flexibility in renaming legal terminology, Clause9 allows to centrally define any number of legal terms. Users can search which clauses or documents use legal terms, and can swap a legal term for any other legal term with the click of a button. Any required grammatical changes to accommodate the change in gender or singular/plural, will be automatically performed.
Users can search for clauses with (a combination of) different criteria, such as filename, clause contents, location in the taxonomy, or user-defined legal attributes.
When searching for content, the software will automatically search for related terms, so that a clause that refers to “suppliers” will also get found when searching for “service providers”.
Using its deep linguistic knowledge, Clause9 will automatically search for linguistic variations of words. For example, a clause that mentions “entities” will also get found when searching for “entity”.
Datafields (“variables”) can host any of the following data types: text, true/false, number, fractional number, date, duration, currency; as well as lists (“arrays”) for any of those data types.
For each datafields, authors can predefine an unlimited number of default values. When values represent internal codes, they can be associated with friendly labels for display towards end-users.
Any number of datafields can be created. Datafields are centrally stored to facilitate reuse across documents and (subject to access conditions) across authors/departments.
The use of datafields in clauses, documents and questionnaires can be interactively queried.
Conditions can be combined any level deep. For example, the warranty clause can be set to only show if the contract value if beyond 500 EUR; while its first paragraph can be set to only show if the jurisdiction is set to France or Germany; while the first sentence of that paragraph can be set to only show if the product being sold is either older than 3 years, or older than 5 years but sold under special warranty; while the three initial words in that sentence only get inserted if consumer protection legislation applies.
To optimise the management of complex conditions, calculations for any data type can be bundled as dynamically configurable datafields that can be centrally stored and updated (“data expressions”).
Tables can be inserted in clauses, with most of the controls offered by MS Word, such as: user-defined backgrounds and borders; merged rows and cells; user-defined row width and table width.
Both rows and columns can be dynamically hidden when certain conditions are not met. For example, in a pricing table, rows with optional services that were not yet effectively ordered, can be hidden to create a more compact table. Similarly, when no discounts are granted, the column with the discounted value can be hidden.
Similar to rows, columns can be dynamically hidden in a table.
Table rows can be dynamically repeated to create lists, such as a lists of all shareholders with their respective share amounts and contact details.
Go beyond simple tags, by dynamically defining any number of “legal attributes” (number, amount of stars, date, true/false or tags) that can then be assigned to clauses and documents. Such legal attributes can facilitate searches and help users select the right clause for the right context. For example, an employment department could create a specific rating for “employer friendly” versus “employee friendly”.
Associate internal comments with clauses, to advise colleagues on the origin or optimal use of the clause (e.g., “don’t use this clause in the aviation sector”). Comments can be translated into any supported language.
Clauses can refer to relevant case law and legal doctrine, to guide legal professionals in their use of a clause.
Clause authors can attach legal checklists to clauses, so that relevant optional subclauses are just a click away. When clicked, the software will direct the user to a list of relevant subclauses to insert.
Clause authors can store richly formatted memos, and associate these memos with clauses, documents, questions or predefined answers. Memos can be translated into any supported language, and can optionally be exported as footnotes.
Store any number of hyperlinks to external or internal legal knowledge bases.
Clause queries can be centrally stored by each user — e.g., to search for all IP clauses that are written in French and are maximally protective for the licensor.
Users can interactively compose legal documents by searching, stacking and configuring clauses as intelligent building blocks.
Clauses can be composed in any order up to nine levels. Numbering, bullets and cross-references will automatically adapt.
Any set of documents can be combined together into a “binder” of documents. Binders can be either exported as a single .DOCX/.PDF file, or as separate subfiles in a .ZIP file.
Through conditions with any level of desired complexity, subdocuments can be automatically shown or hidden under specific circumstances. For example, depending on the product being sold, different pricing schedules and warranty annexes could be shown.
Subdocuments can be set to repeat any number of desired times. For example, in a share transfer agreement, the consent letter that need to be signed by each of the current shareholders, can be set to repeat, so that it is automatically included in multiple copies, adapted to each of those shareholders.
At any point in a template, other DOCX files can be inserted — either individual paragraphs, or as standalone pages (together with their original styling, header and footer). Where relevant, such insertions can be subjected to any number of desired conditions. Furthermore, placeholders in those DOCX files can be replaced by any dynamically selected content.
Documents can include PDF pages that are prepended (e.g., a corporate branded title page) or appended to the main document (e.g., certain product annexes), optionally subjected to conditions to only show in certain circumstances.
Any change to a document — clause move/insertion/removal, layout or numbering changes — will be applied in real-time, no matter how complex or long the document. Changes are typically applied in less than 0.2 seconds.
Clause9 encourages authors to separate legal content from layout, so that the styling becomes separately configurable, fostering reuse of legal content. For example, an inhouse team drafts documents could assign different layouts to documents of different subsidiaries, while the legal contents would be identical.
Styles can be centrally defined on the level of the company/firm, department and user. With a single click, user can then apply any of the styles available to them — from the character’s font and colour, to numbering, paragraph spacing, bullet style, header/footer style, etc.
To balance the drive towards uniform layouts, while granting styling flexibility to specific departments or users, all style settings can be configured at any of the available levels (firm/company, department, user, document, clause, paragraph). Style settings “cascade” and get merged with overriding styles at deeper levels.
To ensure uniformity across the firm/company, individual departments and/or users can be disallowed to change central styling or apply certain types of styling.
Centrally define whether defined terms should be shown with initial caps, all capitals or with no capitalisation.
Finetune the layout of automatically compiled definition lists, e.g. to configure whether terms are highlighted in bolds, how terms are quoted (straight / curly / single / double, and whether definitions are shown as bullets, numbered paragraphs or table rows.
Configure how enumeration suffixes should get styled, e.g. whether to use a common or semicolon; whether to put “and” / “or” after each bullet, or instead only after the penultimate bullet; and whether to include a period after the final bullet.
The numbering style and surrounding symbols of inline enumeration prefixes is fully configurable. An enumeration such as (1) alpha; (2) beta; (3) gamma can therefore also be styled as [i] alpha, [ii] beta, and [3] gamma.
Inline enumerations can be centrally converted into bullets, either everywhere, or as from a certain threshold (e.g., only when the enumeration contains more than 10 words in total).
Optionally include the title of a cross-reference’s target clause — e.g., “... as defined in article 4.5 (Intellectual Property)”.
Automatically include clarification expressions when desired, e.g. “... as set forth in article 1 up to and including 3” when referring to clause ranges.
The styling of placeholder text can be centrally defined, including setting such as the background color and the symbols — any combination of square brackets and/or ● blobs.
Centrally define whether to print numbers as mere digits (“30”), letters (“thirty”), or letters and digits (“thirty (30)”). Also define, on a per-language basis, how to print large numbers and fractional parts (“1.234,56” or “1,234.56”).
Centrally define how to print currencies: as symbols (€) or words (EUR), before or after the number.
Centrally define, on a per-language basis, the quotes styling: curly, straight or French-style guillemets? Lowered or raised?
A Q&A is an optional and independent layer on top of a document. Authors can therefore create multiple Q&As for the same document — e.g., a simplified version for business users and an extensively detailed version for lawyers, or completely different versions per jurisdiction.
All questions can be translated into any of the supported languages; such translations can optionally be prepared by the integrated machine translation engine. Fully translated Q&As for example allow central service entities to serve multiple jurisdictions, even when the staff in such entities do not master the output language.
Unlike most other editors, in Clause9 end-users are not required to sequentially answer a Q&A. Except when specifically disabled by the questionnaire author, end-users can provide answers in any order they like, reducing “questionnaire fatigue”.
Questions can be set to compact mode, can be set to indent, have their text color changed. Groups of questions can be indented and/or initially minimised, to avoid distracting users with details. In addition, questions can be optionally (also) inserted directly into the legal document — which can be useful to allow end-users to more directly manipulate documents, further reducing questionnaire fatigue.
Any question can be set to mandatory, and administrators can prohibit exporting documents until all mandatory questions are answered. Text-based answers can be validated using advanced rules, to for example ensure that VAT numbers and company numbers are inputted in a certain way.
An extensive range of conditions can be associated with question groups, questions and predefined conditions: visibility of other questions or groups; current language; user or customer profile.
Legal logic can be stored where it is most appropriate —at the level of an individual clause, individual document and/or specific Q&A. For example, when a clause should never be shown in a certain jurisdiction, it is probably most appropriate to store a condition at the level of that clause, so that clause will never show up inadvertently in any document or Q&A. Conversely, for other clauses, it may depend on the specific document and/or Q&A whether it should be shown or hidden.
Conditions can be nested any level deep, in any combination (all/any/no condition in a subsection is satisified). For advanced scenarios, a dedicated expression language is available for calculations within conditions.
Frequently used questions can be centrally stored, to facilitate reuse across questionnaires and lower maintenance. Similarly, frequently used conditions can be centrally stored.
Any questionnaire can be “simulated” at any moment, to interactively test what it will look like for end-users. Interactive inspectors and consistency checks are available to analyse the behaviour of truly complex Q&As.
Q&As can include user-configurable disclaimers (optionally accompanied by user-configurable dismiss buttons), to be shown towards users when the Q&A is first opened.
Authors can dynamically disable and/or replace clauses based on a user’s answers, or offer knowledgeable users the possibility to disable/select alternative clauses themselves. For example, a Q&A author may decide that (only) senior business users are allowed to replace a government court by ICC arbitration, if the value of the deal remains below a certain threshold.
Authors can dynamically change the layout, depending on the user’s profile and/or answers. This allows to, for example, automatically apply the corporate entity of a subsidiary when that subsidiary is chosen as the contracting party.
Q&A authors can dynamically swap legal terminology, depending on an end-user’s profile and answers. For example, instead of generically referring to “the customer”, the document can refer to the actual name of the customer, as indicated by the end-user.
For those situations where small lists of items are not enough, administrators can create user-configurable mini-databases, hosting data of any type. Examples include lists of lawyers to include in engagement letters; lists with entity names and VAT numbers; lists of function titles and signatory names; etc.
Any question and any group of questions can be accompanied by either short hints (“Don’t include data before 2012”), or elaborate documentation of any length (e.g., legal background information for business users). Q&A authors can dynamically configure whether to export such documentation as footnotes.
Whether and how Q&As are exported, is fully configurable. For example, authors can define that business users should only be able to export to PDF; or that export is disabled when the value exceeds a certain threshold; or that only some fields can be changed in the resulting DOCX file when a junior lawyer would complete the Q&A.
To facilitate cooperation, administrators can define shared folders, accessible to multiple users, either on a read-only basis, or with both read and write rights.
To facilitate cooperation, administrators can define shared folders, accessible to multiple users, either on a read-only basis, or with both read and write rights.
Users can highlight clauses, mark clauses for deletion, and add comments to any clause.
Both empty and partially completed questionnaires can be made available to external users, who can then get access with a simple URL — no username/password necessary. Multiple questionnaires can be bundled, access rights can be configured (e.g., no access beyond a certain date), and export options can be finetuned (e.g., only form filling; only PDF output; both PDF and DOCX output).
Define any number of custom portals, and apply them to either on a per-user or per group basis. This allows to, for example, welcome business users for one product branch with a different set of recommended templates than business users for another branch.
A portal can be subdivided into one or more tiles of over 15 different types: photos, videos, text boxes, recent files, Q&A lists, iframes, and so on. Tiles can be reused across portals.
Optionally host the software on your own domain or subdomain. Insert your own logo, and/or apply custom CSS files to approximate your own branding.
We offer an exhaustive manual, hands-on tutorials, and over 50 short tutorial videos explaining specific features of the software.
Contextual help facilities can be interactively accessed from within the product.
Users can submit a support request from within the product, including the option to include screenshots or even a full memory snapshot of the browser tab.
Documents can be exported to PDF, DOCX, Outlook attachment and the computer’s clipboard (for direct pasting in other software). Administrators can optionally those export facilities for a user in general (e.g., to allow junior sales representatives to only export to uneditable PDF), or subject the export of a Q&A to certain conditions (e.g., to allow export DOCX only if the value of the deal is below a certain threshold).
Clause9 builds DOCX from the ground up, starting from a “clean slate”, so you don’t have to deal with semi-corrupted DOCX templates whose numbering and layout behaves erratically. The resulting DOCX files are built in a technically sound manner, with correct styling and cross-references.
When clauses are available in multiple languages, documents can be exported with multiple languages lined up in parallel columns. For clauses where multiple columns make no sense (e.g., signature blocks) can nevertheless be set to span all columns.
With the click of a button, documents can be exported with all changes vis-à-vis the original template (or any other version of the document) visible. Users can even export a ZIP-file containing both a “clean” version and a version with changes.
DOCX files can be optionally protected with a password, to avoid any change, allow only form-filling, or enforce track changes.
Any document can be exported to an Excel-file, containing columns with all the variable datafields. Upon upload, each row of the Excel-file will become a separate PDF/DOCX file, fully customised for the inserted data. Typical use cases include bonus letters, corporate restructuring letters and GDPR data breach letters.
Under no circumstances do we want you to feel like you are locked in with Clause9 due to the unique intelligence you added to your clauses. At any point in time, you can create a complete back-up of all clauses and documents in your library, both with Clause9 intelligence and without it.
Files can be stored in multiple libraries — a personal library, a department library, a cross-department library and a firm-wide library.
Every folder and every file can be assigned different bundles of access rights, allowing for ultimate flexibility — e.g., allowing certain templates to be editable by department X, but read-only for department Y, with the exception of specific templates that should only be accessible by a limited number of senior lawyers of department X.
Over 200 buttons and software features can be disabled on a per-user basis — e.g., to avoid that junior lawyers would be able to tweak the layout of documents, business users would be able to export to DOCX, external users would be able to mark clauses for deletion, etc.
Administrators can define any number of groups of users, and users can be a member of any number of groups. In addition, groups can optionally have their own clause libraries and be subject to specific access rights. This allows to accurately model the nuanced “political” reality present in many firms and companies, where specific departments/users/seniority levels should (not) have (write) access to certain content.
With the possibility to enable/disable them, or change the preferred display order, on a per-user basis.
Clause9 is designed and developed by former data protection lawyers — so you can rest assured that GDPR compliance is a priority, applying security and data minimisation principles everywhere.
As you would expect from an automated legal document composer, Clause9 is "deterministic" (unlike large language models, which usually generate variated content). However, Clause9 makes smart use of artificial intelligence and LLMs in several locations, e.g. to name new clauses, find clauses and assist users to understand the grammatical analysis of a sentence.
All clauses can be exported to a format that can be ingested by large language models. In particular, you can export content into the prompt - answer structure required for finetuning by large language models from OpenAI and Microsoft (GPT3, ChatGPT and GPT4), as well asCohere). Full documents can be exported into plaintext and/or an entire JSON breakdown structure that combines text with metadata, for easy ingestion by other databases.
Clause9 and ClauseBuddy are ideal companions to LLMs. They allow you to structure your legal content using multiple types of metadata, and offer advanced search capabilities, including semantic search. You can store your legal content now, and be prepared for the future, when plugins to LLMs become a necessity.
Clause9 and ClauseBuddy can instruct LLMs to output paragraphs in a structure format (with numbering, bullets, placeholders, and so on). Such output is ideal to combine with your own structured clauses.
Clause9 is the most advanced layout engine for legal documents on the market. You can use its capabilities to compose contracts, memos and legal briefs that combine the output from LLMs with your own legal content.
Read our in-depth article about finetuning large language models.
Right from within iManage's interface, you get access to all the templates created in Clause9 and ClauseBuddy. In addition, with the click of a button, you can send interesting documents (or even entire folder hierarchies) to ClauseBuddy's Truffle Hunt.
Clause9 includes an open API, accompanied by a Swagger sandbox for experimentation, that can be used for server-to-server user management and document generation. API functions currently include customised PDF/DOCX generation, creating temporary user logins, and querying the contents of dynamic templates. More functions can be added at request.
For server-to-server communication, industry-standard REST & OData is offered, with either HTTP basic or (preferably) OAuth 2.0. For example, for HR departments, integrations include Workday, SAP SuccessFactors, SalesForce and O365.
We offer a GraphQL implementation, together with an interactive GraphiQL console, for advanced integrations. The GraphQL implementation allows, for example:
retrieving clauses through file ID, filename, attributes or legal content
retrieving specific versions of a particular clause
rendering clauses as structured JSON, or (when combined with datafields) as a Base64 DOCX
stacking specific (sub)clauses together into a PDF or DOCX file, with customised styling and legal terms
decomposing clauses as rich structured content, up to the level of individual words
SSO is available through Azure, for both Clause9 and ClauseBuddy. An easy step-by-step guide is provided for configuring an Azure SAML connection.
By way of “light” integration, particularly to avoid security questions, Clause9 offers various types of drag & drop integrations that are easy to implement, yet quite flexible:
find data in text extracted from MS Word files, text files and text-based PDF files
machine-recognition (OCR) of text in screenshots and scanned PDF files
extract data from individual cells or ranges of cells in Excel files
extract data from a JSON file using JSONPath
extract data from web pages using XPath
To cope with the most complex situations in clause composition and/or conditions, Clause9 hosts the world’s most powerful programming language (Clojure).
