Project
The Process Insight application is designed to deliver essential insights into pharmaceutical manufacturing. It features intuitive dashboards that detect deviations, send alerts, and facilitate data-driven decisions. It offers seamless access to essential parameters and attributes, consolidating data for real-time monitoring and reporting. Due to strict industry regulations, Process Insight emphasizes precision and clarity. Its main users, Quality SMEs, need a tool that simplifies complex data workflows while adhering to industry standards.
My role
- UX Lead/ Team lead
- Responsible for running design workshops to define the scope of the project.
- Creating a customer journey map for the product
- Conducting stakeholder interviews with key team members
- Refining the design operation processes and iterating over designs
Objective
The primary goal of the project was to:
- Integrate Process Insight into NGP(Next Generation Platform) to provide a unified experience.
- Maintain consistency in functionality while adapting the User Experience to NGP standards.
- Simplify the complex workflows of CPV (Continued Process Verification) management.
- Enhance information architecture for better navigation and usability.
- By achieving these objectives, we aimed to improve user efficiency, reduce cognitive load, and establish a robust decision-support system.
Research methods
- Customer and stakeholder interviews (Product, SMEs, Validation Engineers)
- Comparative analysis of IA and chart comprehension
- Moderated usability testing with scientific experts
- Archetype development for role-based behavior modeling
In a regulated environment, complexity is not just inconvenient. It is a risk.
01 Tension
Brilliant experts, buried in fragmented screens
Before the redesign, the CPV experience in ValGenesis was a landscape of fragmented screens, dense statistical output and workflows that had grown over years of incremental development.
Quality SMEs, some of the sharpest domain experts in pharmaceutical manufacturing, were working around it every day.
01
CPV plans scattered across multiple views.
02
Statistical charts that needed deep interpretation.
03
Batches and variables buried under non-intuitive tables.
04
Slow, error-prone configuration steps.
05
No cohesive information architecture.
06
Compliance-critical actions, such as electronic signatures, hidden inside inconsistent flows.
My mandate as UX Lead was not to redesign screens. It was to create a unified, navigable and reliable system that supports scientific decisions, regulatory compliance and operational confidence, inside the Next Generation Platform (NGP).
6
Steps in the redesigned configuration wizard
4
Discovery methods used before designing
3
Roles in the approval flow: author, reviewer, approver
4
UX metric categories defined
02 Discovery
Understand the system before changing it
I led several discovery cycles before touching a screen. The goal was to learn how these experts reason about reference batches, variability, PCA, CpK thresholds and deviation detection.
On-site sessions
With Quality SMEs and Validation Engineers, to understand their mental models.
Stakeholder interviews
From product managers to statistical SMEs to regulatory experts, to learn what “trust” means in a CPV system.
Heuristic evaluation
Of the legacy workflow, to find where the interface and user expectations diverged, especially around dense data and statistical visualisation.
Archetypes
Role-based behaviour patterns across SMEs, reviewers and approvers, to drive role-specific decisions.
The CPV workflow produced insights, but the system did not help users understand, validate or trust them.
The problem, once discovery was done
The information architecture did not support the scientific reasoning SMEs needed. The tool executed steps but did not explain them, and its navigation did not reflect CPV as a cyclical, decision-driven process.
That gave me the direction: redesign CPV as a guided, intelligible, end-to-end system, not a collection of isolated screens.
03 Alignment
Turning a complex domain into a shared mental model
Leading this redesign took more than craft. It took orchestration across teams, personas and constraints.
Cross-functional workshops
Mapping the CPV lifecycle from plan creation through reference batch selection, variability assessment, limit configuration, routine monitoring and reporting, and finding where users lost context or confidence. We also drew the boundary between NGP standards and CPV-specific needs.
Pattern sessions with developers
Converting complex statistical logic into reusable visual patterns, with loading, threshold and alert behaviour that scales across variables and plans, while staying feasible in the legacy architecture.
Guiding junior designers
Giving them the structural rules of the design system and the interaction logic, then delegating components while keeping the whole cohesive.
Designers, developers and SMEs ended up sharing one mental model of the CPV system.
04 Decisions
Four calls that reorganised the experience
Decision 01
Turn a scattered setup into a guided wizard
Context
Creating a CPV plan was spread across screens and did not follow how SMEs actually decide.
The call
Redesign creation as a structured, six-step flow that mirrors real SME decision-making: new plan, workflow, variables, reference batches, outputs and inputs, finalization.
What it did
The wizard became the backbone of CPV clarity.
Step 2 · Configure Workflow
Reviewer and approver roles are set up visually, so responsibility is clear from the start.
Step 3 · Select Variables
Critical quality attributes are flagged and selectable in one list.
Step 4 · Reference Batches & Preview
Variability charts sit beside the settings that change them.
Step 4 · Preview
A preview step lets experts verify variables and reference batches before setting limits.
Step 5 · Configure Outputs & Inputs
Control chart type, sigma, CpK threshold and distribution tests, configured beside live charts.
Step 5 · Response to Signals
Responses to signals are defined as a flow: what triggers, what is notified, what happens next.
Step 6 · Finalization
One summary of everything in the plan before it is submitted.
Decision 02
Make variability understandable, where decisions are made
Context
PCA, Hotelling’s T² and Q residual plots were visually overwhelming and detached from the workflow.
The call
Bring all variability indicators into one interpretable layout, with consistent legend behaviour across the charts, thresholds and confidence intervals emphasised in NGP’s visual language, placed exactly where SMEs decide.
What it did
Better comprehension, and less back-and-forth navigation.
Decision 03
Give routine monitoring a hierarchy, and dense tables a structure
Context
The routine dashboard mixed charts, tables and signals in a flat layout, and plan and batch tables had to balance density, regulatory accuracy and fast scanning.
The call
A clear hierarchy: Overview, Control Charts, Histogram, Capability Metrics. Modular chart containers that follow one pattern, breadcrumbs for batch-level deep dives, and tables with column grouping, anchored headers and adaptive row density.
What it did
Monitoring went from hunting and guessing to a predictable, navigable flow, and dense scientific data stayed readable without losing precision.
Decision 04
Design compliance moments as patterns
Context
Electronic signatures, reviewer workflows and audit-trail entries are the moments where trust is won or lost.
The call
A unified approval card for author, reviewer and approver actions, modal-driven electronic signatures with clear legal messaging, pending separated from completed work, and consistent signals for responsibility and status.
What it did
Less ambiguity around regulatory actions, and a system people can trust.
05 Proof
Defining what “good” looks like
To make the value measurable, I defined a compact metrics framework tied to CPV’s scientific and operational goals. It let me evaluate clarity, efficiency and trust with evidence, not only perception.
Efficiency
Time to complete CPV plan configuration. Navigation steps per task.
Comprehension
Correct interpretation of PCA and other complex indicators. SME confidence in variability assessment.
Trust and compliance
Reduced use of external tools. Completion success of e-signature workflows.
Consistency
Adoption of NGP components. Fewer UI inconsistencies across CPV.
01
Baseline benchmark of the legacy experience.
02
Moderated usability sessions with SMEs, testing comprehension and workflow fit.
03
Prototype interaction analytics to confirm efficiency gains.
04
Targeted debriefs with Validation Engineers and SMEs on decision confidence and trust.
The specific numbers are confidential. The direction is not:
Faster configuration
A structured wizard and fewer navigation jumps between steps and charts.
Clearer statistics
SMEs reported better understanding of PCA and CpK indicators, with less reliance on external statistical tools.
A reusable pattern library
A CPV-wide library usable across the NGP ecosystem, and a scalable information architecture for future features such as trending and automated deviation insights.
Stronger compliance
Clearer audit-trail visibility and signature patterns, with compliance-critical actions clearly separated.
06 Reflection
What this shows about how I lead
CPV is a domain where clarity is foundational. A misread deviation or a hidden signal can have real consequences. Leading it asked for four things:
Systems thinking
To reorganise complexity instead of decorating it.
Scientific UX
To make statistical insight understandable to the people who own the science.
Orchestration
To align designers, developers and SMEs on one model.
Vision
To see a unified CPV experience inside NGP, and bring the team there.
Next
To transform systems, not screens.
That is how I approach design leadership: guide teams toward clarity in the places where clarity matters most.
