AI transparency,
page by page.
We use AI for research assistance, initial drafting, and structured-data generation. Every page is then reviewed by a human editor and, for state-specific content, a licensed reviewer in that state.
Where AI helps
Research synthesis (pulling together federal and state regulatory sources), initial drafting of explainer paragraphs, structured-data generation (JSON-LD schema), and internal-linking suggestions. These are editor-assisting tasks.
Where AI doesn't help
AI does not decide whether a number is current. AI does not substitute for a licensed reviewer's read on whether a recommendation drifts into unauthorized practice of law. AI does not write atomic answers that we publish without human editing.
How the review flow works
1. Researcher gathers primary sources (state Medicaid manual, federal CFR, recent case law). 2. AI helps draft and structure. 3. Human editor revises for voice, accuracy, and AEO hygiene. 4. Licensed reviewer in the applicable state signs off. 5. Publish with dated "last verified" stamp.
Why we disclose this
Medicaid planning is a high-stakes domain where wrong information costs families real money. We'd rather be clear about AI involvement than pretend every word was typed by hand. The test is whether the final output is accurate and useful — not whether AI was involved in producing it.