Sunday, June 23, 2024

Let’s talk Data Models: Designing AI Based Knowledge Brain for Law Firms. – Part 2

Must read

In the first post, we talked about concept of knowledge graph and then about Legal Ontologies. We also went through various considerations in designing such knowledge brain for law firms like security.

Designing AI Based Knowledge Brain for Law Firms. – Part 1

https://biworld.ca/2023/05/designing-ai-based-knowledge-brain-for-law-firms-part-1/

Now in this post, let’s explore various models we will need as foundations for designing and developing such AI based knowledge brain.

Person

This entity can represent anyone associated with the firm, including lawyers, paralegals, clients, and other staff.

Attributes might include name, role, location, contact information, areas of expertise, years of experience, eduction, language

Case

This entity represents a legal case that the firm is involved in.

Attributes could include case ID, case type, involved parties, client, assigned lawyers, court, status, and outcome, start date, end date

Document

This entity represents any document related to a case or the firm’s operations.

Attributes might include document ID, document link (link in DMS), type (e.g., contract, court filing, internal memo), associated case (client and matter reference), author, security map, creation date, and last modified date. Attributes can be further expanded to include summary of the document, keywords, topics etc.

Event

This entity represents a significant event, such as a court date, meeting, or deadline.

Attributes could include event ID, type, associated case, involved persons, date, location, description.

Billing Record

This entity represents a record of billable work.

Attributes might include record ID, associated case, person who performed the work, description of work, category, hours, rate, total amount billed, and date.

Time Record

This entity represents a record of time spent on a case or task.

Attributes could include record ID, associated case (client and matter), person who performed the work, billable/non-billable, description of work, and hours.

HR Record

This entity represents a record related to human resources, such as hiring, promotions, or performance reviews.

Attributes might include record ID, person, type of record (e.g., hiring, promotion, performance review), details, and date.

See also  Top Free Resources To Learn Artificial Intelligence

Practice Area

This entity represents a specific area of law that the firm practices.

Attributes could include area name, leadership, description.


These are just example models that we might need, to build scalable foundations for this Knowledge Brain system. Now the question is where can we find the data behind these models? These can be find in various applications

Documents management system – majority of Law firms have some sort of DMS. Categorizing the documents in DMS is really important. Use metadata like Client and Matter, Practice Area, Department, Document Type etc.

Time & Billing System – considering data from both time keeping system and then billing system.

Ethical Walls System – walls can be on client or case level. This information is so important as base for maintaining security and integrity at levels of this Knowledge brain.

Closing Library – this is the place we will generally have outcomes related information for the cases.

CRM. – CRM is generally source of truth for data like contacts, opportunities. Even before a firm becomes a client, its information is captured in CRM. Hence for Client information lifecycle, CRM is very important.

Work Allocation Tool – data related to how work got requested and allocation resides in Work Allocation tool. Data from this tool can be useful for analysis related to DEI, Work Allocation, Insights on best fit resource based on case, removing biases etc.

In upcoming posts, we will design some of the applications that can be developed based on these models like Case outcome, How similar are two cases?, Predict risks & profitability for cases based on trend from historical similar cases etc.

Once that’s done, we will try building knowledge graph and then build LLM/Generative AI application with context of Law Firm using the discussed models. Stay tuned!!!

More articles

Latest articles