Installation and configuration
This page outlines how to install the aiR for Privilege application, set security permissions, and configure general settings that apply to all aiR for Privilege projects within a workspace.
Instance level installation
The aiR for Privilege application will automatically be installed within your RelativityOne instance. This process takes 1-5 business days from the completion of aiR for Privilege contracting.
Workspace level installation
To install the aiR for Privilege application to a workspace follow the steps outlined in Installing applications.
Security permissions
The aiR for Privilege application relies on standard workspace permissions.
We suggest creating two user groups with the following permissions to complete an aiR for Privilege project:
| User group | Object security | Tab visibility | Other settings |
|---|---|---|---|
| aiR for Privilege Lead
A Lead creates aiR for Privilege projects, manages settings, defines input items, runs pipeline steps, monitors annotator progress, QC’s annotator decisions, and populates final results. |
|
|
Mass Actions
Admin Operations
|
|
aiR for Privilege Annotator An annotator confirms AI predictions as part of each pipeline step. |
|
|
Mass Actions
|
General settings in a workspace
General settings apply to all aiR for Privilege projects within the workspace and must be populated prior to starting any aiR for Privilege project.
Note: You should only update the fields in the General Settings section between projects, as changes won’t apply to projects that have already completed the Prepare Project pipeline step.
To update general settings:
- Hover over the aiR for Privilege icon on the navigation bar.
- Hover over Setup (Priv).
- Select General Settings (Priv).
- Complete the General Settings form.
Categories and fields for each section are listed below. - Click Save.
There are three sections in the General Settings section:
Client Brain
Knowledge from past projects is stored in a Client Brain. Each client, as defined by the Client object linked to a workspace, will have its own segregated Client Brain. Privilege projects can only use or update the Client Brain associated with the Client object linked to the workspace they exist in. The Client Brain can kick start a project and create consistency across projects.
Note: Consider whether you currently align the Client Objects to a Law Firm or a Corporation to understand the implications when using the Client Brain.
Client Brain fields
| Field | Field type | Required field | Description | Notes |
|---|---|---|---|---|
|
Use Client Brain |
Yes/No | Y | This setting determines whether you want to use knowledge stored in the Client Brain associated with the Client object linked to this workspace, and update the Client Brain with new learnings from any project you complete in this workspace. |
It is highly recommended that you use the Client Brain as it will reduce the annotation time for your project and create consistency across projects. Consider how you currently align the Client Object to understand the implications before electing to use the Client Brain. If this is your first project for a client, you should select Yes to make sure a Client Brain is created for the client and knowledge is retained. |
General field mappings
These settings specify the document fields used in a project for analysis.
Caution: It is important that these fields are both mapped correctly, and that documents appropriately include values for these mapped fields. Incorrectly mapping or missing document values may cause poor results or failures. Double check these mappings and the documents before starting any project. It is suggested to create a document view that includes your mapped fields so you can quality check (QC) the data completeness and quality before starting a project.
| Field | Field type | Required field | Description | Notes |
|---|---|---|---|---|
|
Document ID (Control Number) |
Fixed-length text | Y | Used to identify the document. | The control number. Should always be unique, and should always exist on a document. |
|
Parent Document ID |
Fixed-length text | Y | Used to determine if a document is an attachment and what document is it an attachment of. |
Should contain the control number of another document. Must be populated for attachments. |
|
Group ID (Family) |
Fixed-length text | Y | Used to determine all documents that are associated to one another by way of parent or attachment. | Populated for all documents, with a unique value per group. |
|
Sent Date |
Date | Y | Used to understand when an email document was sent. |
Should include both date and time normalized to UTC time. Must be populated on every email document. Should include both date and time normalized to UTC time. |
|
Sent Time |
Fixed-length text | N | Used to determine the time an email document was sent if the time was not included in the sent date setting. | Should be in the following format in UTC time: {04:46:02 PM} |
|
Record Type |
Single choice | Y | Used to determine whether a document is an email or not. |
A value of Email or email is classified as an email, while all other values are considered a non-email. Must be populated on every document. |
|
File Name |
Fixed-length text | Y | Used to determine the name of the file. | |
|
File Extension |
Fixed-length text | Y | Used to determine the type of file to drive file specific analysis. | |
|
Author |
Fixed-length text | Y | Used to determine the individual who created the document. | Should be populated on non-email documents with the individual who created the file. |
|
Custodian (single value) |
Single object | Y | Used to determine the individual who this document was collected from. |
Must be a single custodian value rather than a multiple custodian deduplicated value. Must be populated on every document. |
|
Email Subject |
Long text | Y | Used to determine the subject of an email. | |
|
Email From |
Fixed-length text | Y | Used to determine the email address that sent the email. | Must be populated on every email document and include only a single email address. |
| Email To | Long text | Y | Used to determine the email addresses that the email was sent to. | Can include multiple semi-colon delimited email addresses. |
|
Email CC |
Long text | Y | Used to determine the email addresses that are carbon copied (cc) on the email. | Can include multiple semi-colon delimited email addresses. |
|
Email BCC |
Long text | Y | Used to determine the email addresses that are blind carbon copied (bcc) on the email. | Can include multiple semi-colon delimited email addresses. |
|
MD5 Hash |
Fixed-length text | Y | Used to determine the unique MD5 hash value for each document. | Must be populated for all documents. |
|
Extracted Text |
Long text | Y | Used to determine the Extracted Text field where the document text exists. | Most of the time this will be the Extracted Text field but could be an OCR Text field or Translated Text field. |
|
Extracted Text Size in KB |
Decimal | Y |
Used to determine what documents to exclude from text analysis because they are greater than our 170KB limit. See Considerations for more details. |
If you do not have this field populated, you can use the Set long text field size mass operation to populate this information. Must be populated for all documents. |
Validating general field mappings
To prevent downstream issues, we recommend manually validating general field mapping selections. This is very important as there are times when the validation we have implemented doesn’t catch a document, metadata, or extracted text. This can cause issues with our pipeline steps.
To validate general field mappings:
- Navigate to the General Settings (Priv) tab.
- Scroll to the Document Field Mappings section.
- Take note of the fields mapped for each setting.
- Navigate to the project in question.
- Take note of the saved search used for the project.
- Navigate to the Document tab.
- Create a new saved search.
- Add the search condition of Saved Search=[saved search linked to project].
- Add each field that was mapped in settings as Fields on the new saved search.
- Save and Search.
- Perform the following manual validation. If any fail, that field may be missing data or incorrectly mapped.
Validate that each field has at least one document populated with a value on the field
Validate that your extracted text field has real data.
Validate that you have a Record Type value of email on documents.
Validate that email addresses are semicolon delimited in the email specific fields.
Validate the Time on your Sent Date field to see if they are all 12:00AM, which is the default if time is provided.
Validate that all documents have a value under Record Type.
Validate that all documents have a value under Extracted Text in KB.
Known items in a workspace
A known item is information that you already know about the matter or about how to identify privilege. This information will be used to start a project and create consistency across projects, similarly to the Brain. There are two types of known items:
-
Known Attorneys
-
Known Law Firms
These known items apply to all projects within the workspace and you must populate them before starting any project.
Adding known items is completely optional and purely improves consistency and time to results.
Note: You should only update known items between projects. Changes will not apply to projects that have already completed the Prepare Project pipeline step. Similar to changes in general settings.
Creating known items
Below are instructions on how to create known items manually. You can also create known items through a bulk import using Import/Export.
Note: When performing a bulk import using the Import/Export application, use the Import Structured (Processed) Data and then the RDO setting and select the Object Type of the known item you want to import. Always confirm that you are not creating new choices when importing to single choice fields. New choices will not be recognized by aiR for Privilege and will cause errors in pipeline steps.
Known attorneys
To create a new Known Attorney manually:
- Hover over the aiR for Privilege icon on the navigation bar.
- Hover over Setup (Priv).
- Select Known Attorneys (Priv).
- Click New Known Attorney.
- Complete the Known Attorney form.
- Click Save.
Known Attorney form fields
The following fields appear on the Known Attorney form:
| Field | Field type | Required field | Description | Notes |
|---|---|---|---|---|
|
Email Address |
Fixed-length text | Y | The full email address for the attorney. | Must be unique. Must be a single value. |
| Full Name | Fixed-length text | Y | The full name of the attorney. | Should be in the format [First Name] [Last Name]. |
|
Role |
Single choice | Y | Designates whether an attorney is playing a legal role or not playing a legal role in this matter. | Must be a value of Legal or Non-Legal. |
|
First Name |
Fixed-length text | N | The first name of the attorney. | Should align with Full Name. |
|
Middle Name |
Fixed-length text | N | The middle name of the attorney. | |
|
Last Name |
Fixed-length text | N | The last name of the attorney. | Should align with Full Name. |
|
Maiden Name |
Fixed-length text | N | The maiden name of the attorney. | |
|
Nicknames (Comma Delimited) |
Fixed-length text | N | Any nicknames for this attorney. | Multiple values may be added if separated by commas. |
|
Start Date |
Date | N | The date in which this attorney started playing a legal role. | If left blank it will be assumed open ended. |
|
End Date |
Date | N | The date in which this attorney stopped playing a legal role. | If left blank it will be assumed open ended. |
Note: If an attorney has multiple emails addresses you should create multiple items with different email addresses but the same Full Name
Known law firms
To create a new Known Law Firm manually:
- Hover over the aiR for Privilege icon on the navigation bar.
- Hover over Setup (Priv).
- Select Known Law Firm (Priv).
- Click New Known Law Firm.
- Complete the Known Law Firm form.
- Click Save.
Known Law Firm form fields
The following fields appear on the Known Law Firm form:
| Field | Field type | Required field | Description | Notes |
|---|---|---|---|---|
|
Firm Name |
Fixed-length text | Y | The complete legal entity name of the law firm. | Example: Jane & Doe LLP |
|
Domain |
Fixed-length text | Y | The domain used within the email addresses for employees of the law firm. | Must be unique and formatted without the @ symbol Example: law-firm.com |
|
Alternative Name 1 |
Fixed-length text | N | Another variation of how this law firm may be named in a document. | Example: Jane |
|
Alternative Name 2 |
Fixed-length text | N | Another variation of how this law firm may be named in a document. | Example: JD LLP |
|
Alternative Name 3 |
Fixed-length text | N | Another variation of how this law firm may be named in a document. | Example: JD |
|
Alternative Name 4 |
Fixed-length text | N | Another variation of how this law firm may be named in a document. |
Example: JDoe |
Note: If a law firm has more than one domain, you should create many items with different domains but the same Firm Name.
