AI-Assisted Analysis: Select Case Examples
Client support in early-stage agency investigation
In support of a client undergoing enforcement agency investigation, our litigation consultants used AI tools to enhance their analyses. Our professionals:
- Applied LLMs for text classification and information extraction to identify at-issue activities among millions of free-form notes from a customer and sales database
- Applied large-scale record linkage to collapse nearly three million unique payer records into homogenous clusters, each representing a single individual
LLM-assisted price fixing analysis
With multiple data scientists embedded into the case team, Cornerstone Research supported counsel in an antitrust matter alleging price fixing. Our professionals navigated a complex data environment comprising over one hundred databases from separate parties. We assisted the team in adopting AI-driven technologies for coding assistance and document review by utilizing multiple tools.
We also used AI tools to assist with review and triage of information across thousands of documents spanning emails, presentations, user guides, and technical specifications. The tool efficiently identified likely matches for complex queries—which substantially reduced manual review time.
Further, we developed custom cloud infrastructure to scale specialized regression models. This capability enabled the performance of many regression sensitivities, which was critical to informing early case strategy and analysis direction under a sudden, short deadline.
AI-assisted coding in monopolization litigation
In an antitrust matter with allegations of monopolization, Cornerstone Research employed AI coding assistance. Using an AI tool that adheres to the firm’s high standards of both quality and data security, our professionals instructed the tool to:
- Clean and reorganize code for a complicated and unusual code production process
- Generate exploratory code for our team to test different approaches and methodologies
- Review and analyze the opposing expert’s code
By using AI to enhance these analyses, our team was able to test more code, review more outputs, and explore more arguments in support of litigation. To test and confirm the conclusions of the AI-assisted outputs, all analyses were replicated by one of our professionals.