Generative AI-Ready Structured Datasets

The iManage conference ConnectLive was back in New York City last week for the first time since 2019. Next post, I'll focus on a few key take-aways from the conference. This week I focus on an iManage-Hearst partnership in its 4th year.

CEO Neil Araujo and team have done an excellent job at envisioning legal's future to maintain iManage's relevance, including the acquisition of RAVN and nurturing its integration and evolution. As of 2024 a well-structured data set is essential to leveraging Generative AI against your internal information and iManage can provide that infrastructure for its 4,000 strong client base.

In 2021-2022, Hearst partnered with iManage to train a few RAVN models against our 500,000+ strong database with Hearst resources tagging results alongside the iManage team.

In 2023 building on what we learned, we launched a data enrichment initiative that applied several dozen RAVN-based models against our corporate contracts data and Thomson Reuters' Westlaw data against our litigation data. We conducted 1:1 user interviews, a follow-up survey, feedback sessions, and after intensive analysis came forward with a series of recommendations, including a training series with quick guides and brief videos and draft iManage use principles.

In 2024 we are piloting the new iManage INSIGHT+, which brings nearly 100 models, and over 100 new meta data fields, and the next generation Thomson Reuters integrations. Users continue to participate in feedback sessions. We are in process of upgrading the training into a badged learning journey  and incorporating evolving use principles into new employee onboarding.

Over this evolution we have seen 5 principle benefits:

1. Auto-application of meta data (replacing manual entry at time of creation) makes it much easier for law practitioners to locate the right document or document clause and incorporating specific data points such as governing law/jurisdiction or counterparty.

2. Having a shared taxonomy, consistently applied, means that it is easier to find specific relevant documents others have created.

3. Having the entire production library processed, consistently classified and enriched makes us Generative AI ready to incorporate our own internal legal information later this year.

4. Drafting briefs and researching in Westlaw is a more seamless experience reducing duplication and the toggle tax.

5. New features in the iManage roadmap are directly responsive to our law practitioner needs. We're particularly looking forward to trying Ask iManage, the practice area landing pages, and the locator services for matters and deals.

Honestly, in 2019 I was not sure what the long term prospects were for document management systems since the onset of Microsoft's One Drive and similar products.  However, the need for GenAI to draw on well-structured data and the increased team satisfaction with iManage demonstrates otherwise. While we have had longstanding 100% adoption in the US, international offices have now requested to join, and will be fully onboarded by end of year.  Onboarding new team members is also easier, as they can become more quickly self-reliant in locating examples to see how Hearst, or a specific Hearst business or location, responds to a specific case type, and can filter for a colleague's or supervisor's precedents.

At age 16 my first legal job was as interim law librarian at Rogers & Hardin in Atlanta. On Monday mornings, I would find books piled on the floor in stacks taller than an associate's head (and sometimes the associate face down asleep at the table). My first task was to re-shelve the books so they could be located quickly when next needed, and sometimes to go across downtown to return books to a law school of obliging law firm.  I would look up entries on the firm's sole Lexis terminal.  Our taxonomy was the Dewey Decimal System.

 When I saw my first document management system (Hummingbird founded 1984) with versioning and cross-filing I was hooked. Now married to a master magician, I understand that magic is science not yet commonly understood. Generative AI is a genji that has unleashed a new era of wonder in legal technology. Will GenAI be more exciting than the invention of the Apple II, SQL database, or Internet? I aim to find out in a hands-on way.

Image Credit: From presentation on INSIGHT+ at iManage ConnectLive 2024.