Skip to content
How I AIHow I AI

How this PM streamlines 60k-page FDA submissions with Claude, Streamlit, and clever AI workflows

Prerna Kaul is a product and platform leader who has spent over 14 years turning machine-learning research into consumer and B2B products at Amazon Alexa, AGI, Moderna, and now Panasonic Well. In today’s episode, she explains how she’s using AI to slash some of the most time-consuming, expensive tasks in life sciences—from generating 60,000-page FDA submissions to crafting communication frameworks that help product managers navigate complex stakeholder dynamics. Her innovations are saving millions of dollars and helping lifesaving treatments reach the market faster. *What you’ll learn:* 1. How Prerna built an AI system that automates the creation of 60,000-page regulatory documents for the FDA—reducing a process that took 4 to 6 months and 20 specialists to just minutes 2. A step-by-step system for detecting and redacting PHI (protected health information) in clinical trial data using Claude 3. How to build user-friendly interfaces for non-technical colleagues using Streamlit to democratize AI tools 4. How to use Claude’s prompt generator to create powerful communication frameworks that help PMs navigate complex stakeholder situations 5. Why transparency about AI costs is crucial for gaining organizational buy-in and tracking ROI 6. A practical framework for approaching AI safety and ethics in highly regulated industries *Brought to you by:* CodeRabbit—Cut code review time and bugs in half. Instantly: https://www.coderabbit.ai/ Lovable—Build apps by simply chatting with AI: https://lovable.dev/ *Where to find Prerna Kaul:* LinkedIn: https://www.linkedin.com/in/prernakkaul/ *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo *In this episode, we cover:* (00:00) Introduction to Prerna (03:01) The FDA submission challenge: 60,000 pages, months of work, millions in costs (05:20) Getting started in Claude: from prompt to production-ready prototype (10:13) How Claude selected the right models for medical entity recognition (12:04) Using Streamlit to create accessible UIs for non-technical users (16:04) Detecting and redacting PHI in unstructured clinical notes (18:44) Generating the Common Technical Document (CTD) for FDA submission (21:54) Tracking and displaying AI operation costs for stakeholder buy-in (24:38) Real-world impact on vaccine development timelines and costs (26:12) Creating an AI communication coach for product managers (30:22) Training Claude on classic literature and persuasion techniques (31:53) Analyzing a complex stakeholder scenario with multiple competing priorities (34:40) Getting personalized communication strategies inspired by tech leaders (35:40) Summarizing strategic approaches (38:26) Conclusion and final thoughts *Tools referenced:* • Claude: https://claude.ai/ • Streamlit: https://streamlit.io/ • Anthropic Console: https://console.anthropic.com/ • Claude Sonnet 4: https://www.anthropic.com/claude/sonnet *Other references:* • Claude project chat (AI Product Management Stakeholder Challenges): https://claude.ai/share/caba4ab0-b28a-480c-8633-71920b12999e • XML: https://www.w3.org/XML/ • Python: https://www.python.org/ • RegEx: https://regex101.com/ • Moderna: https://www.modernatx.com/ • FDA: https://www.fda.gov/ • Project Gutenberg: https://www.gutenberg.org/ • FDA Biologics License Application: https://www.fda.gov/vaccines-blood-biologics/development-approval-process-cber/biologics-license-applications-bla-process-cber • Protected health information (PHI): https://www.hhs.gov/hipaa/for-professionals/privacy/laws-regulations/index.html _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Claire VohostPrerna Kaulguest
Jul 13, 202545mWatch on YouTube ↗

At a glance

WHAT IT’S REALLY ABOUT

AI workflows cut months from 60k-page FDA submission drafting time

  1. The episode shows how life-sciences regulatory submissions (e.g., a ~60,000-page BLA/CTD package) can be accelerated using a GenAI-driven workflow that produces strict XML outputs and redacts PHI from messy clinical notes.
  2. Prerna starts by prompting Claude like a “software engineer,” using a PM-style narrative (impact, stakeholders, demo goals) to get a full implementation plan, setup guide, and code scaffolding.
  3. She then wraps the workflow in a Streamlit UI for non-technical users, adds synthetic-data generation for testing, and includes trace-and-cost transparency to address organizational concerns about scaling AI usage.
  4. In a second workflow, she uses Anthropic’s Prompt Generator and Projects to create a structured stakeholder-influence coach, grounded in public-domain persuasion/literature sources, that outputs meeting prep, agendas, and anticipated exec questions.

IDEAS WORTH REMEMBERING

5 ideas

Treat the LLM like an engineering partner, not just a chatbot.

Prerna’s strongest results came from giving Claude a PM-grade problem statement (why it matters, expected product, demo goals), which yielded not only code but also setup docs and a demo narrative—accelerating initial solution shaping.

Strict output formats (XML) must be a first-class requirement.

Because CTD/BLA submissions require rigid XML structure, she made “structured, schema-like generation” a core constraint from the start—reducing downstream rework and making the workflow viable for regulated submissions.

PHI redaction is harder in free-text notes than in tables.

Clinical data contains both structured fields and unstructured clinician notes; effective redaction needs medical NER-style detection beyond column-based masking, especially to catch names, dates, and other identifiers embedded in prose.

A simple UI can be the adoption unlock for non-technical stakeholders.

Wrapping the pipeline in Streamlit turned a script into a usable internal tool—critical for cross-functional teams like medical writers and regulatory specialists who need buttons and previews, not notebooks.

Use synthetic data to test privacy and formatting workflows safely.

Generating synthetic trial data enabled validating PHI detection/redaction and document outputs without exposing real patient information—an important practice in regulated environments.

WORDS WORTH SAVING

5 quotes

We had to develop a nearly 60,000-page document. Would have taken about four to six months of effort and nearly 20 specialists...

— Prerna Kaul

The thinking I had in mind is that Claude is a software engineer... trying to tell them why it matters... what end product we want to produce.

— Prerna Kaul

The first is... the BLA is a structured document... XML-based... The second was... detect PHI... because it's patient data.

— Prerna Kaul

If you're saving on time, you're bringing life-saving vaccines in the hands of people who actually need it.

— Prerna Kaul

If you're getting internal resistance to cost... bring true transparency [to] ROI and investment.

— Claire Vo

Regulatory documentation burden in biotech (BLA/CTD)Structured XML generation requirementsPHI detection and redaction in unstructured clinical notesStreamlit as lightweight internal product UISynthetic data for validation/testingToken/operation cost transparency and ROI narrativesPrompt Generator + Projects for stakeholder communication coaching

High quality AI-generated summary created from speaker-labeled transcript.

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

Add to Chrome