Evaluating the Current and Future State of eDiscovery and its Processes

By Nick Inglis

In my book Advancing from eDiscovery to Prediscovery, I provide a deep dive into the current state, process evaluation, and future state of Information Governance and eDiscovery, as well as how these disciplines can be aligned.

In this second in a three-part blog series about the book, I will summarize the second section, focusing specifically on the current and future state and process of eDiscovery.

The basis for the eDiscovery profession is, as has been said, best reflected in the EDRM, a linear step-based process released by George Socha, and Tom Gelbmann in 2005. It begins with identifying data related to a case, then moving along a straight line path to collecting, preserving, analyzing, processing, and finally presenting it.

While eDiscovery processes have remained chiefly static since this model was developed, new technologies like Artificial Intelligence (AI) and Machine Learning are gradually being adopted by more companies and firms and presenting an array of intriguing results that are shaking it up.

The professionals who are leading in the adoption of new technologies tend to view eDiscovery as more of a dynamic profession than the linear EDRM presents. Meanwhile, countless lawyers remain on the sidelines with these advancements and continue to perform their role with little changes to the process.

This dichotomy will not last, as those who refuse to adapt are slowly starting to miss out on more prominent cases, and their inability to perform the tasks as rapidly as their competitive firms are beginning to take a toll.

Why paper eDiscovery is moving to digital eDiscovery

There are many new technologies available now that can improve and expedite the discovery process than were available years ago, or even when the EDRM was developed. We now can drive efficiency in ways that we never could with paper.

At the most basic, because most paper records are now scanned to create digital copies of all data, we can parse and analyze most of our information more efficiently. Unlike paper that would need to be walked down a hall for the eDiscovery process to continue, one button push in the digital world can mean more than one action can take place at the same time.

As such, in this digital world, we no longer need to think of EDRM processes as single steps – we can layer steps on top of one another with digital data. We can eliminate multiple steps by simply moving that stage’s activities in line with other stages.

In fact, the first three steps of the eDiscovery portion of the EDRM when done with paper, could be described as, searching through a file room to identify, pulling papers out and putting them into your own folder to collect them, then walking down the hall and handing off to a Records Manager or Paralegal to preserve them.

With the right technologies and automation, these steps are no longer necessary.

The future state of eDiscovery

In reimaging eDiscovery with these new technologies, organizations are starting to see that the future of eDiscovery is no longer limited to a linear process.

Using secure, limited, and provisioned access to data, they are using technology that can search information in place without having to copy or move it. That view across their various repositories, allows them to follow the chain of actions in the EDRM in a single step as they move between systems – e.g. a chat in Microsoft Teams, leads to an email, leads to a collaboratively created document with three authors, leads to a signed version of that document that’s a contract sitting in a Records system (probably with 15 other copies in various email boxes and other systems).

Through one view of all that data, they can better understand the context and get a Live Early Data Assessment of all information related to the case, allowing them to merge EDRM processes through a rapid identify-analyze recursive process.

Preserve and analyze during collection

Preservation is another process step can be automated and folded into another part of the EDRM.

There is no reason that when organizations first cast their net for relevant information that their software should preserve all potentially pertinent and responsive information. Preservation, putting information on legal hold, should happen at the earliest stage of the process and protect relevant information from deletion or change – and as we find information to be not relevant, technology now can automatically release the legal hold.

All of this is easily automated but must start with moving preservation to happen during collection, the very first step. I also believe collection should occur automatically behind the scenes, which should happen during the analyze phase.

As information is analyzed, there are two potential areas where the collection could happen in the background: either right at the start as we “identify” or later when we “analyze.”

Active Learning is a Machine Learning technology that leverages software to automate the identification and analysis portion of the EDRM. In Active Learning, the software identifies potentially relevant content to your matter and leverages human intervention only on items where the software is less sure of relevance.

This Relevance First movement is drastically shifting the conventional approaches to eDiscovery, and with advanced technologies including AI and Machine Learning, it could flip the entire eDiscovery approach on its head.

Improving the processing stage

Through greater consistency in information and a reliance on Artificial Intelligence and Machine Learning, the processing is yet another step of the EDRM that can be shifted in the future for eDiscovery. Some processing will be required in the early stage of identification (spot checking files and quality assurance), while the rest can be withheld until later in the process, during the analyze phase.

It’s clear that the current state of eDiscovery is changing as the old snapshot approach to finding eDiscovery is time consuming and creates potential duplication.

The future state, through the use of active learning, AI, and Machine Learning technology, focuses on a shorter path to identifying, preserving, and analyzing relevant data.