READ+

READ+ Team members and divisions

Problem Statement

Public officers spend significant time on sensitive-high documents which involve repeated and tedious manual steps to extract the required information. Due to the sensitive nature of the data being worked with, we need to develop the solutions in-house. Currently, READ is offered as an in-house built service that automates the manual workload on processing documents for officers, and is able to process documents like Passport, Practicing Certificate and Transcripts for foreign nurses. As part of the expansion process, in this hackathon we propose READ+ to expand this service to more agencies, document types and use cases.

Problem Formulation Process

Key Insights Gathered from the business owner Interview
We engaged with various Agencies to understand their requirements, including DSTA, CPF, MSF, MOM. From these, we Identified overlapping document requirements:

• Marriage certificates for MOM's EP/SP/WP
• Marriage certificates for ICA's LTVP/PR/SC applications
• Receipt for DSTA applications

Based on the insights gathered, we are confident that we need to expand READ to be more general in order to tackle the new use cases.

Proposed Solution Details

Vision for READ+ We envision READ+ end goal to be an in-house built solution, developed entirely within the team. It must be designed to be generic so it can serve multiple agencies’ use-cases. The insights drawn from the documents must be specific and useful to each agencies.

Ideation

• Continue Stakeholder Conversations: Engage with various stakeholders from multiple agencies
• Build PoCs for New Customers: Target customers such as ICA and DSTA with confirmed interests
• Design Web Front-End for WOG: Create a web interface for quick evaluation
• Develop and Deploy New ML Model: Focus on key information extraction for general documents.

Impact and Outcome Analysis

Future Plans