When Two Minutes Changes Everything: A Real-World AI Discovery Demonstration

blank

By John Tredennick

Yesterday, I demonstrated Merlin Alchemy to a litigation partner and head of ediscovery from a major national law firm. What happened next perfectly illustrates why the legal profession stands at a fundamental inflection point—from passive document searching to active intelligence generation through AI collaboration.

The Challenge: Testing the Framework Under Real-World Pressure

For context, the UK Post Office Horizon scandal represents one of the most significant miscarriages of justice in British legal history. Between 1999 and 2015, over 700 sub-postmasters were wrongfully prosecuted based on faulty Horizon accounting software. Many lost their homes, livelihoods, and reputations. Some went to prison. Several died before seeing exoneration.

Thanks to several UK barristers, we had the opportunity to load about 15,000 documents relating to the scandal along with a years-long UK government investigation into all aspects of the scandal from how it happened to its impact on sub-postmasters, their families and the nation’s trust in the government itself.

I offered the firm representatives a choice: pick any one of 54 difficult analytical topics on the UK Post Office Horizon scandal to investigate. They selected Topic 28:

"Examine all evidence regarding changes to Post Office policies, procedures, and governance structures implemented in response to the scandal, including reform initiatives, new oversight mechanisms, and measures to prevent similar injustices."

We discussed how you'd approach this with traditional keyword searches. Where would you even start? "Reform"? "Governance"? "Policy changes"? How many false positives? How many missed documents using synonyms you didn't anticipate?

The Alchemy Process

Instead of struggling further with keywords, we simply pasted the topic into Alchemy.

In less than 2 minutes: Alchemy deployed three AI search methodologies simultaneously—natural language processing, keyword algorithmic search, and CAL classification—to identify the top 300 most relevant documents from the 15,000 documents uploaded into the Alchemy platform.

It then analyzed and summarized each of the retrieved documents based on our topic questions, linking its summary point to the original source text and giving each a title and relevance score.

From there, and within the original 2 minutes, Alchemy began generating a comprehensive 67-page analytical report with direct citations to the source information, which consisted of reports, transcripts and about 15,000 other documents. 

All in all, the process took less than five minutes from the initial topic request to the ultimate streaming of the 67-page report. From there I downloaded the report into Word and sent it to our prospective clients. You can see the actual report I created here: http://bit.ly/47ekhuD

The Alchemy Interactive Framework: From Passive Search to Active Intelligence

Instead of wrestling with Boolean logic, we applied Alchemy's Interactive Framework—a fundamental paradigm shift that moves legal professionals from reviewing search hits to collaborating with AI to generate answers.

Find → Analyze → Answer: A New Discovery Methodology

Step One: Find (Under 2 Minutes)

Alchemy deployed three AI search methodologies simultaneously:

  • Semantic Search (NLP): Understands intent and meaning, not just literal word matching
  • AI-Enhanced Keyword Search: Intelligently expands and maps related concepts, reducing the risk of missing synonyms or contextual variations
  • Machine Learning (CAL heritage): Continuously improves relevance ranking with feedback, surfacing more of what matters

The result: identification of the top 300 most relevant documents from thousands of pages—accomplished without manual query construction, Boolean operators, or iterative keyword testing.

Step Two: Analyze (Minutes, Not Days)

Lightweight AI models (Haiku 3.5, GPT-4o Mini) performed rapid first-pass review, reading and comprehending document content, extracting key facts and relationships, generating context-aware summaries, and scoring relevance. This automated analysis—work that traditionally required armies of reviewers over days or weeks—was completed in minutes while maintaining consistency and comprehensive coverage.

Step Three: Answer (Comprehensive Synthesis)

Advanced large language models synthesized findings across the entire document collection, generating a 67-page analytical report with full source citations.

What the Report Revealed: Evidence of Systematic Intelligence

The system identified and synthesized 15 distinct reform categories, including:

  • Board Governance Transformation: Appointment of two serving Postmaster NEDs in June 2021, with succession planning underway; skills matrix completed June 2024 identifying gaps in technology and transformation expertise
  • Risk Management Overhaul: Establishment of Group Assurance function; appointment of Group Risk and Assurance Director (June 2024); planned root-and-branch review of control environment
  • Whistleblowing Infrastructure: Creation of NED Whistleblowing Champion role, dedicated Whistleblowing Manager, centralized investigations unit (CIU), with policy maturity reaching 80% by end-2021 per external assessment
     

Each finding included specific document citations, dates, and cross-references to related governance changes.

════════════════════════
⏱️ Total Time: Under 5 minutes
📄 Output: 67-page analytical report
🎯 Coverage: 15 reform categories
🔗 Citations: Full source traceability
════════════════════════

This Isn't a Summary—It's Synthesis: Understanding the Distinction

What Alchemy delivered differs fundamentally from traditional document review outputs. Traditional approaches produce lists of potentially relevant documents and individual summaries requiring manual compilation. Alchemy delivers cross-document pattern recognition, temporal analysis tracking implementation timelines, identification of relationships between initiatives, and coherent narratives with maintained source traceability.

This represents synthesis—connecting dots across hundreds of documents to create actionable intelligence while preserving the evidentiary foundation required for legal defensibility.

Why This Represents Workflow Transformation, Not Mere Acceleration

The efficiency gains are dramatic, but the strategic implications extend further.

Analytical Capabilities Enhanced
  • Comprehensiveness: Analysis across entire document sets rather than sampling or targeted review
  • Consistency: Uniform analytical standards applied across all materials
  • Depth: Multi-dimensional analysis (temporal, relational, thematic) performed simultaneously
  • Iteration: Rapid refinement through follow-up questions exploring new angles

Professional Practice Elevated

Critical Distinction: This technology amplifies legal judgment rather than replacing it.

Attorneys must still evaluate findings for legal significance, apply substantive legal reasoning, develop case strategy and theory, exercise professional judgment on tactics and ethics, and counsel clients on risks and opportunities.

The transformation: Instead of spending days merely locating and reading potentially relevant documents, legal professionals can dedicate their expertise to higher-value analytical work—the tasks that truly require human judgment, creativity, and strategic thinking.

Proven Across High-Stakes Scenarios

We've demonstrated this methodology across diverse matters: comprehensive 60-page closing argument for BP in Deepwater Horizon trial (90 minutes), detailed medical records chronology analysis (seconds to minutes vs. weeks), regulatory investigations processing 100,000+ documents (weeks instead of months), and rapid internal investigations identifying compliance failures across disparate sources.

The Competitive Divide: Why Timing Matters

The legal profession is experiencing a fundamental transformation—not in some distant future, but now. Document intelligence platforms like Alchemy aren't emerging technologies; they're operational realities already deployed by forward-thinking organizations.

For Law Firms: Significantly reduced discovery costs enable more competitive fee structures. Accelerated timelines improve client satisfaction. Enhanced analytical depth strengthens case outcomes. Differentiated service offerings create market position in increasingly competitive landscapes.

For Corporate Legal Departments: Dramatic efficiency gains in internal investigations and compliance matters. Better visibility into legal risks across document repositories. Faster, more comprehensive responses to regulatory inquiries. Improved cost management while maintaining quality standards.

The Critical Choice: Early adopters gain compounding advantages—not just operational efficiency, but market positioning, client confidence, and the organizational learning required to maximize AI-augmented workflows. Those who delay adoption face a widening gap as competitors deliver faster, more comprehensive, and more cost-effective services.

Built for Modern Requirements

The Alchemy framework operates on infrastructure purpose-built for contemporary demands: single-tenant security architecture eliminating data commingling, cloud utility pricing (pay-by-the-minute with 70% reduction when idle, cutting hosting costs up to 60%), and sustainable computing practices reducing energy consumption while supporting corporate ESG commitments.

From Information to Intelligence: The Train is Leaving the Station

The legal profession is transforming before our eyes. Document intelligence platforms like Alchemy aren't coming—they're here. The question isn't whether to adopt these tools, but how quickly you can integrate them into your practice.

The demonstration illustrated a fundamental shift: from searching for needles in haystacks to receiving detailed, defensible answers backed by comprehensive analysis of the full documentary record.

This isn't incremental improvement—it's a workflow transformation that preserves and elevates professional judgment while dramatically accelerating the analytical foundation upon which that judgment operates.

The firms and legal departments that embrace this transformation now will have a significant competitive advantage. Those that wait will find themselves struggling to compete on both efficiency and cost-effectiveness.

The train is leaving the station. The question isn't whether AI will transform legal practice—it's whether you'll be positioned to capitalize on that transformation or left struggling to catch up.

Want to see the full 67-page Post Office analysis? Download the complete report here: http://bit.ly/47ekhuD 

Ready to explore how this applies to your practice? Contact me directly or click here to schedule a demonstration using your own challenging scenarios.

blank

About the Author

John Tredennick (jt@merlin.tech) is the CEO and Founder of Merlin Search  Technologies, a software company leveraging generative AI and cloud technologies to make investigation and discovery workflow faster, easier, and less expensive. Prior to founding Merlin, Tredennick had a distinguished career as a trial lawyer and litigation partner at a national law firm. 

With his expertise in legal technology, he founded Catalyst in 2000, an international ediscovery technology company that was acquired in 2019 by a large public company. Tredennick regularly speaks and writes on legal technology and AI topics, and has authored eight books and dozens of articles. He has also served as Chair of the ABA's Law Practice Management Section.

Scroll to Top