Insights ECI (Early Case Intelligence)

Overview

Early Case Intelligence (ECI) is an AI-powered tool designed to provide rapid insights into your document collection before initiating a full document review. Unlike traditional Early Case Assessment tools that rely on basic keyword frequencies and metadata, ECI uses generative AI to understand the substance of your case and group documents accordingly.

This tool enables legal teams to:

Prerequisites

System Requirements

Before beginning the Early Case Intelligence process, ensure you have the following:

1. Document Requirements: 
   • Text-Based Documents: All documents (emails, Word documents, PDFs, etc.) must have extractable text
   • File Size Limitations: Text-based documents should have an extracted text size of less than 300 KB
   • Supported File Types: The tool supports emails, Word documents, PDFs, Excel files, PowerPoint presentations, and text files

2. Case Information: 
   • Have a clear understanding of the key issues, relevant date ranges, and parties involved in your matter
   • Be prepared to describe the case and relevance criteria in detail

Components

Upload Interface

The upload interface allows you to submit your documents for AI analysis. You can either:        

1. Send documents via FTP directly to eDiscovery AI
   

• Select documents from an existing Relativity workspace through a simple mass operation

Case Background

The Case Background section is where you provide context about your legal matter to eDiscovery AI. This is a critical component that directly impacts the quality of the analysis:

Limitations

1. Case Background - 20,000 characters
   
2. Relevant Description - 5,000 characters
   
3. Not Relevant Description - 5,000 characters

6. Core legal issues and claims
7. Key parties and their relationships
8. Relevant time periods
9. Industry-specific terminology or jargon
10. Types of documents expected to be relevant
11. Examples of what would and would not be considered relevant
    

Document Analysis

After review, ECI provides several types of analysis:

1. 1. Relevance Categorization: Documents are automatically sorted into three tiers:
   
1. Likely Relevant: Documents that strongly match the matter description and key issues
2. Potentially Relevant: Documents with possible connections to case issues that merit further review
3. Not Relevant: Documents that are completely unrelated to your matter

2. Keyword Correlation Analysis: Identifies which keywords and terms are most strongly correlated with relevant documents in your specific matter. Like all of the visuals, the word cloud information can be downloaded in a table format by clicking the download button in the upper right corner. 

3. Document Type Breakdown: Visual representations showing the composition of your document set based on the content of your documents and the Case Background. 

4. Relevant Document Breakdown: Visual breakdown of the various subjects and topics found in the relevant document set. This is unique for every data set and informed by the user input in the Case Background so making the organization much more useful for each matter. 

5. Not Relevant Document Breakdown: A Visual breakdown of the various subjects and topics found in the not relevant document set. This allows users to review and ensure the not relevant topics are in fact not relevant to the matter.  
   

5. Inappropriate Content: Using generative AI, your documents will be analyzed based on sentiment for various in appropriate content types. This allows users to identify and review concerning documents or communications even if not directly related to the par legal matter. 

7. Narrative: A comprehensive summary that: 

4. Describes the overall composition of your document set
5. Identifies exemplar documents that illustrate key issues with document citations
6. Highlights areas of concern requiring further investigation
7. Provides strategic recommendations for review approach and prioritization

Process

1. Preparing Documents

1. Ensure your documents have been properly processed with extracted text
2. Verify that documents fall within the supported file types and size limitations
3. If working with non-English documents, no special preparation is needed as ECI is language-agnostic
4. Writing Effective Matter Descriptions
5. The matter description is crucial for accurate AI analysis. Follow these best practices:

1. Be Specific and Comprehensive: 

3. Include as much detail as possible about the legal issues and context
4. Define the key terms and concepts relevant to your case
5. Describe the specific document types or content that would be considered relevant

2. Provide Examples: 

2. Include examples of the types of communications or documents that would be relevant
3. Specify what would not be considered relevant to help the AI understand boundaries

3. Include Industry Context: 

2. Define any industry-specific terminology or acronyms
3. Explain company-specific terms that might appear in documents

4. Set Clear Parameters: 

2. Specify relevant date ranges if applicable
3. List key individuals or departments whose communications are of interest

Example of an effective matter description:

Copy

This case involves allegations of patent infringement of US Patent #12345 by Acme Corp against XYZ Inc. 

The patent covers methods for processing digital payments through mobile devices. Key invention features include: 

secure tokenization, biometric verification, and real-time fraud detection.

Relevant documents would include:

- Technical specifications of XYZ's mobile payment systems developed between 2018-2022

- Communications between XYZ's engineering team and payment processing partners

- Product development meeting notes where mobile payment security was discussed

- Competitive analysis of Acme's patented technology

Not relevant would be:

- Routine HR communications unrelated to the technology

- Marketing materials that don't describe technical functionality

- General financial reports not specific to the mobile payment products

      

Reviewing AI-Generated Analysis

After processing, review the ECI analysis to understand your document set:

Examine Document Categories

2. Review the distribution of documents across relevance tiers
3. Sample documents from each category to verify accuracy

2. Study the Case Memo: 

3. Review the AI-generated case summary for key insights
4. Evaluate exemplar documents identified by the AI
5. Consider strategic recommendations for review

3. Analyze Keyword Correlations: 

2. Note which terms are most strongly associated with relevant documents
3. Consider using these terms to refine search strategies in full review

4. Review Document Type Breakdown: 

2. Identify which file types contain the most relevant information
3. Note date ranges with high concentrations of relevant material
4. Using Insights in Review Strategy
5. Apply ECI insights to optimize your document review strategy:

1. Prioritize Review Batches: 

2. Begin review with documents identified as "Likely Relevant"
3. Use ECI's thematic groupings to create focused review batches

2. Develop Review Protocols: 

2. Use exemplar documents to train reviewers on relevance criteria
3. Create review guidelines based on AI-identified patterns and themes

3. Refine Search Methodology: 

2. Incorporate AI-identified keywords into search strategies
3. Focus on date ranges and custodians with higher concentrations of relevant material

4. Inform Case Strategy: 

2. Use early insights to guide deposition preparation and other strategic decisions
3. Identify potential strengths and weaknesses in your document collection and merits of your case
4. Relativity Integration
5. Field Mapping
6. ECI seamlessly integrates with Relativity:

1. Automated Field Creation: 

2. ECI will create the necessary fields in Relativity to store its analysis results
3. These fields include relevance determinations, document summaries, and category assignments

2. Custom Field Mapping: 

2. Users can map ECI's categories and insights to custom fields in their Relativity workspace
3. This flexibility allows integration with existing review workflows
4. Transitioning to Document Review
5. Move from ECI analysis to full document review:

1. Transfer of Insights: 

2. All insights generated by ECI are directly mapped to Relativity fields
3. No manual transfer of information is required

2. Creating Review Batches: 

2. Use ECI's relevance categorizations to create prioritized review batches
3. Focus initial review efforts on documents identified as most likely relevant

Review Acceleration

2. Pre-categorized documents and identified themes significantly reduce review time
3. Reviewers can focus on documents with highest likelihood of relevance first