Why AI matters at this scale

Franklin County Children Services (FCCS) is a public agency responsible for child protection, foster care, and family preservation services in Ohio's most populous county. Founded in 1877, it operates with a staff of 501-1000 employees, handling thousands of complex cases annually involving abuse, neglect, and family crisis. Its mission-critical work is data-intensive, governed by strict regulations, and perpetually strained by high caseloads and limited resources.

For a mid-sized public agency like FCCS, AI presents a transformative lever to enhance decision-making and operational efficiency. At this scale, manual processes for risk assessment, documentation, and resource coordination consume vast amounts of caseworker time that could be redirected to direct family engagement. AI can process the agency's extensive historical and real-time data to uncover patterns invisible to human analysis, enabling a shift from reactive intervention to more proactive, preventative support. This is crucial for improving child safety outcomes while managing public funds responsibly.

Concrete AI Opportunities with ROI

1. Predictive Triage for Investigative Response: An AI model analyzing historical referral data (call text, prior history, demographic factors) can assign risk scores to new cases. This allows supervisors to prioritize the most severe allegations for immediate response, potentially reducing harm and improving investigator efficiency. The ROI is measured in better child safety outcomes and optimized use of limited investigative staff.

2. Administrative Automation for Caseworkers: Natural Language Processing (NLP) can automatically summarize lengthy case notes, draft routine reports, and extract data for compliance filings. For an agency with hundreds of caseworkers, this could reclaim thousands of hours annually from paperwork, directly increasing time available for family visits and support—a clear ROI in staff capacity and job satisfaction.

3. Intelligent Resource Matching: A recommendation system can match families with appropriate community-based services (e.g., mental health, housing, job training) by analyzing family needs, service eligibility, and provider availability. This improves service uptake and outcomes, creating ROI through more effective use of community partnerships and potentially reducing repeat system involvement.

Deployment Risks for a 501-1000 Employee Agency

Deploying AI at this size band carries specific risks. Budget constraints may limit investment in modern data infrastructure required for AI. Legacy systems common in government can hinder data integration. There is a acute risk of algorithmic bias; models trained on historical data could perpetuate systemic disparities in which families are investigated. The agency must ensure any AI tool is explainable, especially when informing high-stakes decisions about child removal. Furthermore, a workforce with varying tech literacy may resist or misuse tools without comprehensive change management and training. Success depends on starting with low-risk, high-support use cases that demonstrate clear benefit to frontline staff.