In my years of leading child support programs as a Director in Riverside County, at the State of California and now as a technology partner with CGI, I've learned that most challenges in this area don't stem from lack of effort or commitment by those engaged. The challenges stem from the sheer complexity of the work and the systems we ask people to navigate every day.
Caseworkers and attorneys operate at the intersection of law, policy and human circumstance. They interpret statutes, apply policies consistently, meet federal requirements and engage families with empathy and fairness, often simultaneously. It's demanding work, even on the best days.
Lately, artificial intelligence has entered the conversation as a potential solution. Agencies are stretched thin. Caseloads are high. Training resources are limited. Turnover is constant. The promise of technology delivering faster, more precise answers is understandably attractive.
But in child support, enthusiasm must be tempered by responsibility.
The AI paradox in child support
Child support programs handle some of the most sensitive information in government: personally identifiable information, federal tax data, financial records and court documents. Protecting that data isn't just best practice, it's foundational to public trust and federal compliance.
This creates a fundamental tension: we want AI's benefits, but we can't afford its typical risks.
Many public sector leaders tell me they're interested in exploring how AI can drive outcomes across their agencies. Still, these same leaders are hesitant to act out of worry about data exposure, legal uncertainty and unintended consequences. That hesitation isn't resistance to innovation; it's sound judgment.
The real pain point isn't information, it's guidance
One of the biggest challenges is accessing clear, consistent guidance.
Policies, statutes and procedures are scattered across policy manuals, training materials, memos, legal guidance and institutional memory. Even experienced staff struggle to find answers quickly, especially for nuanced situations that don't fit textbook examples.
I've watched highly capable caseworkers spend precious time hunting for information they know exists, or second-guessing answers because they're unsure whether policies have changed. This friction doesn't just slow work; it creates inconsistency and stress.
Why consistency matters more than ever
Consistency has always been important in child support, but it's critical now. Families expect fair treatment regardless of which office handles their case. Attorneys need confidence that guidance aligns with current law. Supervisors need assurance that staff apply policies correctly.
Yet achieving consistency across teams, counties or states is difficult when guidance gets interpreted or applied unevenly. Training helps, but it doesn't scale, and it can't always provide answers in the moment when a real case demands a decision.
A more practical vision for AI
The most promising role for AI in child support isn't automation or decision-making. It's decision support.
This means using AI in a constrained, intentional way to help staff navigate the policies, statutes and training materials they're already expected to follow. Not replacing judgment. Not accessing live case data. Not making determinations.
Think of AI as a modern extension of existing reference tools:
- Access relevant policy faster when staff need guidance
- Interpret questions in plain language rather than requiring precise search terms
- Deliver consistent, current answers across the entire organization
When designed thoughtfully, this approach aligns with how child support actually works: grounded in rules, guided by discretion, executed by professionals who understand the human impact of every decision.
Questions leaders should ask now
As agencies explore AI, the most important questions aren't about technology; they're about governance and intent:
- How do we support staff without increasing risk?
- How do we improve consistency without removing human judgment?
- How do we make guidance more accessible without exposing sensitive data?
- How do we help new staff ramp up faster while supporting experienced staff equally well?
These aren't abstract questions. They're operational ones, rooted in reality.
Moving forward carefully and confidently
Child support has always evolved cautiously, and for good reason. But caution doesn't mean paralysis. With the right guardrails, we can explore AI in ways that respect the unique responsibilities of this work.
The key is choosing the right partner, one who understands both the technology and the unique constraints of child support work. Not every AI solution is built with government compliance, data security and human judgment at its core.
The goal isn't chasing trends. It's making the work more sustainable, more consistent and more supportive for the people who do it every day. That requires partners who understand the stakes, not just the technology.
From where I sit, that's a conversation worth having, carefully, thoughtfully and grounded in the realities of child support, not the hype of technology. Reach out today to begin the conversation.
About this author
David Kilgore
Director, Consulting Services
David Kilgore is a Director for the Child Support Practice with CGI, bringing over 20 years of experience, most recently leading California’s Child Support Program. Under his leadership, the department achieved significant technological and programmatic advancements to improve children's lives. Notably, he spearheaded a full ...
