Intelligent Analytical Document Control for Steps Biosciences: Claude AI and AWS

About the client

Steps Bio (Steps Biosciences) is a sterile product development company with operations in the United States and India, specialising in complex generic and specialty sterile dosage forms for human and veterinary applications, including injectables, ophthalmic and otic products. Its Formulation and Analytics function manages a growing portfolio of active pharmaceutical projects, each requiring structured test data submission, second-person quality review, and stability reporting across multiple test types and timepoints.

Problem Statement

The Analytics team managed pharmaceutical test data and stability reporting through fragmented, manual processes. Instrument data was transcribed by hand, submitted over email, and reviewed informally, with no structured workflow enforcing data integrity or review accountability. Generating a final stability report could take days after the underlying data was already ready.

Challenge

Manual Data Transcription and Error Propagation
Analysts manually transcribed instrument readings into calculation sheets with no automated validation layer to catch inconsistencies before submission.
No Structured Submission or Review Workflow
Analytical submissions were routed informally through email and shared folders with no defined ownership, status tracking, or accountability at any stage.
Dual Submission Format Complexity
Excel-based calculation outputs and direct PDF instrument data both needed to flow through the same review process, but no system existed to handle either consistently.‍
No Enforced Quality Gate
Nothing prevented an analyst from reviewing their own data. Compliance with second-person verification depended entirely on manual discipline.
Manual, Disconnected Report Generation
Final stability reports were authored individually after review, with no automated pipeline connecting the reviewer approval decision to report creation.

Solution

Pace Wisdom built an enterprise Analytical Document Control Platform on AWS, with Anthropic's Claude powering automated stability report generation across a structured, role-gated workflow.

Role-Based Workflow with Strict RBAC
Three defined roles (Administrator, Analyst, Reviewer) with platform-enforced access boundaries. Each role can only perform actions within its scope.
Auto-Populated Analytical Templates
On issuance creation, the system generates a pre-filled Excel calculation sheet carrying project metadata, analyst details, batch references, and instrument IDs, eliminating manual transcription at source.
Multi-Format Submission via Amazon S3
Analysts submit via Excel mode (CALC file and RAW PDFs) or PDF-only mode, both using Amazon S3 pre-signed URLs for secure, traceable file transfer.
Enforced 4-Eyes Principle at the Reviewer Gate
Reviewers access a dedicated Reviewer Gate, separate from the Analyst view. The platform blocks self-approval and requires a mandatory comment on rejection, returning submissions for correction.
Claude-Powered Report Generation on Amazon Bedrock
On approval, Anthropic's Claude automatically synthesizes submitted data into a structured, audit-ready stability report. Report status is tracked in real time and accessible to all roles on completion.
End-to-End Audit Trail
Every issuance, upload, review decision, and report generation event is logged with full traceability from project creation to final report download.

Intelligent Analytical Document Control for Pharma: Claude AI and AWS

No items found.

Technology used

React.js, Node.js, AWS, Amazon S3, Amazon Bedrock, and Anthropic Claude for AI-powered analytical report generation.

Impact

37+

Active Pharma Projects Managed on the Platform

~90%

Reduction in Stability Report Turnaround Time

2

Submission Modes Unified Within One Governed Workflow

Transform your business with Pacewisdom

Talk to us