Hope Biosciences runs clinical trials on one login instead of five.

Hope Biosciences did not have a technology problem so much as a connective-tissue problem. Salesforce, a Skuid-built partner community, Zoom, Tableau, Google Workspace, and a research database each did its job well and none of them spoke to the others. The most telling symptom was a login screen that said "logged in" while quietly loading nothing - the byproduct of six systems with six different session timeouts. Our brief was not to rip and replace, which no regulated research operation can afford, but to make the tools they already trust behave as one.

Document generation was the first proof it could work. Consent forms and the Texas HB300 authorization were being assembled by hand, so we built a Google Picker integration directly onto the Skuid pages: a staff member picks a template from Drive, the system replaces placeholders - including nested fields like a contact's account name and picklist labels - with live Salesforce data, and saves it back to Drive, entirely in the browser with no Apex in the path. Along the way the HB300 template was rebuilt from four pages to two without losing a required field.

Identity was the structural fix. We made JumpCloud the central identity provider so the community, Zoom, and Tableau all authenticate through one sign-on, then added a session-expiry detector and aligned every timeout to two hours - which is what finally killed the "logged in but blank" failure. On top of that single front door we layered the pieces a trial actually needs: a Twilio integration that texts participants and captures their replies through a webhook into an auditable engagement history, a deliberately one-way calendar sync that pushes study visits out with a Meet link but never lets a calendar event create a scheduling conflict, and Tableau dashboards rendered inside the community through a JWT-authenticated connected app.

Underneath all of it was an org that had grown to roughly 1,600 objects with undocumented triggers and flows and several inherited features quietly broken. We wrote data dictionaries for the objects, documented triggers with their handler summaries and flows element by element, and repaired the broken copy-to-file and study-search features - the unglamorous work that makes a life-sciences Salesforce org safe to keep building on. Integration code shipped at 92% test coverage.