AI-led Automation of API Change Request Management
Know how a Fortune 500 pharma company cut API change-request time from days to hours without changing the core systems.
90%
Automation of ACR lifecycle
100%
Reduction in turnaround time from days to hours
100%
Auditable, standardized workflows
0
Core system replacements
Overview
One of the world’s largest pharmaceutical companies — operating across 80+ countries with a portfolio of 1,500+ medicines, relied on a manual intense , email-driven process for managing API change requests. Vendor notifications arrived by email, teams classified and documented them manually, and updates were coordinated across disconnected systems. Compliance risk grew with every hand-off. Turnaround times stretched into days.
Saxon AI deployed an AI assistant built on its AIssist platform to automate the end-to-end ACR lifecycle. Today, 80–90% of the process runs autonomously, from reading a vendor notification to updating the Quality Management System, with humans involved where judgement matters.
“We didn’t want a point solution that patched one step. We needed something that understood the entire change lifecycle and could act on all of it intelligently.”
— Head of Global Quality Operations
The Challenge
Scale adds pressure in every direction. A single specification change from a key API supplier can affect multiple finished products, multiple markets, and multiple regulatory filings simultaneously. Getting it right — and getting it done quickly — is not optional.
What was breaking down
API change request management is complex. On the surface, it appears to be information-routing: a supplier notifies the company of a change, the appropriate teams evaluate it, and a decision is recorded. In practice, it is a multi-step, multi-system coordination challenge where errors carry regulatory weight.
API vendor change notifications — covering process, specification, site, and administrative updates — arrived as unstructured emails with no standard intake channel. Quality teams manually classified each change, chased supporting documents across time zones, and re-entered data across disconnected QMS and ERP systems.
With no automation to verify completeness, flag ambiguous cases, or maintain a real-time audit trail, the process created compounding compliance risk, unpredictable turnaround times, and quality professionals spending the majority of their time on coordination rather than judgment.
Challenge | Business Impact |
Unstructured email intake | No consistent record, high misrouting risk |
Manual change classification | Inconsistency, missed regulatory triggers |
Document chasing | Days of delay per request, unpredictable timelines |
Disconnected QMS and ERP | Duplicate data entry, missed supply impact assessments |
No real-time audit trail | Compliance risk, difficult to defend in inspections |
The solution
Saxon AI deployed an intelligent AI assistant across the full ACR lifecycle — binding intake, classification, document management, system updates, and human escalation into a single governed workflow.
The assistant:
- Monitors vendor inboxes and interpret unstructured notifications
- Classifies changes (process, spec, site, admin) using enterprise logic
- Collects and validates documents proactively
- Creates and updates ACR records directly in the QMS
- Runs automated supply, product, and revenue impact assessments via ERP integration
- Sends only exceptions or ambiguous cases to reviewers with complete decision packages
Under the hood
The AI assistant is built on Saxon AI’s AIssist platform and integrated via secure API connections into the existing QMS and ERP systems — with no changes to either system’s configuration or validated state. A data layer logs every agent action, system call, and data transformation, ensuring full auditability for regulatory inspection.
Quality professionals were embedded in the design process, with their domain knowledge directly shaping the classification logic and exception-handling rules the assistant operates by.
Key capabilities deployed:
- Natural language interpretation of vendor notification emails
- Automated change classification with full audit trail
- Proactive document collection, tracking, and validation
- Direct QMS record creation and real-time lifecycle updates
- ERP-integrated supply and revenue impact assessment
- Human-in-the-loop escalation with structured decision ladders
- Real-time audit logging across all agent actions and data flows
“The AI assistant gives us a complete picture. Our reviewers tare freed up for more productive and decision-making work.”
— Director of Quality Assurance