Robert Keeling Shares Insights in the eDiscovery Today “2026 State of the Industry Report”

In the “2026 State of the Industry Report” published by eDiscovery Today, Robert Keeling shares insights on key trends shaping the eDiscovery landscape. 

The report draws on responses from a twelve-question survey completed by 559 professionals from law firms, service providers, corporations, consultancies, and government organizations, along with analysis from more than sixty industry experts.

Below are Robert’s perspectives on several important questions facing the eDiscovery community, as highlighted in the report. 

…

What do you believe will be the impact of large language model (LLM) and generative AI technologies on eDiscovery solution offerings and workflows in 2026?

Discovery workflows using GenAI will gain substantial “market share” in 2026.  We will see an increase in matters using GenAI across all major workflows – including responsiveness review, privilege review, key document identification, and privilege logging.

Which use case for large language model (LLM)/generative AI technologies do you think is not being discussed enough and why?

Privilege logging.  The quality of GenAI logs have increased substantially over the past few years.  The quality is now comparable to contract attorney review – but much faster and arguably cheaper.

In your opinion, what is the single biggest roadblock to adoption of large language model (LLM) and generative AI technologies in eDiscovery?

The single biggest roadblock is the current lack of a court decision similar to Judge Peck's Rio Tinto opinion that helped legitimize TAR.  There is nothing inherently different between the use of TAR and GenAI, and both should receive court approval.  But the lack of an explicit decision endorsing GenAI review inhibits adoption.

Given the focus on large language model (LLM)/generative AI technologies by eDiscovery providers, do you think organizations are already phasing out predictive coding technologies and approaches?  Why or why not?

No. TAR still has a place in discovery.  In addition, we are seeing very good results combining TAR with GenAI to achieve even greater efficiencies depending on the matter.

In your opinion, what is the single biggest roadblock to discovery of mobile devices in litigation and investigations, given the ubiquity of use of those devices by potential custodians?

The single biggest roadblock to the discovery of mobile devices is the lack of a collection technology that can seamlessly (and with minimal custodial involvement) perform searches on the device itself and collect just the responsive messages while minimizing privacy and burden concerns.

What do you think the landscape of chat and collaboration app discovery will be in five years in terms of adoption, technologies and approaches?

The industry would benefit from increased investment to the back-end legal hold and discovery capabilities of chat and collaboration apps to help reduce discovery burden and cost.

What do you think will be the biggest business-related trend in eDiscovery in 2026?

Costs associated with processing and reviewing AI-created summaries will significantly increase despite the minimal relevance of such summaries.

Given recent case law rulings, do you think courts will expect parties to identify solutions to produce contemporaneous versions of hyperlinked files in the near future?  Why or why not?

Much of the current caselaw is actually getting hyperlinked files correct – namely, most of the hyperlinked files are not relevant, and producing parties should have to recover only a limited number (e.g., 100-200) identified by the receiving party.

What do you think will be the biggest eDiscovery related trend that we will see organizations address in 2026 and why?

AI summaries are expensive to include in discovery and largely irrelevant and should be subject to corporate roll-off policies.

Other than lack of eDiscovery competence, what do you think is the eDiscovery challenge that not enough people in the industry are talking about and why don’t you think more people talk about it?

I’m going to fight the question and argue that the biggest eDiscovery challenge today centers on competence—not in the traditional sense, but in navigating the complexity of emerging AI issues, including AI-generated data.  These topics are so intricate that even seasoned eDiscovery professionals often struggle to grasp all the nuances and risks.  As a result, defining where the competence line should fall for general litigators on these matters will pose a significant challenge.

More information about this report can be found here.

Observations expressed in this report by thought leaders are their own and do not represent the views of their employers, clients, or partners.