Case Study
The Hawaii State Judiciary’s Criminal Justice Research Institute (CJRI) partnered with Cloudwick to implement the Amorphic Intelligent Document Processing (IDP) Solution, harnessing the power of generative AI to transform case document management and analysis. With critical information embedded in unstructured text fields within its case management system, CJRI deployed Amorphic IDP to automate the extraction, structuring, and analysis of specific data points. The solution enabled CJRI to efficiently process complaint documents, providing essential insights and empowering faster, data-informed decisions in judiciary processes.
CJRI’s objective was to unlock vital information hidden within freeform text fields in its case management system. While structured data highlighted general case details, specifics—such as the exact type and quantity of substances involved in drug-related offenses—were only available in unstructured text. Cloudwick’s Amorphic IDP Solution provided the capability to automatically extract, organize, and make this information accessible for analysis, supporting CJRI’s mission to drive informed judicial decisions and policy recommendation.
Critical case details, like specific drug types and quantities, were locked within unstructured text fields in CJRI’s case management system, making it challenging to retrieve and analyze these insights effectively.
Variations in documentation and terminology complicated data sharing across departments. Amorphic enabled CJRI to standardize and aggregate data across disparate sources, making comprehensive trend analyzes possible.
Generative AI was essential in identifying and structuring complex data points buried within lengthy, unstructured text, such as offense codes, substance amounts, and contextual details around each case.
With the LLM’s advanced contextual capabilities, CJRI could overcome limitations of traditional search methods, retrieving data accurately and enabling faster, more reliable analysis.
Using Anthropic Claude 3.5 Sonnet from Amazon Bedrock, Amorphic IDP processed the complaint documents related to marijuana offenses, leveraging generative AI to interpret, extract, and structure complex information.
Tailored prompts guided the model to identify key fields accurately, such as specific drug types, quantities, and relevant legal codes, enabling CJRI to retrieve essential data quickly and reliably.
Manual review of a subset demonstrated a 98% accuracy rate, successfully showcasing the reliability of the LLM in transforming unstructured text into structured data.
The structured exports enabled CJRI to analyze data trends within an interactive dashboard, complete with visual summaries and downloadable reports.
Future processing costs were estimated at $1 per 20 documents — demonstrating the scalability and cost-effectiveness of the Amorphic platform.
The Amorphic IDP Solution addressed CJRI’s challenges by transforming unstructured data into structured, actionable insights:
Within Amorphic, CJRI utilized a dedicated document store to experiment with prompts, refining data extraction accuracy and adapting to new document types.
The Amorphic IDP transformed complex text fields into structured, queryable data, thanks to the nuanced interpretation capabilities of generative AI.
Processed data was organized into relational tables, easily searchable and reviewable, with each document’s results accessible for individual or bulk analysis.
The bulk-exported data was visualized in dashboards that highlighted key metrics like offense timelines, distribution trends, and possession quantities, allowing CJRI to gain valuable insights from unstructured data.
Customized reports provided CJRI with accessible summaries, supporting informed discussions and data-driven policy analysis.
Cloudwick’s Amorphic IDP Solution, driven by generative AI, delivered transformative improvements in data accessibility and efficiency for CJRI.
A By transforming unstructured freeform text into structured data, Amorphic IDP gave CJRI fast, accurate access to essential case details, supporting more timely judicial insights.
A 98% accuracy rate in data validation demonstrated the precision and reliability of the extraction process, allowing CJRI to confidently use the data in decision-making.
Automation minimized the need for manual data handling, allowing CJRI to focus on analysis and strategic decision-making.
The solution’s affordable processing costs highlighted its scalability, ensuring CJRI could continue to analyze documents within budget constraints.
Interactive dashboards and downloadable reports provided CJRI with the ability to visualize trends and patterns, supporting data-driven discussions and enhancing the judiciary’s operational efficiency.
The Amorphic solution is deeply integrated with the AWS ecosystem, leveraging over 80 native AWS services. Some of the key AWS services used include: