Smart Demographics (ML)
Asking a participant to re-enter their date of birth, ethnicity, and postcode for the fifth time is a sure way to lose them. Smart Demographics solves this with a fully on-server ML pipeline that identifies demographic answers, normalises them, and writes them back to the participant record — ready to pre-fill the next study.
Privacy by architecture, not by promise
Demographic recognition runs inside our own infrastructure. No participant data is sent to OpenAI, Anthropic, Google, or any other public LLM provider. The models are small, focused, and auditable — and that auditability is one of the reasons clinical teams choose ProNova52 over generic AI survey tools.
Recognises the answers, not just the questions
Whether you ask 'What is your gender?' or 'How do you describe your gender identity?', the platform recognises the answer pattern and maps it to the standardised demographic field. The original question and answer remain in the response — the normalised value is stored alongside on the participant record.
Pre-fill that respects consent
On the next study, demographic questions appear pre-filled and clearly marked — the participant can edit or clear them before submitting. Pre-fill never bypasses consent or the participant's right to update their record.
Capabilities
- On-server machine learning — no public LLM calls
- Recognises 12+ demographic fields (DOB, gender, ethnicity, address, etc.)
- Normalises free-text answers into structured fields
- Writes back to participant record via secure RPC
- Pre-fills future studies with editable values
- Researcher-role PII masking still applies
Typical use cases
- Reducing drop-off on repeat PROM check-ins
- Building a longitudinal demographic dataset across studies
- Standardising free-text demographic answers for analysis