The Promise of GenAI in Litigation: From Skepticism to Reality
When discussions turn to Generative AI’s impact on litigation, skeptics often dismiss it as more hype than substance. But sometimes, real-world examples speak louder than theoretical debates.
Friday Night Lights
Consider this example from this Friday night (12/13/2024): Our client is a large, national law firm. They were taking over an important matter from another firm. The discovery materials included 102 deposition transcripts, totaling 18,294 pages of testimony that needed to be summarized and analyzed. Traditionally, this would have been a weeks-long project, consuming countless attorney and paralegal hours, with costs easily running into six figures.
The client sent them to Merlin on Friday afternoon, with a request that we summarize the transcripts as quickly as possible. We offer this as a side service in conjunction with the Merlin Discovery Platform and were happy to provide the service. We were asked to finish work by the end of next week but we didn’t see any reason to wait that long.
Deposition Summaries to Go
Instead, using advanced GenAI technology, these depositions were transformed into comprehensive, actionable summaries in just 131 minutes. That’s not a typo — just over two hours to accomplish what would typically take weeks of dedicated human effort.
Each summary wasn’t just a basic outline. They included detailed testimony abstracts, organized key sections with explicit statements and references, and most importantly, linked page-line citations that allow attorneys to instantly jump to critical testimony in the original transcripts.
We delivered the summaries, each with links to the underlying transcript testimony, in both PDF and Word formats, making them immediately useful for case preparation and strategy development. Notwithstanding the deadline seven days later, we delivered them that same night. And no human did any work on the summaries other than to send them to our automated transcript analysis software.
How did we accomplish this?
We developed serverless computing software that runs in AWS. Once we receive transcripts from our partners, we run the software in batch mode. With the vast computing power offered by AWS, we could send each of the transcripts up to their serverless computing resources to run in parallel. The software ran the software we created, sending the text up to Anthropic’s Claude Sonnet 3.5.
Each transcript was divided into sections for analysis and summarization. Our software then took the output to create the summaries in both PDF and Word format. Each summary included a copy of the underlying transcript testimony with page and line number links along with a transcript overview, section overview and a topical table of contents.
Want to see what one looks like?
For obvious reasons, we can’t provide copies of the transcripts we received. However, you can take a look at this summary of Joe Nadeau’s deposition which the EPA posted on its website right here: https://drive.google.com/file/d/1zhGceOWQPeRIjBwGGH4Q2d2uXSxWHX8Z/view
We regularly include an overall summary that looks like this:
After providing a linked, topical table of contents, we provide section summaries with individual statements with pages and line numbers. Like this:
The blue links take you directly to the actual testimony.
What does this mean for the skeptics?
The implications are profound:
- Time Efficiency: What once took weeks now takes hours, allowing legal teams to focus on strategy rather than administrative tasks.
- Cost Reduction: With summaries costing as little as $1.00 per page, clients benefit from dramatic cost savings compared to traditional manual summarization.
- Consistency: The AI-driven process ensures uniform quality across all summaries, eliminating the variability that comes with multiple human reviewers.
For those still skeptical about GenAI’s impact on litigation, these numbers tell a compelling story: 18,294 pages, 102 transcripts, 131 minutes. This isn’t theoretical — it’s a real-world demonstration of how technology is revolutionizing legal practice, one deposition at a time.
This is just the beginning. As these tools continue to evolve, they’re not just making existing processes more efficient — they’re enabling entirely new approaches to case preparation and strategy development. For forward-thinking legal professionals, the message is clear: the future of litigation is here, and it’s being shaped by GenAI.
