Policy & Practice | Fall 2026

case records, one-by-one, to check for this. By automating the identification of documentation delays, the organi zation transformed a labor-intensive audit process into a real-time man agement tool that helps workers stay current and allows supervisors to intervene earlier. Completing the Promise of Data-Driven Child Welfare The transition from the Statewide Automated Child Welfare Information System (SACWIS) to the Comprehensive Child Welfare Information System (CCWIS) reflected an important recognition: compliance reporting and operational decision making are not the same thing. CCWIS represented a significant step forward by emphasizing interoper ability, flexibility, and analytics. Yet these alone do not solve the challenge facing most child welfare agencies. Agencies can connect systems, exchange data, and modernize infra structure, but if the most important information remains trapped within narrative documentation, decision makers will continue to operate with only a partial view of reality. The examples of NLP use emerging from agencies across the country suggest that the next evolution in child welfare data infrastructure is not simply better reporting. It is the ability to integrate narrative analytics along side structured case management data as part of everyday operations. The information agencies need to identify unmet needs, monitor compli ance, assess practice quality, support family preservation, and improve outcomes is already being generated. It already exists within the case record. For agency leaders, the question is no longer whether the data exist. The question is whether they have the tools to see it. As child welfare continues its journey toward becoming truly data driven, unlocking the narrative may prove to be the final step in fulfilling the promise that decades of tech nology modernization efforts were intended to achieve.

practice, support compliance efforts, and uncover trends that would other wise require extensive manual review. The sibling placement story dem onstrates how narrative analytics can support permanency and family preservation efforts by surfacing relationships hidden within years of documentation. Similar approaches are helping agencies address a wide range of operational challenges. At MA DCF, staff used narrative analysis to identify transition-aged youth experiencing housing insecurity across thousands of cases, including young adults who were not already connected to housing services. The agency also implemented an ongoing process to identify references to Native American ancestry within case documentation to support compliance with the Indian Child Welfare Act. Information necessary for tribal noti fication often exists only in narrative records, making traditional reporting methods insufficient. The Illinois Department of Children and Family Services (IL DCFS) has used narrative analytics to identify pregnant women within intact family cases, uncovering 621 additional indi viduals statewide whose needs had not been captured in structured data. This enabled the agency to move from reactive service delivery to a more pro active approach. IL DCFS also applied NLP to docu mentation associated with “good faith efforts” requirements for investigators. While contacts with alleged victims and perpetrators were frequently doc umented in case notes, those activities often could not be captured through traditional compliance reporting. By analyzing narrative documentation at scale, the agency verified thousands of compliant investigations that had previously gone uncounted because the evidence existed only in free text. Nonprofit providers are seeing similar benefits. At Aspiranet, a large California based nonprofit, staff supporting transition-aged youth previously relied on manual reviews to determine whether required documentation updates were overdue. Supervisors and staff had to examine individual

throughout the 1990s accelerated auto mation and standardized reporting through AFCARS and NCANDS. More recently, the transition to CCWIS has emphasized interoperability, analytics, and operational decision making. Through Child and Family Services Reviews (CFSRs), Continuous Quality Improvement (CQI), and many other frameworks, agencies are increasingly expected to use data not simply for compliance, but to improve practice and outcomes. Yet despite these investments, one fundamental challenge has persisted. Most child welfare information systems were designed around struc tured, quantifiable data: checkboxes, drop-down menus, date fields, and categorical codes. These systems excel at supporting reporting requirements and compliance monitoring. They are far less effective at capturing the com plexity of child welfare practice. When a caseworker documents a home visit, they are not simply recording an event. They are describing family dynamics, emerging risks, strengths, housing concerns, service barriers, parent-child interac tions, and observations that inform professional judgment. Those nar ratives contain some of the most important information available about children and families. They also repre sent the largest source of data within the child welfare system. Historically, that information has remained inaccessible at scale. Supervisors reviewing individual cases can read it. Caseworkers can reference it. But agencies have had no practical way to analyze millions of narrative records across thousands of cases. As a result, the information most reflective of day-to-day practice has remained largely invisible to system leaders. Turning Narrative Into Actionable Data Advances in AI and natural-lan guage processing (NLP) are changing that equation. Rather than relying exclusively on structured fields, agencies can now analyze narrative documenta tion across an entire organization’s caseload to identify needs, monitor

Marty Elisco is the CEO at Augintel.

Fall 2026 Policy & Practice 41