The AI Crime Analytics Agent is an AI-powered assistant designed to help a large metropolitan police department quickly query data and generate actionable insights. The solution enables identification of anomalies, trends, and surges in criminal activity, supporting proactive crime prevention and strategic decision-making.

By replacing manual, siloed processes with natural language queries and automated reporting, the AI Crime Analytics Agent reduces analysis time from 90–120 minutes to just 18 seconds per query—a 350x efficiency improvement.

The police department faced significant challenges:

• 17+ siloed systems with conflicting data sources.
• Manual, keyword-based searches taking 90–120 minutes for trend analysis.
• Tactical, reactive approach to crime analysis.
• Limited ability to identify patterns and anomalies quickly.

The goal was to create a centralized, intuitive, and AI-driven solution that:

• Accelerates information retrieval.
• Enables proactive crime prevention.
• Improves productivity and resource allocation.

The AI Crime Analytics Agent leverages IBM watsonx Orchestrate and Db2 to:

• Aggregate diverse data sources (Places, Vehicles, Events) into a unified view.
• Enable natural language to SQL queries for intuitive data access.
• Detect anomalies, identify cross-system linkages, and analyze temporal trends.
• Generate visual reports (tables, bar charts, line graphs, pie charts) for actionable insights.

1. Discovery & Workshop

• Led a scope-defining workshop with police department stakeholders to align on goals and success criteria.
• Facilitated exercises to understand user needs and define MVP scope.
  
  

2. Design Activities

• Created architecture diagrams to illustrate system components and data flow.
• Designed user journeys in Figma to map investigator workflows and ensure intuitive UX.
• Developed personas (e.g., Detective Jordan) to guide design decisions and demo storytelling.
  
  

3. Build

• Stored police department data in IBM Db2 for querying.
• Built AI agents:
• Detective Agent for natural language to SQL queries.
• Anomaly Detection Agent for pattern identification.
• Reporting Agent for visualizations.
• Iteratively refined configurations based on feedback from working sessions.
  
  

4. Demo

• Produced a commercialized demo using the Adobe Suite, showcasing real-world investigative scenarios.
• Scripted a compelling walkthrough to highlight the AI Crime Analytics Agent’s impact on crime analysis.

• Reduced trend analysis time from 105 minutes to 18 seconds (350x faster).
• Enabled law enforcement to shift from reactive to proactive strategies.
• Improved productivity, allowing analysts to focus on higher-priority tasks.
• Positioned the department as a leader in AI-driven public safety innovation.