INSIGHT: GCs Should Turn to AI to Save Time, Money, Manage Data
By: Haresh Bhungalia | Casepoint
The time has come for corporate legal departments overseeing large litigation portfolios or facing frequent inquiries and investigations to move beyond the limitations of case-by-case matter management.
In the age of big data and process automation, a siloed approach to serial litigation is a huge missed opportunity. When departments analyze data from multiple matters over time, significant patterns and trends that were once invisible suddenly become clear. The insights that emerge can inform a range of decisions and strategies that enable smarter litigation, more accurate forecasting and budgeting, and more efficient distribution of resources across legal workflows.
For example, data analysis suggests that the most important custodians in certain types of litigation, or in matters focused on certain products or business units, often overlap from one case to another. Some organizations have turned to proactive collection of data from “hot” custodians for specific categories of litigation or investigation so they can act more quickly and decisively as new matters emerge.
In a similar vein, some organizations have set up multi-matter data repositories to create cross-matter efficiencies spanning the entire e-discovery process. Legal departments are finding that reuse of data, coding and attorney work product in document review can lead to substantial cost reductions over multiple related matters.
A combination of artificial intelligence (AI) technologies make these and other initiatives both possible and practical by identifying with increasing precision where overlaps occur and which data and work product to focus on.
At the heart of a rigorous, data-based approach to matter management is a cluster of AI technologies, including data analytics, machine learning, predictive modeling and natural language processing. These are not simply additional software applications to add to an increasingly fragmented legal technology landscape.
Instead of thinking of AI as another “application,” it makes more sense to view it as a group of underlying processes that detect patterns by analyzing raw data. When data is subjected to AI-powered analysis across functions and legal matters, it yields new insights that can help legal departments save time and money by pinpointing the causes of inefficiencies and providing a road map for process improvement.
Examples of Efficiencies
Repeat litigants can apply advanced analytics to perform early case assessments (ECAs) focused on data collection. If you litigate a series of similar matters simultaneously or over a period of years, analysis of culling activities in one matter can inform culling strategies in subsequent matters.
For instance, data analysis may reveal you consistently culled specific data types such as brochures or presentations or PDFs in the first matter. That insight then informs your approach to culling as you perform multiple process iterations across multiple matters. When you apply machine learning to each new iteration and each new matter, culling requirements become clearer much earlier in the process, and that means you can significantly reduce data volumes downstream, saving a lot of money as you go.
Thinking again about ECA, perhaps you wish to focus on addressing the merits of similar cases earlier in the discovery process to arrive at a negotiate-or-litigate decision before collecting and processing large amounts of data at steep per-gigabyte prices. Certain categories of litigation will repeatedly hinge on a small number of specific legal topics or issues.
Data analytics and natural language processing will help you identify the cases that exhibit these similarities, and then identify the documents and document types that have historically proven most relevant to those issues. With each new case, you have the means to find such documents more quickly and with greater accuracy, and you are in a position to make strategic decisions before having to make significant investments in e-discovery and the services of outside counsel.
Limitless Possibilities in the Law Firm
Once you begin to think this way, the possibilities are nearly limitless. Predictive modeling technology can take billing and invoicing data from a series of similar matters and help increase the accuracy of litigation cost forecasting, planning and budgeting in subsequent matters.
You can analyze the same data to inform decisions regarding the hiring and management of outside counsel, quantifying overall and categorical spending from case to case, but also more granular factors like the relative speed and efficiency of a firm’s junior and senior associates performing specific tasks. You can also use these and other metrics to determine, for example, which outside firms are most cost-effective working on certain types of matters or in specific practice areas.
AI technology can even be used to test its own effectiveness in particular contexts. Let’s say you aren’t certain which legal matters benefit the most from the use of technology-assisted review (TAR), a relatively early application of AI technology in e-discovery. In recent years, your company has litigated 12 intellectual property matters in a specific product category, seven of which conducted document review using traditional methods and five of which applied TAR to the review process.
AI technologies can quickly digest the data related to these 12 review projects and sort the variables, leaving you with actionable information that reveals which matter characteristics—whether that points to data volumes, data types, data complexity, number of custodians, specific legal issues or whatever—are most relevant when determining whether TAR represents the most efficient approach.
AI in the legal department is not a passing fad, and multi-matter portfolio management is only one way in which organizations stand to benefit. Eventually we can expect to see many legal departments consolidate on a single, extensible technology platform with AI built in, but AI can also be profitably applied to relatively fragmented technology infrastructures. After all, each of those applications is producing mountains of data related to your entire portfolio. Why not use it?
This column does not necessarily reflect the opinion of The Bureau of National Affairs, Inc. or its owners.