First-Run Onboarding for a Platform You Can't Learn by Poking

Some products you can learn by clicking around. A bank's risk-intelligence platform is not one of them. BlackSwan serves analysts doing day-to-day reviews and investigators researching flagged entities, across five surfaces and a Risk Engine dense enough that trial-and-error isn't a safe learning path. For a PM, that reframes first-run onboarding from a courtesy into a path to competent first use. The design principle I followed: orient by surface and role, not by feature tour. An analyst's first run should land them in triage understanding how to read a flagged result and act on it; an investigator's should make the Entity page's relationship mapping immediately legible. Throwing a single linear walkthrough at both users would serve neither. I leaned on the design system's consistency to do quiet onboarding work too — once a status pill or a drill-down pattern is learned on one surface, it transfers, so every screen after the first costs less to learn. The PM payoff is faster time-to-value and fewer support escalations from confused new users. In a domain where the cost of a confused user is a missed or mishandled risk, onboarding isn't the welcome mat. It's the first line of getting the work right.