BridgeOS

Why I built it

The work from the first plan to the submission was in many separate places. These included integrated schedules, Module 3 document plans, evidence files, CDMO follow-up, health authority readiness, and late-stage review cycles. BridgeOS puts these steps in one workspace.

What it does

• Mission Control shows the route, blockers, owners, and dates in one place.
• Evidence links the status of each source to the work products that use it.
• CMC Studio uses the source documents to make cited draft text.
• Submission and Lifecycle show the status of the submission pack and the impact of each change.

What goes in

Program route assumptions, public regulatory information, private evidence files, document plans, and the current review and submission needs.

What comes out

Route options, milestone workbacks, evidence maps, cited draft sections, controlled review packs, and notes on the impact of lifecycle changes.

What it changes

The route logic, evidence status, drafts, submission packs, and lifecycle work stay in one chain that the team can see. They are not spread across separate tools and email handoffs.

How it stays inside the rules

What it is for

Program planning and evidence support for a CMC PM and a regulatory CMC lead. The tool shows filing route options, dependency exposure, and draft work products for human review. It makes no quality, release, or disposition decision.

What it never does

This tool is not a critical GMP application. It does not do batch disposition, release decisions, specification decisions, or process control. It sits above the governance layer, not inside the manufacturing layer. Under the Annex 22 draft, that is the only place for an LLM.

Who decides

The IPT makes the route decision and records it in the decision register. The tool ranks the options and shows the evidence for each option. The named accountable owner selects one.

Risk level

Medium under FDA's draft AI framework (how much the output influences a decision × how serious that decision is)

Guidance it follows

FDA draft guidance · EU GMP Annex 22 · ICH Q8 to Q14

Where companies can use AI in regulated CMC work →

What already exists

• Veeva Vault RIM: The system of record for regulatory information, submission content, and registration status
• ArisGlobal LifeSphere (NavaX): A generative AI layer for safety and regulatory work: dossier QC, intelligence summaries, and label change impact
• MS Project, Smartsheet, Planisware: Schedules, dependencies, critical path, and resource plans

The gap it targets

These tools record status and schedules. My focus is the filing route decision. To select a route, a CMC PM weighs guidance precedent, evidence readiness, schedule exposure, and the later lifecycle burden at the same time. This demo shows those four things in one decision chain.

What it is not

It is not a RIM system and not a system of record. It reads data from these systems and never replaces them.

Do you build tools like this?

I am glad to compare notes: what worked, what a reviewer did not accept, and where AI should stay out of the decision.