Think LP

The Project

Crime Link

Overview

Improving partnered with Think LP to enhance their Crime Link application, a platform designed to track and aggregate criminal activities across various retail chains. The primary aim was to streamline data processing, improve the system's efficiency, and enable better identification of repeat offenders. Using AWS-managed services and advanced AI techniques, we re-engineered the existing solution to ensure that Think LP could better serve their clients, which include major retail chains across North America.

The Challenge

Think LP faced significant challenges in tracking and aggregating criminal activities across multiple retail chains due to inconsistent reporting and an over-engineered, error-prone system. Individual thefts often fell below prosecution thresholds, allowing repeat offenders to evade justice by targeting different stores. The existing system was slow, complex, and costly—misusing resources and failing to produce timely, actionable insights.

The Solution

Our approach involved a comprehensive overhaul of the Crime Link application. We simplified the system architecture and transitioned from a Kubernetes cluster to AWS-managed services, reducing complexity and cost. We implemented AI to extract structured data from unstructured reports, enhancing data consistency and comparability. By adopting AWS services and Infrastructure as Code (IaC) practices, we significantly improved the solution's efficiency, reliability, and scalability.

Business Benefits

• Improved Efficiency: Data processing times reduced from days to hours, handling large data sets efficiently.

• Cost Reduction: Monthly database costs decreased by nearly 90%, switching from CockroachDB to Aurora DB.

• Enhanced Accuracy: AI-driven data extraction improved the consistency and reliability of crime reports.

• Simplified Maintenance: Managed services reduced the operational burden, allowing the small development team to focus on core functionalities.

• Scalability: The new architecture easily scaled to meet growing data and client demands.

• Rapid Deployment: Infrastructure as Code practices enabled quicker deployments and easier updates.

Technologies And Methodologies Used

• AWS Managed Services: Streamlined system operations and reduced complexity.

• Amazon Bedrock for Data Extraction: Converted unstructured text into structured data.

• AWS CDK: Implemented Infrastructure as Code for efficient resource management.

• Lambda Functions: Improved operational efficiency and reduced code complexity.

• Aurora DB: Provided a cost-effective, scalable database solution.

• DevOps Practices: Ensured smooth integration and deployment processes

Partnerships

The success of this project was built on strong collaboration between Think LP’s technical team, Improving’s experts, and AWS. Think LP’s CTO, Anu, provided crucial insights into business requirements, while our team led the architectural redesign and implementation. AWS's ecosystem facilitated the transition to managed services, ensuring a robust and scalable infrastructure. Minimal third-party involvement ensured focused and streamlined development.

Lessons Learned

1. Simplicity Over Complexity: Simplifying the architecture reduced errors and maintenance overhead.

2. Managed Services Advantage: Leveraging AWS services minimized the need for in-house infrastructure management.

3. Consistent Data Handling: AI-based data extraction improved the consistency and usability of crime reports.

4. Cost Efficiency: Thoughtful resource selection, such as switching to Aurora DB, significantly reduced operating costs.

5. Client Collaboration: Regular communication with the client ensured alignment and timely resolution of unforeseen challenges.

6. Modular Design: While microservices provide benefits, starting with a more monolithic approach can simplify initial deployment and can be further modularized as needed.

Conclusion

Think LP’s Crime Link project exemplifies Improving’s ability to deliver innovative, cost-effective, and efficient solutions. By leveraging AWS-managed services and advanced AI, we transformed a complex, error-prone system into a streamlined, scalable, and reliable application. Our approach not only reduced operational costs but also significantly improved data processing capabilities, enabling Think LP to deliver better insights and outcomes for their clients. This case study highlights our commitment to excellence and our capacity to handle complex technical challenges, delivering tangible business benefits.