Unleashing compassion: AWS for child protection & data governance (IMP210)
Transitioning to a Data-Driven Organization: Compassion International's Journey
Establishing a Platform Mindset
Compassion International, a 70+ year old organization, had a lot of data across various systems that were not compatible with a modern data and analytics approach.
To address this, they focused on transitioning to a "platform thinking" mindset:
Established a solid technical foundation ("the blue part of the pyramid") to support various user needs.
Implemented a "program cycle adoption" process to collect, analyze, and provide data-driven insights to the frontline church staff.
Benefited from common branding, user experience, training, and data governance across the global organization.
Implementing a Data Fabric
Compassion adopted a "data fabric" architectural pattern to enable consistent access to data across the enterprise.
This included an abstraction layer with a common data model and "compassion outcomes framework" to understand relationships between data.
A data catalog allowed people to discover and understand available data.
The storage layer enabled access to data at the required latency.
They decided to focus on integrating new systems into the data fabric, while addressing legacy systems on a case-by-case basis.
Empowering People through Experimentation
Compassion used "Ignite" events and AI hackathons to encourage people to experiment and learn.
These events allowed teams to work on projects based on their passions and ideas, within the broader objectives of the organization.
Examples included using generative AI for grant request communications, a "prayer partner" app, source code improvement, and insights from children's letters.
This hands-on, project-based learning approach helped overcome fear and enabled experts to work alongside frontline staff.
Protecting Children with Data
Compassion developed the "PATCH" project to screen supporters' addresses against known offender lists, using a complex but robust system:
Centralized address data in Amazon S3, using AWS Lake Formation to control access.
Leveraged AWS Step Functions, Lambda, and DynamoDB to perform fast, multi-layered address screening.
Captured both incoming requests and responses to enable continuous improvement and analytics.
Maintained a human-in-the-loop approach to validate potential matches.
For content screening (e.g., letters, images, videos), they followed a similar pattern:
Stored content in S3, performed pre-processing, and passed references through the system.
Used Amazon Comprehend to analyze content for issues like inappropriate language, grooming, or out-of-channel communication.
Kept the processed content in S3 to enable further analysis and innovation.
Lessons for Other Organizations
Data Lake and Permissions: Use AWS Lake Formation to simplify data access, enable cross-account sharing, and provide centralized auditing and governance.
Intelligent Document Processing: Leverage pre-built machine learning services like Amazon Comprehend to automate document processing tasks.
These cookies are used to collect information about how you interact with this website and allow us to remember you. We use this information to improve and customize your browsing experience, as well as for analytics.
If you decline, your information won’t be tracked when you visit this website. A single cookie will be used in your browser to remember your preference.