Legal Data Intelligence Lessons from Relativity Fest: Part 1
Author: LDI Team
In this four-part series, we will reflect on vital insights, teachable moments, and exciting developments shared by founding members of the Legal Data Intelligence project at Relativity Fest 2024
Since its launch in May at CLOC Global Institute 2024, Legal Data Intelligence (LDI) has captured the attention of the legal industry as legal professionals grapple with ever-increasing data volumes, data complexity, and rapid technological shifts such as generative AI.
In a recent panel discussion at Relativity Fest, founding members of the Legal Data Intelligence model gave conference attendees a rare and revealing look at crucial moments in their own careers when they applied the ideas and principles at the heart of Legal Data Intelligence.
Titled How Legal Wins Budget and Influences Stakeholders, the panel discussion featured detailed case studies presented by founding members on how they helped their clients navigate complex data challenges, uncover new insights, and apply their skills to a broad and diverse range of use cases.
Here are some highlights from the discussion.
“Legal Data Intelligence” Has Become an Official Role
Kelly Friedman is a founding member of the Legal Data Intelligence project and a notable lawyer who was once voted “Top 25 Most Influential Lawyers” by Canadian Lawyer. Today, she holds the distinct honor of being the first person to hold an official Legal Data Intelligence job title, as she recently joined the Toronto-based firm Heuristica as their Chief Legal Data Intelligence Officer and Senior Counsel.
As legal professionals face increasing data volumes, complexity, and regulatory demands, the adoption of LDI principles is reshaping career pathways, as well as creating new service lines, like those highlighted in the following case studies.
Editor’s Note: For anyone interested, sample job descriptions around Legal Data Intelligence can be found here.
A School Board Suffers a Major Data Breach
During the session, Friedman recounted a deeply complex project. Not too long ago, a prominent school board in Canada was hit with a data breach potentially impacting sensitive data relating to students, parents, and the board’s employees. The sensitive data ran the gamut: "You have grades, you have disciplinary proceedings, you have financial information, you have passports, you have gender identities, health information...," said Friedman.
Further, there were two complicating factors that added to the complexity:
- Much of the data corpus was unstructured and commingled, which made it significantly harder to identify and flag sensitive information buried in it;
- The data existed in two different languages—English and French.
“I approached this project with the [mindset] that this is a legal data challenge. I have a legal problem: I have to notify a huge number of people that their data might have been compromised; I also have privacy regulators to notify. We are dealing with the personal information of tens of thousands of people mixed in with ROT [Redundant Obsolete and Trivial] data. Technology has got to help me here because this is not going to be a project where we look at one document at a time.”
Kelly approached a US-based AI company to help navigate the problem. “There were some unique challenges. There were no French language identifiers. We had to create identifiers. Likewise, Canadian social insurance numbers have a different format compared to American social service numbers. So we had a lot of work to do to figure how we were going to identify personal information and then link it to the individuals affected in the data breach.”
Even though this project came across her desk prior to the launch of the Legal Data Intelligence model, Friedman devised a new workflow that shared the same fundamentals as the LDI model.
At a high level, the process was designed around three key steps—Initiate, Investigate and Implement—that ultimately allowed Friedman and her team to successfully identify the SUN (Sensitive, Useful and Necessary) data buried in the ROT data.
“It was a huge success. The client was shocked as they knew how messy their data was,” Friedman said.
She also shared that the matter allowed her to open a new service line focused on data breaches. “Up until that point, we had mainly been doing litigation and investigations, regulatory requests and competition-related issues; but we hadn’t done data breach response. But after that we knew that we could handle it, and we knew that the tech could handle it.”
Reviewing Contracts to Find Opportunities to Increase Revenue
In the course of the panel discussion, founding member Scott Milner, who is a Partner and Global Head and Practice Group Leader of eData Practice Group at Morgan Lewis, shared an interesting success story about a contract review project that his team worked on for a client in the transportation space.
“They had no clue what was in the contracts, and they had no clue what they needed to get out of the contracts. So initially we used an AI model to pull out the standard clauses you want out of contracts: notice, terms, parties, type of contract —all the boring stuff,” Milner explained.
But what really piqued Milner’s curiosity was the moment the client made a slightly different ask; they wanted Milner’s team to help ensure they were fully taking advantage of price escalation clauses in their contracts.
“Lawyers are always mitigating risk. But I wanted to show the value of data. Legal is always seen just as a cost center but I want to flip the script [...] Businesses aren’t created to deal with litigation or discovery issues. I wanted to help the client go back and correct their revenue leakage.”
Milner’s team set out to work. The challenge, however, was that the contracts were not on the client’s paper, and the terms most important to the client were not common terms that an out-of-the-box AI model could identify. There were workflows for creating custom AI models for clause extraction, but not for these specific custom clauses. Milner’s team therefore had to build custom AI models. They had to get examples, train the model, and validate its outputs until they were comfortable with the quality of the results.
Ultimately, the team generated a report that enabled the client to act on its results and capture missed revenue opportunities.
After they successfully tackled this project, the Morgan Lewis team was further engaged by the same client to extract clauses around IP rights.
More Insights to Come
This panel discussion offered just a glimpse into the real-world application of Legal Data Intelligence, showing how this evolving model is reshaping the legal landscape. Stay tuned—over the next few weeks, we’ll share more insights and stories from Relativity Fest that highlight the innovative ways LDI is driving change in the legal industry.
Reach out to us at info@legaldataintelligence.org to share your feedback or find out how you can get involved.