Workflow Redesign: From Role-First to Outcome-First
Traditional biotech workflows are designed around roles: regulatory affairs owns the NDA, clinical operations owns the trial, chemistry owns the lead candidate. Each function optimizes its own process, and the outputs get handed off across functional boundaries.
This design made sense when the coordination cost of cross-functional workflows was high — when it was genuinely more efficient for each function to run its own process independently. AI changes the coordination cost equation. Cross-functional coordination that previously required dedicated project management can now be handled by shared infrastructure. The role-first design is no longer the most efficient design.
The outcome-first alternative
An outcome-first workflow starts with the desired outcome — IND submission, Phase 2 start, NDA approval — and asks: what is the optimal process to achieve this outcome, given the current capabilities of AI-enabled teams?
The answer is usually a hybrid workflow that cuts across traditional functional boundaries:
IND preparation example. Traditional: CMC prepares Module 3 independently, pharmacology prepares Module 4 independently, clinical prepares Module 5 independently, regulatory integrates. Outcome-first: a cross-functional working session at the start defines the submission strategy. AI maintains a shared dossier with real-time consistency checking across all modules. Each function contributes to the shared dossier rather than to independent documents that are later integrated.
Trial protocol example. Traditional: clinical drafts protocol, operations reviews for feasibility, regulatory reviews for compliance. Sequential, with significant rework at each handoff. Outcome-first: AI generates a protocol draft from a clinical brief and a feasibility data input simultaneously. Cross-functional review happens in parallel on the shared draft. Rework at integration is minimal because integration was designed in from the start.
What workflow redesign requires
Outcome-first workflow redesign requires three things: leadership alignment on what the outcome actually is (not the functional deliverable, but the program milestone), explicit decision about who owns the cross-functional process, and AI infrastructure that can support cross-functional collaboration on shared documents.
The last is easy. The first two are organizational questions that AI cannot answer.