This post is part of a series on Health Samurai sessions from HL7 FHIR DevDays 2025. In this article, “Enhancing SDC workflows with AI: Automation and beyond,” Aidbox Forms Product Manager Maria Ryzhikova shows how AI can reduce manual work in FHIR SDC Questionnaire design, multilingual support, and other data-capture workflows.
What you'll learn:
FHIR SDC builds on Questionnaire and QuestionnaireResponse, adding conditional logic, UI hints, pre-filling, and structured extraction so complex forms can be implemented in a consistent way. Designing these forms is labor-intensive: implementers must define nested item structures, author FHIRPath expressions, configure prefill logic, and map responses into data models. Many workflows still start from paper forms, making manual digitization repetitive and error-prone, which motivates using AI to handle structure and layout while humans keep control of rules and validation.
The session demonstrates converting the GAD‑7 anxiety screening PDF into a FHIR Questionnaire. A PDF is uploaded to the Aidbox Public Form Builder, OCR extracts text and layout, and an AI model proposes a Questionnaire structure that mirrors questions and answer options. The generated JSON is then validated, issues are corrected, and the Questionnaire is stored on a FHIR server and opened in a visual builder, where layout, widgets, and scoring (for example, total score via FHIRPath) are refined. AI supplies a structured starting point instead of requiring a fully manual build.
Maria also shows how the same form can be maintained in multiple languages without duplicating Questionnaires. A default language is set on the Questionnaire, a target language is chosen, and an AI-based interface generates translations for item text and answer labels. These translations are stored using standard FHIR translation extensions and can be reviewed and edited. Switching the active language re-renders the form (for example, English vs French) from the same underlying Questionnaire definition.
Planned and experimental uses of AI in SDC include: generating expressions, prefill rules, and extraction mappings from textual prompts; building complete forms from narrative descriptions; mining existing forms for reusable components; guiding users through forms with conversational or voice interfaces; and turning PDFs into QuestionnaireResponses to capture completed paper forms. The overall aim is to keep SDC as the technical backbone while using AI to lower the effort required to design and fill structured forms.
Watch Maria's full presentation from FHIR DevDays 2025 and the complete Health Samurai playlist:
Watch the full video
Full playlist on YouTube
Connect with Maria on LinkedIn.
Try Aidbox Forms and its Public Builder to experiment with PDF-to-Questionnaire conversion, validation, and multilingual SDC forms.
Get in touch with us today!
