Company: Novo Nordisk
Role: Lean Partner / Transformation Lead
Context: Regulated pharmaceutical manufacturing
Scope: Batch release execution after process re-engineering
Technology: Robotic Process Automation (RPA)
Context
After reducing release lead time and stabilising deviation handling, the batch release process had become:
- Stage-based and predictable
- Governed by clear ownership and KPIs
- Aligned across Production, QA, and QC
- Stable under audits and leadership scrutiny
Only at this point did automation become a real opportunity.
Before this, release execution involved manual, repetitive steps that added no decision value but consumed significant time and attention.
The Critical Decision
There was pressure to “automate release” early.
I deliberately did not.
Based on experience and data, automating earlier would have:
- Accelerated rework loops
- Embedded misaligned decision logic
- Increased compliance risk
- Delivered speed without control
So the decision was explicit:
First re-engineer and stabilise the system.
Then automate what remains purely mechanical.
What Was Automated (and what was not)
Automated
- Manual execution steps once all release criteria were met
- Data checks that were already standardised and validated
- System interactions that required no human judgment
Not automated
- Quality decisions
- Deviation assessments
- Risk evaluation
- Any step requiring interpretation or context
Automation was applied only where the process was already right-first-time.
My Mandate & Authority
I was asked to lead the Instant Release RPA implementation, with responsibility to:
- Ensure automation addressed validated constraints, not symptoms
- Maintain full compliance and audit readiness
- Deliver speed without introducing hidden risk
- Integrate RPA cleanly into existing governance and KPIs
I worked hands-on with QA, IT, and operations, ensuring business ownership remained clear.
What I Did
1. Validated automation readiness
- Confirmed release flow stability and predictability
- Ensured deviation handling was under control
- Verified that automated steps required zero judgment
Only after this validation did automation proceed.
2. Designed RPA around the re-engineered process
- Automated execution of release steps once criteria were fulfilled
- Embedded automation into existing KPIs and routines
- Ensured transparency rather than black-box execution
RPA followed the process — not the other way around.
3. Secured compliance and sustainability
- Ensured traceability and audit readiness
- Avoided hard-coding assumptions into automation
- Maintained human control over decisions
Automation reduced effort, not accountability.
Results & Impact
Hard outcomes
- Release execution time reduced from ~7.5 hours to seconds
- Manual effort removed from a critical but non-value-adding step
- Faster product availability without quality or compliance trade-offs
Structural outcomes
- Automation built on a stable, governed process
- No increase in deviations or audit findings
- Clear ownership and control preserved
- RPA became a scalable enabler, not a fragile workaround
Key Insight
Automation does not fix broken systems.
It only makes them fail faster.
By re-engineering first and automating last, speed, control, and compliance improved together — without creating technical debt or hidden risk.
Öner TANK 