Public Sector – AWS Innovation Islands

Public Comment Analysis

The USDA faces challenges in efficiently managing and analyzing public comments on regulatory proposals. With thousands of submissions on Regulations.gov, manually categorizing and summarizing these comments is time-consuming and labor-intensive. The ASU AI Cloud Innovation Center built an automated system to conduct topic modeling, sentiment analysis, and AI-generated content detection while ensuring accurate clustering and summarization of public feedback.

Key features:

Automated Topic Modeling – Categorizes comments for quick analysis.
Sentiment Analysis – Assigns overall sentiment to comment groups.
AI-Generated Recommendations – Suggests data-driven actions.
AWS-Powered Backend – Step Functions, Lambda, Amazon SageMaker, Bedrock (Sonnet 3.5), and API Gateway for real-time insights.
User-Friendly UI – Built with React and Material UI, deployed via AWS Amplify.

Learn more and access the solution on our blog.

Assessor Assistant

Understanding property tax regulations in California is complicated and confusing for most homeowners. This conversational AI assistant uses natural language processing and semantic search to provide accurate, cited answers from sources like the California Assessor’s Handbook, Tax Codes, and Letters to Assessor documents. The chatbot also features an intelligent exemption form finder that guides citizens through tailored yes/no questions, recommending appropriate exemption forms.

AWS services utilized:

Amazon Bedrock
Amazon Titan Text v2 Embedding
Anthropic Claude Sonnet
AWS Lambda
Amazon OpenSearch Serverless
Amazon Textract

This solution significantly enhances citizen interactions, reduces assessor workloads, and streamlines training, transforming regulatory navigation.

GenomicsMapper

In today’s rapidly evolving public health landscape, standardizing genomic data for federal submissions poses a significant challenge for laboratories. GenomicsMapper addresses this by leveraging generative AI to automate the translation of laboratory-specific genomic terminologies into standardized formats required by national repositories.

The innovative solution uses natural language processing to match diverse local terminologies with standardized NCBI BioSample definitions, significantly reducing manual processing time and improving submission accuracy.

AWS services used:

Amazon Bedrock
AWS Lambda
Amazon API Gateway
Amazon S3

Innovation Sandbox on AWS

Innovation Sandbox on AWS, a new solution addressing the global shortage of cloud skills and the need for experimentation and innovation. The solution provides self-managed sandbox environments through a web-based portal, optional approval flows, cost and duration control via lease templates, and automated account cleanup and recycling.

The solution allows for customizable budget limits, duration settings, and notification thresholds.

Technology:
AWS, AWS Solution Library, CloudFormation, AWS Organizations, AWS STS, AWS CLI

More information:

Landing page
Implementation guide
Workshop (currently only works with customer provisioned accounts):

Procurement Scope Builder

In today’s complex procurement landscape, creating comprehensive and accurate statements of work scopes can present a challenge for university procurement teams. Scope Builder uses generative AI to assist procurement specialists during the work scope development process, helping them produce detailed and comprehensive documents. Through a conversational AI interface, the system guides users through the process of creating work scopes, drawing from established best industry practices and relevant past contracts.

AWS services used:

AWS Lambda
Amazon API Gateway
Amazon Cognito
Amazon S3
Amazon Bedrock

Enrollment Prediction

Higley Unified School District serves approximately 13,500 students across 16 schools. Predicting school enrollment is crucial.

According to local school administrators, simply getting static predictions can cost as much as $20,000, an amount that can be a significant burden for the school districts and does not have the flexibility to accommodate changing circumstances.

The ASU Cloud Innovation Center created an open-source prototype to predict future enrollment and integrate real time data, such as attendance data to assess enrollment predictions. This solution uses a wide pool of data from various sources and provides analytics with an easy to use dashboard.

Open-source code: Download

PDF Accessibility

Many organizations have document collections containing hundreds of thousands of PDF documents, many of which do not meet the Web Content Accessibility Guidelines (WCAG) 2.1 Level AA standards, making it difficult or impossible for individuals relying on assistive technologies to access those documents. To address this issue, the ASU Cloud Innovation Center developed an innovative, artificial intelligence-driven solution designed to remediate documents. Some readily available remediation solutions cost $3-$15 dollars per page, but with this solution, expenses for AWS services are less than a penny per page.

AWS services used:

Amazon S3: Used to securely store and manage the documents being remediated
AWS Lambda: Automates the file processing workflows
ECS (Fargate): Handles document processing efficiently
AWS Step Functions: Coordinates the various processes involved in splitting, processing, and merging documents
Amazon Bedrock: Generates alt text for images and charts using advanced LLM capabilities

This solution also integrates Adobe Auto-Tag APIs which are designed to automatically clean metadata, apply appropriate tags, and further enable document remediation.

More information:

My eCISO

The My eCISO project revolutionizes cybersecurity assessments by leveraging generative AI to assist organizations in strengthening their cyber resilience. Addressing the challenges posed by evolving cyber threats and the complexities of evaluating cybersecurity infrastructure, the Cal Poly Digital Transformation Hub (DxHub), powered by Amazon Web Services (AWS), developed an innovative AI-driven application. My eCISO automates the assessment process by conducting natural language interviews based on the NIST 1.1 framework, enabling organizations to accurately gauge their security posture and implement necessary safeguards. This tool significantly enhances operational efficiency and provides a robust, user-friendly approach to maintaining cybersecurity compliance.

Learn more:

Blog + Open Source Code

Query Police Reports

The CaseQuery Pro project transforms data access in the law enforcement sector by automating data ingestion across various formats, saving thousands of hours in manual data entry and analysis. Addressing the inefficiencies and delays caused by manual searches through extensive case data, CaseQuery Pro integrates AWS services to develop a user-friendly, conversational chatbot. This tool allows lawyers to quickly and securely access crucial case information, significantly enhancing operational efficiency and ensuring compliance with stringent legal data handling requirements.

9-1-1 Helper

The 911 Call Taker Classifier project revolutionizes emergency call handling by employing advanced AI to assist dispatchers in rapidly and accurately categorizing incoming calls. Addressing the critical need for accurate response categorization to prevent misclassification and delays, the DxHub team in collaboration with Seattle Police developed an innovative application. This tool analyzes the caller’s spoken words in real-time and displays the most probable police response category for 911 call takers in near real-time. By integrating AI technologies, the 911 Call Taker Classifier significantly enhances dispatcher efficiency and ensures timely and appropriate police responses, ultimately improving public safety outcomes.