Generative AI is easily the hottest topic in eDiscovery in years. Given its potential to impact the legal industry, lawyers and legal professionals are particularly keen on trying to see what lies ahead. At Venio, we have always been committed to leading edge technology and pride ourselves on thought leadership. Part of this leadership is listening to clients, potential clients, and our peers in the industry. (And, yes, that includes our competition.)

Based on these conversations, we find that the following five questions are at once most asked and least understood.

Where will GenAI be used?

We see GenAI having the biggest impact on processing, culling, and analysis with large effects on review.

Thinking about the eDiscovery process, the simple goal is to produce the minimum number of documents that are relevant, responsive, and not privileged. Look no further than the claims of the software and service provider that promise an 80%, 90%, or more reduction in the number of documents you must review.

Consequently, by the time documents are sent to review – more on that below – the document set should be quite small. GenAI is a Large Language Model. Emphasis on the “large.” Our view is that GenAI will have the biggest impact analyzing the large data sets present before or at the start of the discovery process and that strategic use of GenAI will drastically reduce the workload on the right-hand side.

How will it be used?

We see GenAI taking over the functions of ECA, but as a background technology.

Properly trained GenAI will be able to take a complaint and/or request for documents and quickly give a natural language answer to questions posed by the legal team based on the collected data set. The key will be the design of the GenAI tool used and the process used by the software provider. Providers cannot simply offer raw GenAI capabilities; the tool will have to be seamlessly integrated into the software process to make it easy to use.

This gives software providers the flexibility to integrate in a manner that best takes advantage of the platform’s strengths and will offer the industry choices to best fit their needs.

What about human review?

Reports of the death of human review have been greatly exaggerated, but we see the combination of GenAI and advance searching eliminating that vast majority of human review.

We have used managed review for so long, we forget why. The reasons are to confirm that documents are relevant, responsive, and not privileged. With GenAI and other AGI technologies, the only thing that is keeping human reviewers employed is a lack of trust. GenAI has been linked to some serious errors (called “hallucinations”), but with the right training set and properly phrased queries, GenAI can eliminate first pass review.

Of course, it may be against the interest of review platform providers to develop reporting to prove AI review is as good or better than human review. But for client corporations interested in attacking the cost of eDiscovery, this remains the easiest way to do that.

Which tool will be the standard?

This is very much a “that depends” answer.

If you are in a cloud environment with software built in that environment, then you will be using that cloud’s tools. If you are in AWS, then you are going to be using AWS’ GenAI tool. In an Azure cloud? Then ChatGPT will be your platform. If you are using on-prem, then you will need to ask pointed questions about what GenAI your software is integrating and what is the roadmap for GenAI in the platform.

Cost is the second element. Cloud based AI services will be an additional cost to the provider – who will pass that along directly or indirectly. For on-prem software, providers will consider it an add-on or a core component and price accordingly. The bottom line is that someone will be footing the bill and, consequently, clients need to ask the hard questions about value and ROI.

Will GenAI affect the EDRM?

We see the EDRM evolving into intersecting cycles. We call this the GRCe (Governance, Risk, Compliance, eDiscovery, pronounced GRACE) cycle.

The adoption of the governance model on the left of the EDRM and connecting it to the classic eDiscovery model was a step in recognizing that eDiscovery is part of the larger GRC challenge faced by organizations. eDiscovery is not the end of the process, but a part of a cycle to manage risk – and the most expensive part of that cycle. GRCe best practices promote the transfer of discovery back into the organization to reduce similar risk from recurring in the future.

Final Thoughts

We have been around technology predictions long enough to be very humble about our prognostication skills. Adoption in legal services seems to take far longer than in other corporate functions, and so it may be with GenAI as well.

This feels a bit different though. Different in the sense that adopters who become proficient in its use may create an unfair competitive and financial advantage. Competitive in the sense that their work product will be more accurate, produced faster, and at lower cost to the client. Financial in the sense that GenAI promises to be a force multiplying technology allowing legal departments and legal service providers to expand their strategic reach at a significantly lower operational cost than the competition.

What Do You Think?

We at Venio have further thoughts about the implications of GenAI and want to hear from you as well. We would welcome the opportunity for a 30-minute talk with you and your team about the above or on other issues. Please feel free to reach out to us via email to michael.andris@veniosyatems.com.