Onboarding had to introduce an unfamiliar idea before asking for anything in return. A short introductory sequence explains what Clinsight does with a lab report and what a patient gets back, so the first thing a new user meets is an explanation rather than a form.

The decision that shaped the rest of the flow was offering two ways in rather than one. Alongside the standard sign-up, the same screen offers a guest route. Rather than forcing registration upfront, users could immediately upload a lab report, receive an AI interpretation, and ask a limited number of follow-up questions before deciding whether to sign up, letting patients judge the platform on its own output instead of on a promise made before they had seen it.

An account is what the rest of the product depends on: saved reports, interpretation history and doctor reviews all need somewhere to return to. The guest route is bounded for that reason rather than built as a parallel version of the product. It exists to earn the account, not to replace it.

Uploading a report is the first meaningful interaction patients have with the platform, so simplicity was essential. Users can upload PDF or image files, after which OCR extracts test names, values, reference ranges, and units for AI analysis.

Instead of presenting patients with raw clinical values, Clinsight transforms their results into a structured explanation consisting of a plain-language summary, a breakdown of abnormal values, and suggested next steps.

We framed the output as an interpretation of laboratory results rather than a diagnosis. The language focuses on helping patients understand what their results could mean without making definitive medical claims, and the option to request a verified doctor's review stays visible throughout, reinforcing that the AI provides context rather than replacing professional medical judgment.

That boundary also decided what we chose not to show. We avoided severity scores, definitive diagnoses, and statements suggesting a patient had a specific condition, all of which would make the AI appear more medically authoritative than intended. Instead we focused on giving patients useful context while keeping a clear line between AI interpretation and professional clinical opinion.

Understanding health information often raises additional questions, so we designed a contextual AI chat that allows patients to continue the conversation based on their uploaded report. By maintaining context throughout the interaction, users receive explanations tailored to their specific results rather than generic health information.

The Insights page gives patients a centralized view of AI-generated health insights based on their uploaded laboratory reports. Each insight presents a concise summary of key findings, making it easy for users to revisit important information without reopening individual reports.