Legal professionals can utilize the Legal Data Intelligence model to identify data relevant to litigation, as well as to determine which documents should be withheld from production for reasons such as attorney-client privilege or the work product doctrine.

Document Review Optimization

Uses analytics and AI systems to identify potential gaps or areas of focus

Enables analysis of the merits of the case and comparisons to similar or historic cases

Captures and organizes objective metrics for each matter (e.g. fees, costs incurred, time to resolution) to enable informed decision-making

How Technology Can Assist

Use of analytics and AI should be subject to local regulation, transparency requirements, and ethical constraints.

Notes

Scope Project

Tracks data sources from identification through preservation, collection, and processing to ensure nothing is missed

Quickly identifies sources and locations of potentially relevant data when using a well-integrated and up-to-date knowledge management or information governance system

Analyzes common data to identify additional data sources or off-channel communications

Determine if a client's in-house systems have search and identification functionality that is sufficiently robust.

Identify Data

Provides legal hold functionality to preserve data in place to ensure it cannot be deleted (where appropriate) 

Tracks the implementation and oversight of each legal hold, including periodic reminders and audits to ensure compliance

Ensures data that should be preserved is correctly secured as required (note that once preserved, data need not be processed until or unless needed) 

Ensures data is preserved and collected while maintaining both file and system application and system metadata (if appropriate)

Take care to ensure that data is collected in line with legal requirements in the relevant jurisdiction(s).

More challenging sources, or sources containing potentially ephemeral data like mobile devices, may need to be imaged swiftly to ensure preservation.

Preserve and Collect Data

Applies filters and other criteria to minimize data to be extracted and processed, thereby reducing the processing of ROT data 

Quickly manages large data volumes  

Reads and records metadata of files for searching and analytics 

Extracts attachments and expands compound files 

Indexes content to enable searching and analytics 

Identifies encrypted files and files that could not be read for exception handling 

Eliminates manual workflows and reduces human error

Reduces hosting of ROT data and thus hosting cost through de-duplication, de-NISTing, etc.

Extract and Process Data

Allows for complex Boolean searches and filtering using content and metadata 

Performs complex searches and summarizes large data sets using generative AI 

Enables custom tagging and foldering to classify and stratify data  

Reduces review volume/ROT data with tools like email threading, deduplication or suppression of near-duplicates, categorization, and clustering 

Codes documents based upon training or prompts using traditional and generative AI tools 

Creates custom chronologies using traditional and generative AI tools

Search and Evaluate Results

Performs complex searches, including regular expressions 

Identifies specific data sets and privileged documents through prompts and training using generative AI 

Conducts visualization methodologies and gap analysis 

Summarizes files for potential use in listings (e.g. privilege logs) 

Provides explanations and citations to help validate output

Analyze Data

Prioritizes data for review based upon training or prompts using traditional or generative AI tools

Exports fielded data in custom formats for metadata privilege logs, key document chronologies, and so on using review and production platforms

Apply Strategic Decision-Making

Applies defensible verification of any predictive coding or generative AI methodologies

Tracks document counts and tags files  

Automates and/or applies redactions  

Performs quality control by identifying errors or inconsistent coding, standardizing production rules, etc.

Synthesize and Redact

Applies Bates or other stamping, if needed 

Creates output in usable, secure forms for all parties 

Produces data that meets requirements in local regulation, agreed protocols, or court orders

Generate Output

Produce

Record/Document

Litigation and Dispute Resolution

Companies must notify the Federal Trade Commission (FTC) before proceeding with a merger. If the FTC has concerns about the transaction, it may issue a Second Request for additional information. The Legal Data Intelligence model can be employed to identify and produce relevant documents in response to the request, ensuring there are no delays in finalizing the transaction.

Second Request Compliance

Identifies potential gaps or areas of focus using analytics and AI systems

Captures and organizes objective metrics for each matter (e.g. fees, costs incurred, time to resolution) to enable informed decision-making

Determine if a client's in-house systems have search and identification functionality that are sufficiently robust.

More challenging sources, or sources containing potentially ephemeral data like mobile devices, may need to be imaged swiftly to ensure preservation.

Collect Data

This step should happen in parallel with the Collect Data and Extract and Process Data steps. 

Continue to evaluate and assess to ensure the review protocol remains in line with requirements.

Develop Review Protocol

Allows for complex Boolean searches and filtering using content and metadata 

Performs complex searches (potentially based directly on requests from regulators) and summarizes large data sets using generative AI  

Quickly identifies information suggesting that there may be an issue using generative AI-powered early case assessment tools, which can be used to assist clients in swiftly determining the best way to respond to the regulator(s)

Enables custom tagging and foldering to classify and stratify data 

Reduces review volume/ROT data

Reduces the ultimate universe of data to be reviewed using email threading, deduplication or suppression of near-duplicates, categorization, and clustering 

Codes documents based upon training or prompts using traditional and generative AI tools

Creates custom chronologies using traditional and generative AI tools

Recovers deleted material 

Extracts potentially useful or relevant information from non-user files using digital forensic techniques (e.g. internet search histories or location information from mobile devices) 

Performs complex searches including regular expressions 

Identifies specific data sets and privilege through prompts and training using generative AI

Creates visualizations and allows for gap analysis 

Summarizes files for potential use in listings (e.g. privilege logs) 

Provides explanations and citations to help validate output

Applies Bates or other stamping, if needed 

Creates output in formats as specified by the regulators 

Produces data that meets requirements in relevant jurisdiction(s) and agreed protocols

Regulatory Requests

The Federal Rules of Civil Procedure (FRCP) and state equivalents in the U.S. protect Third Parties from undue burden when responding to subpoenas. Scope must be negotiated by limiting the number of custodians, data sources, or by date to minimize burden.

Quickly identifies sources and locations of potentially relevant data when using a well-integrated and up-to-date knowledge management or information governance system

Enables search and analysis of data in place to avoid time-consuming and burdensome collection and hosting

Enables search and analysis of data in place prior to formal collection

Collects data in a forensically sound manner that preserves system and application metadata, as well as folder and file paths, per process requirements

Quickly manages large data volumes 

Applies standard processing configurations (including deduplication) to reduce the total volume for review

Enables email threading and similar tools that can help prioritize review as well as reduce the total number of documents for review 

Eliminates manual workflows 

Reduces human error

Process Data

TAR training will likely require a few iterations until the results meet the agreed-upon level of confidence.

This step should happen in parallel with the Collect Data and Process Data steps.

Enables rapid testing and iteration of search terms and queries so that the most relevant data can be identified while avoiding bringing irrelevant material into the review queue

Search

Enables comparison of coding treatment across duplicate, near-duplicate, and subject matter categories

Facilitates evaluation of reviewer performance and accuracy

Enables review to ensure consistency of coding (including privilege status) across reviewers

Evaluate Data

Provides analytical tools that can expose date gaps in collections, missing email or messaging correspondence, and other modes of communication

Quality control generally involves case team members reviewing sample sets or specific collections of interest to confirm that the coding has been applied correctly.

Programmatically creates the production set defined by the case team in volumes of manageable size

Securely transfers large amounts of data subject to complex encryption 

Creates output in usable, secure forms for all parties

Third-Party Subpoenas

The compliance team at a pharmaceutical company initiated an internal investigation following a regulatory inquiry. Given the unprecedented size and scope of the data that required analysis, the company&rsquo;s Legal Data Intelligence practitioner leveraged the LDI model to identify the SUN (sensitive, useful, necessary) data within both internal and external communications. This approach enabled the compliance team to provide a comprehensive and thoughtful response to the regulator.

Regulatory Inquiry Response Strategy

A company tasked its Legal Data Intelligence leader with helping investigate alleged workplace harassment. Leveraging the LDI model, the leader identified and analyzed relevant employee communications in which employees were referring to the individual involved. Filtering ROT (redundant, obsolete, trivial) data from SUN (sensitive, useful, necessary) data, the LDI practitioner narrowed the dataset to 500 key documents, enabling the firm to make informed decisions on next steps.

Targeted Discovery in Sensitive Investigations

A utility company conducted an internal investigation in anticipation of a class action lawsuit over its role in a major wildfire in California.

Proactive Litigation Readiness

Enables analysis of the merits of the case and comparisons to similar or historic cases

Captures and organizes objective metrics for each matter (e.g. fees, costs incurred, time to resolution) to enable informed decision-making

Tracks data sources from identification through preservation, collection, and processing to ensure nothing is missed 

Assists with the identification of relevant data sources and locations 

Analyzes common data to identify additional data sources or off-channel communications

Ensuring sufficient identification may be a particular consideration if there is risk of related litigation or regulatory action pending the result.

Determine if a client's in-house systems have search and identification functionality that is sufficiently robust.

Identifies potential data sources instead of conducting interviews  

Provides legal hold functionality to preserve data in place and ensure it cannot be deleted (where appropriate) 

Collects data covertly, remotely, or both in many cases 

Tracks the implementation and oversight of each legal hold, including periodic reminders and audits to ensure compliance 

Ensures data that should be preserved is correctly secured as required (note that once preserved, data need not be processed until or unless needed) 

Ensures data is preserved and collected while maintaining both file and system application and system metadata (if appropriate)

Applies filters and other criteria to minimize data to be extracted and processed, thereby reducing the processing of ROT data 

Quickly manages large data volumes  

Reads and records metadata of files for searching and analytics 

Extracts attachments and expands compound files 

Indexes content to enable searching and analytics 

Identifies encrypted files and files that could not be read for exception handling 

Eliminates manual workflows and reduces human errorâ€¯ 

Reduces hosting of ROT data and thus hosting cost through de-duplication, de-NISTing, etc.
Automates the preservation of electronically stored data (mailbox and laptop data, network drives, SharePoint, etc.)

Pulls data from sources into a platform for processing

Reduces collection of ROT data by excluding non-relevant documents by using date filters, export rules to filter junk/system files, etc.

Allows for complex Boolean searches and filtering using content and metadata 

Performs complex searches and summarizes large data sets using generative AI 

Enables custom tagging and foldering to classify and stratify data  

Reduces review volume/ROT dataâ€¯with tools like email threading, deduplication or suppression of near-duplicates, categorization, and clustering 

Codes documents based upon training or prompts using traditional and generative AI tools 

Creates custom chronologies using traditional and generative AI tools

Performs complex searches, including regular expressions 

Identifies specific data sets and privileged documents through prompts and training using generative AI 

Conducts visualization methodologies and gap analysis 

Summarizes files for potential use in listings (e.g. privilege logs) 

Provides explanations and citations to help validate output  

Codes information for privilege/privacy, which is useful if it's determined that a chance of litigation could come later 

Analyzes emails/files for signs of manipulation or fabrication 

Enables digital forensic analysis to identify and/or recover deleted material or extract non-user information that may be relevant (e.g., web history information, recently connected devices)

Generates summaries of documents or groups of documents using generative AI 

Generate timelines of actions through prompts and training using generative AI 

Identifies key questions that should be asked of an employee based on findings in the documents and the allegations that have been made using generative AI

Act on Results

Internal Investigations

A Legal Data Intelligence leader in a law firm was engaged to review contracts for a real estate client. During preliminary analysis, they identified pricing anomalies in several agreements. Further investigation revealed that one lease contract included a higher markup due to undisclosed relationships between the parties involved.

Pricing Anomaly Review

Asked to review a corpus of contracts for a client, a Legal Data Intelligence practitioner employed an LDI approach. The analysis revealed that the client was missing out on potential revenue by not enforcing price escalation clauses. In this case, the LDI leader delivered direct business value by empowering the client to realize additional revenue.

Revenue Optimization Analysis

Confirm Budget and Vendor RFP Process

Issue RFP and Review Bids

Evaluate Contract Volume Versus Available Resources

Migrate Data

Quickly manages large data volumes

Eliminates manual workflows and reduces human error

Uses machine learning to auto identify key provisions for efficient review

Uses heat mapping and other culling features to identify provisions for batch exclusion based on template provision

Select Data

Utilizes auto-extracted provisions and other built-in machine learning features to perform lawyer-in-the-loop review

Complete Review

Depending on the outcome of the due diligence, certain steps may be necessary to close the deal

Act on Results of Review

Contract Reviews

A Legal Data Intelligence leader at a financial services company developed an AI-based metric to gauge the level of effort required for negotiating contracts. This metric enabled the company to predict the time needed to finalize a contract, estimate the number of negotiation rounds if it&rsquo;s drafted on the client&rsquo;s paper, and assess the negotiation&rsquo;s business impact.

Contract Negotiation Effort Metrics

Can vary significantly by company, including non-disclosure agreements (NDAs), master service agreements (MSAs), statements of work (SOWs), etc.

Develop Playbook and Model Agreement

Uses machine learning to automatically extract the identified provisions after the requisite training set is identified

Identify Data Extraction

Complete Intake Agreement

Requires inputting certain data and specific data being automatically captured

Tracks the process and includes checklists to ensure agreement deadlines are met

Input and Elevate Agreement

Attorney confirms/tweaks the redlines within the technology or inserts comments, depending on company's editing preference

AI is typically only used during the first turn of a pre-execution negotiation

Redline Agreement

Send Redlines and Continue Negotiations

These fields can be used for future business or M&A due diligence should the need arise

Input Final Agreement into CLM System

Establish Notifications

Contract Lifecycle Management

Legal Data Intelligence practitioners at a law firm scoured the data pools of a soon-to-be divested business to identify and secure confidential information before handing over the company&mdash;including the underlying data&mdash;to the acquirer. The stakes were very high&mdash;the fate of the deal hung in the balance because it was contingent on both parties feeling confident about their process and the accuracy of their output.

Securing Confidential Information in a Divestiture

A Legal Data Intelligence practitioner conducted research and due diligence for a possible acquisition of a company to determine whether the transaction would pose any potential regulatory risk.

Acquisition Due Diligence

Automates review and assessment of deal documents

Tracks relevant deadlines and milestones

Analyze Deal

Automates review and assessment of deal documents

Manages personnel information and organizes relevant data regarding key individuals

Assess Individuals

Understand Implications

Organizes and manages corporate infrastructure

Tracks and groups data sources and related information in organized format

Manages relationships with vendors and compiles related information

Identify Data Sources

Collates documents in organized fashion for review and coordination

Inventory Documents

Transmits legal hold notices to custodians

Tracks legal hold acknowledgments

Manages personnel information and organizes relevant data regarding custodians

Explore Custodian Implications

Analyzes, tracks, and reviews relevant contracts

Organizes contracts and relevant documents

Monitors changes in regulator structure and requirements

Consider Regulatory and Contractual Issues

Organizes and manages corporate infrastructure

Tracks and organizes information related to specific custodians and preservation obligations

Monitors data system usage and revisions

Execute Changes

Manages and collates information regarding disputes

Handles discovery in an efficient and effective manner

Analyze Impacts

Organizes and manages corporate infrastructure

Tracks and organizes information related to specific custodians and preservation obligations

Establish Protocols

Mergers, Acquisitions, and Divestitures

An environmental agency in North America routinely handles FOIA requests. To ensure timely and accurate responses, the agency relies on Legal Data Intelligence practitioners fluent in discovery and technology to efficiently identify relevant data and redact personal and confidential information before production.

FOIA Request Data Management

Receive and Review Request

Provide a First Response to Request

Uses enterprise search technology and data collection tools

Identify Potentially Responsive Records

Utilizes search and AI technology to assist in identifying the responsive information from the potentially relevant set

Uses tagging, comments, and notes to mark whether a document is responsive to the request and record decisions

Review Potentially Responsive Records

Employs redaction technology to redact material that must be excluded from otherwise responsive records

Uses tagging technology to record the specific legislative exception or exemption relied upon for each redaction

Assess Exemptions for Exclusions from Production

Consult with Stakeholders

Leverages a review platform to identify and locate the responsive documents including any redactions, print or download the documents for transfer, and provide the needed explanations for withheld documents and redactions

Prepare Response

Uses technology to share the proposed documents and response to others for review and approval

Review and Approve

Provide Response

Monitor and Follow Up

Utilizes the tagging, comments, and note functionality of the review platform to record all steps and decisions when reviewing documents to have a complete record of all steps

Document the Process

Freedom of Information Act (FOIA) Requests

A global consulting company relies on its in-house Legal Data Intelligence team to manage external Data Subject Access Requests (DSARs). While the HR department typically handles DSARs from current and former employees, the in-house legal team is responsible for all other DSARs.

In-house DSAR Management

Receive Request

Automates the verification process and timeline creation

Helps track and catalog available data sources

Future generative AI technologies may make it easier to identify relevant data using LLMs and prompts

Pulls data from sources into a platform for processing

Reduces collection of ROT data by excluding non-relevant documents by using date filters, export rules to filter junk/system files, etc.

Quickly manages large data volumes

Eliminates manual workflows and reduces human error

Reduces hosting of ROT data through de-duplication, de-NISTing, etc.

Looks across internal systems and databases to locate the data that must be provided under the DSAR, resulting in greater consistency across requests

Reduces review volume/ROT data

Evaluate Results

Provides explanations and citations to help validate output

Tracks document counts and tags files

Automates redactions to protect personally identifiable information (PII) and adhere to privacy regulations

Helps with quality control by identifying errors or inconsistent coding, standardizing production rules, etc.

Creates output in usable, secure forms for all parties

Fulfill Request

Data Subject Access Requests (DSARs)

Following a data breach, a school board engaged a Legal Data Intelligence leader to identify impacted data and compile a list of affected individuals. Leveraging AI and LDI practices to analyze intricate data belonging to students, faculty, and parents, the breach&rsquo;s scope was accurately assessed, enabling the school board to take necessary next steps.

Data Breach Impact Assessment

After a company&rsquo;s email server was breached, exposing hundreds of thousands of emails, a Legal Data Intelligence leader leveraged AI to assess the compromised data. By ruling out personally identifiable information in over 90% of the emails, the review burden was effectively reduced to a manageable dataset.

PII Filtering for Review Management

Scope Project and Set Goals

Identify Potentially Compromised Data

Pulls data from sources into a platform for processing

Upload Potentially Compromised Data

Provides a variety of functionalities to conduct the initial assessment, including search terms, regular expressions, Boolean operators, AI, and analyses and categorization of document types, titles, sizes, and subject matter

Carry Out Initial Assessment

A variety of functionalities -- ranging from search terms, regular expressions, Boolean operators, and AI -- may be chosen to carry out this process.

Identifies sample documents from the Review Set to provide as appendices to the Protocol to demonstrate the nature of the information reviewers are likely to come across and how to treat different information types

Develop a Review Protocol

AI enables the quick extraction of personal information and the association of entities (affected individuals) with linked PI, PHI, and PSI

Review Data

AI tools populate the affected individuals in a column and all their associated PI traces in rows

AI easily normalizes duplicate records belonging to the same individual

Create a Notification List

AI helps prevent duplicate notifications to the same affected individual and consolidates all notifications

Notify Individuals and Regulators

Data Breach Response

A pharmaceutical company utilized AI to develop a defensible strategy for identifying and redacting the personal information of European data subjects after their data was subpoenaed in the United States.

Cross-border PI Redaction

A Legal Data Intelligence leader at a pharmaceutical company used AI to identify and redact personal information (PI) and protected health information (PHI) within documents related to clinical drug trials.

AI-driven Redaction of PI and PHI

Analyzes legal filings against a structured reference data set of laws using text recognition / AI models

Automates purpose and lawful basis processes by using an exhaustive list of options for lawful basis to automatically populate data entered in a workflow

Automates risk tolerance processes by using an exhaustive list of risk levels and situational strategy decisions

Automates chosen action(s) based on operational guidelines and rules that are programmed into the software

Create Operational Guidelines and Rules

Pulls data from sources into a platform for processing

Reduces collection of ROT data by excluding non-relevant documents using date filters, export rules to filter junk/system files, etc.

Locates data containing personal or sensitive information

Reduces review volume/ROT data

Provides opportunity for human input and analysis recording

Leverages human input and analysis to update any project and/or client-level active learning or automation

Visualizes and dashboards statistics regarding both analysis of and actions applied to the personal or sensitive data in the data set

Documents decision points for defensibility

Tracks document counts and tags files

Automates redactions to protect PII and adhere to privacy regulations

Helps with quality control by identifying errors or inconsistent coding, standardizing production rules, etc.

Sets disclosure of protected data through secure password-protected transmission tools

Establishes audit and access logs using permission-oriented transfer technology

Protect Data

Personal Data Identification and Anonymization

During litigation, a global technology company headquartered in the Netherlands advocates for a Protective Order provision recognizing the applicability of the GDPR to custodial and non-custodial data collected, processed, and produced as relevant to U.S. discovery requests. The provision is included in a Protective Order entered by the court. Upon case closure, both parties must certify deletion and destruction of opposing party data to comply with the Protective Order. The technology company must also ensure deletion of the personal data it produced to comply with the GDPR.

Protective Order and Regulatory Compliance

A U.S.-based pharmaceutical company enters into an agreement with a U.S.-based legal technology company to provide document review technology and service support for U.S. discovery workflows. A provision in the contract requires both companies to destroy each other's confidential information within 30 days of contract termination. The contract terminated 10 days ago.

Contract Reviews

USE CASE IN PRACTICE

Pricing Anomaly Review

Revenue Optimization Analysis

Initiate

Confirm Budget and Vendor RFP Process

Initiate

Issue RFP and Review Bids

Initiate

Evaluate Contract Volume Versus Available Resources

Initiate

Migrate Data

Initiate

Process Data

How Technology Can Assist

Investigate

Analyze Data

How Technology Can Assist

Investigate

Select Data

How Technology Can Assist

Investigate

Complete Review

How Technology Can Assist

Implement

Act on Results of Review

Notes

Contract Reviews

Initiate

Investigate

Implement