Policy & Practice | Fall 2026

technology speaks

By Jeremy Gwin, MBA, MPsy

Modernizing in Place: Building an Intelligence Layer Around Human Services Systems

H uman services modernization has often been framed as a replace ment decision: keep the legacy system or replace it. Agencies are managing aging technology, constrained budgets, workforce shortages, growing nar rative data, and expectations for faster, more consistent decisions. Yet many lack the funding or capacity to launch another multiyear core system replacement. A more practical question is emerging: What can agencies improve now, without starting over? That question is especially timely in child welfare. Federal interest in predictive analytics, including dem onstration and grant opportunities, is encouraging agencies to explore how data can better support safety, perma nency, well-being, supervision, and continuous improvement. The oppor tunity is not to replace professional judgment with automation. It is to extend familiar analytics environments so they provide earlier signals, better context, and more actionable insight. The Architecture Matters Most agencies already rely on dashboards and reporting tools to monitor intake volume, investigation timeliness, placement stability, case contacts, permanency progress, and other performance measures. These tools answer retrospective questions: what happened, where performance changed, and which activities are overdue. Predictive and practice-facing ana lytics can connect those environments

whether a child is safe or whether a report should be investigated. Its role is to identify reports that may warrant faster review, additional context, or supervisory attention. For high-volume intake operations, predictive prioritization can surface relevant signals, support consistent application of policy and judgment, and help leaders see emerging patterns across reports, regions, or populations. IntelliGuide applies a related concept to field practice. Agencies collect large volumes of narrative documentation through contact notes, visit summaries, case plans, super visory notes, and service updates. Historically, much of this informa tion is reviewed through supervision, quality assurance sampling, audits, or incident-specific review.

more directly to current operational decisions. Operating as an intelli gence layer around existing platforms, these capabilities use data exchange, natural-language processing, model monitoring, and dashboard integra tion to add value without replacing the system of record. A capability can be introduced at a defined decision point, connected to existing data and workflows, evaluated against a specific operational problem, and expanded only after value is demonstrated. IntelliTriage and IntelliGuide illustrate this pattern. IntelliTriage applies predictive ana lytics and natural-language processing to intake information, including structured history, report character istics, prior involvement, allegation patterns, household context, and nar rative content. Its role is not to decide

Illustration by Chris Campbell

Policy & Practice Fall 2026 38