The AI System Helping Veterans Get Help When It Matters Most
Protagona partnered with a Houston-based veterans services nonprofit to design and deploy an AI-powered priority scoring proof of concept on AWS — automating triage across hundreds of veteran records and surfacing the cases most urgently needing human intervention.
Industry
Nonprofit
Teams & Services
AI/ML Engineering, Back-End Engineering, Cloud Architecture, Engagement Management
Tech & Tools
AWS Bedrock, AWS Lambda, Amazon API Gateway, Amazon DynamoDB, Amazon S3, AWS CloudWatch, Amazon Lex, Typesense, Python
Key Data Points
The Vision
A Houston-based nonprofit connects veterans and their families to partner organizations across employment, housing, healthcare, and other critical services. As their caseload grew, prioritizing which veterans needed immediate attention became increasingly complex. Rather than committing to a full build, they chose a focused proof of concept — using Generative AI to assess veteran history and surface the cases most urgently requiring human intervention.
The Goal
Protagona was engaged to design and deploy a working AI-powered priority scoring proof of concept on AWS. The system needed to ingest veteran connection data, apply a configurable severity algorithm, and output ranked priority tiers to guide staff intervention decisions. The POC had to demonstrate three things before any investment in a full production system: scoring accuracy across diverse veteran records, full transparency into AI reasoning, and an architecture sustainable enough for the client team to operate and evolve independently after handoff.
The Challenge
The scoring system had to reason across a veteran's full service history and produce an auditable priority ranking — not just evaluate a single intake field. Records frequently arrive incomplete, so the model needed to handle data gaps without dropping veterans from the queue. Every scoring decision also had to be explainable and cross-verifiable against Salesforce, so staff could trust the output.
Connecting AWS Bedrock to an existing Typesense search infrastructure required close collaboration with both technical staff and frontline intake workers, while balancing token efficiency, inference cost, and scoring accuracy across a diverse veteran population.
The Solution
Protagona built an event-driven, serverless AWS architecture that routes veteran records through a Generative AI scoring pipeline. Bedrock handles LLM inference, Lambda manages orchestration, API Gateway triggers scoring requests, and DynamoDB stores results with change-detection logic. S3 supports bulk batch processing.
Prompt engineering was collaborative — working directly with intake staff to validate scoring criteria, refine priority weights, and handle edge cases like missing data. CloudWatch captures the AI's reasoning for every decision, giving staff full visibility to cross-verify outputs against Salesforce. Knowledge transfer documentation was built in parallel with the final sprint, ensuring the client team could operate the system independently after handoff.
.png)