eDiscovery & legal technology

Accelerating defensible discovery with GenAI and expert workflows

Accelerating defensible discovery with GenAI and expert workflows

A large Australian corporation approached KordaMentha following a document production request linked to historical conduct within an acquired business. The company was involved in proceedings initiated by a national authority, yet had limited visibility into the underlying facts.

2 min read
Overview How we helped Outcomes acheived

A large Australian corporation approached KordaMentha following a document production request linked to historical conduct within an acquired business.

The company was involved in proceedings initiated by a national authority, yet had limited visibility into the underlying facts.

Legal and regulatory leaders needed to discharge discovery obligations and advise stakeholders on unfamiliar subject matter.

KordaMentha was engaged to help the client understand key issues, reduce risk and meet production deadlines.

How we helped

Given the client’s limited insight into the historical conduct, we recommended an early case assessment to identify key themes and reduce the document set. Using analytics, we narrowed the initial pool of over one million files to fewer than 50,000 documents. We then developed GenAI prompts to interrogate the data and surface relevant individuals, behaviours, timeframes and communications.

With this intelligence, we refined the AI prompts to classify documents against discoverable issues. Once validated, the review set was reduced to a manageable size, enabling internal counsel to complete the review without compromising timelines.

To ensure defensibility and completeness, we deployed tailored workflows that enhanced the reliability and transparency of GenAI outputs:

  • Cross-checking AI results against reviewer coding
  • Combining GenAI with traditional technology-assisted review
  • Using trained machine learning models to validate unreviewed sets
  • Presenting AI validation in line with accepted eDiscovery standards

Outcomes acheived

  • Rapid document reduction: Over one million files narrowed to under 50,000
  • Targeted issue identification: Key individuals and behaviours surfaced within hours
  • Internal resourcing enabled: Final review completed by two internal counsel
  • Review completed in one week: No compromise on quality or defensibility
  • Cost savings realised: External legal advice required only in limited cases
  • Improved stakeholder visibility: Enhanced understanding of historical conduct
  • Validated AI outputs: Consistent with accepted eDiscovery standards
  • Integrated workflows: GenAI combined with proven review methodologies