Overview

The Arizona Department of Child Safety (DCS) is responsible for protecting children, strengthening families, and ensuring the safety, permanency, and well-being of vulnerable children throughout Arizona. The department investigates reports of child abuse and neglect, provides support and intervention services to families, coordinates foster care and kinship placements, facilitates adoption and guardianship services, and works closely with courts, law enforcement, healthcare providers, schools, and community partners to achieve positive outcomes for children. DCS serves thousands of children and families across the state each year through a workforce that includes caseworkers, supervisors, support staff, and partner organizations operating in both urban and rural communities. The department's mission is to successfully engage children and families to ensure safety, strengthen families, and achieve permanency. Technology plays a critical role in supporting that mission. DCS operates Guardian, Arizona's Comprehensive Child Welfare Information System (CCWIS), which serves as the system of record for child welfare operations and supports more than 3,500 users statewide. Guardian enables case management, investigations, provider management, court reporting, payments, analytics, and collaboration across the child welfare ecosystem. The department continues to modernize its technology environment through cloud adoption, cybersecurity enhancements, data governance, and responsible artificial intelligence initiatives designed to improve operational efficiency and return valuable time to frontline staff. At its core, DCS exists to ensure that every child has the opportunity to grow up in a safe, stable, and supportive environment. Every program, process, and technology investment is aligned to that mission and focused on improving outcomes for Arizona's children and families.

Supernova Award Category

Tech Optimization and Modernization

The Problem

The Arizona Department of Child Safety relies on Guardian, Arizona’s Comprehensive Child Welfare Information System (CCWIS), to support critical child welfare operations across the state. As the system grew in size and complexity, technical teams faced increasing challenges in rapidly identifying the root cause of application issues, efficiently triaging incidents, and maintaining high code quality across ongoing enhancements and releases. These activities required significant manual effort, resulting in longer resolution times, increased operational overhead, and less time available for innovation and mission-focused work. To continue modernizing its technology environment while maintaining a reliable and secure platform for more than 3,500 users, DCS needed a way to improve productivity, accelerate issue resolution, and enhance software quality without increasing staffing levels. The department sought to leverage responsible AI capabilities to streamline technical operations, reduce repetitive manual tasks, and allow technology staff to focus more of their time on delivering value to frontline workers serving Arizona's children and families.

The Solution

DCS analyzed operational metrics, support trends, incident response activities, and software delivery processes to identify areas where highly skilled technical staff were spending significant time on repetitive manual tasks. After evaluating several modernization options, the department selected AI-powered Root Cause Analysis, Triage, and Code Review agents because they could be integrated into existing workflows, accelerate issue resolution, improve software quality, and deliver measurable productivity gains while maintaining human oversight and accountability.

The results

Before implementing the AI-powered RCA Agent, Triage Agent, and Code Review Agent, highly skilled technical staff spent significant time manually reviewing logs, investigating incidents, categorizing support requests, and performing repetitive code quality reviews. These activities were necessary but often delayed issue resolution, slowed software delivery, and limited the amount of time available for modernization efforts and mission-focused innovation. Today, these AI capabilities serve as force multipliers for the DCS technology organization. The RCA Agent accelerates incident investigations by rapidly identifying likely causes and impacted components. The Triage Agent improves the speed and consistency of issue intake and prioritization, ensuring critical child welfare operations receive immediate attention. The Code Review Agent continuously analyzes software changes for quality, security, and compliance concerns before deployment, reducing defects and improving application reliability. The broader impact extends beyond productivity gains. By embedding responsible AI into core IT operations, DCS has established a repeatable model for how public sector organizations can modernize software engineering and operational support without increasing staffing levels. The approach allows technical professionals to focus on higher-value work while AI handles routine analysis and review activities. As government agencies nationwide face increasing demands, workforce constraints, and growing technology complexity, this model demonstrates how AI can fundamentally transform IT operations, accelerate modernization efforts, improve service delivery, and ultimately return more time and resources to the mission. The disruptive potential lies not in replacing people, but in enabling public sector organizations to achieve significantly greater outcomes with the teams they already have.

Metrics

Before implementing the AI-powered RCA Agent, Triage Agent, and Code Review Agent, technical teams manually analyzed incidents, reviewed support requests, and conducted code reviews. Initial measurements showed that incident investigations often required multiple hours of engineering effort, issue categorization varied depending on the reviewer, and code review activities consumed valuable development time that could otherwise be spent delivering enhancements to Guardian. Following implementation, DCS observed measurable improvements across key operational areas. Root cause investigations that previously required hours of manual analysis can now be completed in minutes, reducing time to identify likely causes by more than 70%. Automated triage has reduced manual ticket classification effort by approximately 60% while improving consistency in routing and prioritization. The Code Review Agent reviews 100% of code submissions and has reduced manual review effort by more than 50%, allowing developers to focus on remediation and feature delivery rather than repetitive inspection activities. Collectively, these capabilities have returned hundreds of staff hours annually to the organization, accelerated issue resolution, improved software quality, and increased operational efficiency without adding personnel. Most importantly, these gains allow DCS to devote more resources to modernizing Guardian and supporting the frontline workforce responsible for serving Arizona's children and families.

The Technology

The solution was built using Microsoft Azure AI services, Azure OpenAI, GitHub Copilot, and custom AI agents integrated into the DCS software development and operational support lifecycle. These technologies were combined with existing DevOps, application monitoring, source control, and ticket management platforms that support Guardian, Arizona's Comprehensive Child Welfare Information System (CCWIS). The implementation included three purpose-built AI agents: a Root Cause Analysis (RCA) Agent that analyzes logs, telemetry, and operational data to accelerate incident investigations; a Triage Agent that reviews and categorizes incoming issues, defects, and enhancement requests; and a Code Review Agent that evaluates software changes for quality, security, performance, and compliance with development standards. All solutions operate within DCS governance and security controls and maintain human oversight for decision making and approval activities.

Disruptive Factor

One of the biggest challenges was introducing AI into core software engineering and operational support processes while maintaining the security, governance, and reliability requirements expected of a child welfare agency. Rather than treating AI as a standalone tool, DCS embedded AI directly into the software delivery and support lifecycle, creating a human-in-the-loop model that augments technical staff instead of replacing them. This challenged the traditional approach to government IT operations, where incident analysis, work triage, and code reviews have historically been manual, time-intensive activities. What sets DCS apart is the application of responsible AI to the modernization of mission-critical government systems, demonstrating how public sector organizations can dramatically increase productivity, improve quality, and accelerate delivery without increasing workforce size.

Shining Moment

What we are most proud of is that this project was never about technology for technology’s sake. It was about finding a responsible way to give time back to our technical teams so they can spend more of their effort improving the systems that support Arizona’s children and families. By successfully integrating AI into mission-critical operations while maintaining human oversight, governance, and accountability, DCS demonstrated that innovation and public trust can coexist. The result is a modernization model that not only improves how we support Guardian today but also establishes a foundation for the future of responsible AI adoption across government.