Technical Excellence Centre

The Challenge

Despite having tools in place, PETRONAS faced four core problems:

1. Scattered systems: Existing platforms held key data, but none provided a holistic view. Users couldn’t monitor issues or trends across the enterprise in one place.
   
2. Reactive decisions: There was no predictive layer to anticipate failures or inefficiencies. Most actions were taken after losses occurred.
   
3. Low financial visibility: There was no clear monthly view of how much money was lost due to unresolved technical issues nor visibility into which OPU (Operating Unit) was most responsible.
4. Disconnected cause and effect: Without clarity on how one unit’s performance impacted others, it was difficult to solve problems at the root level.
   

Objectives

• Unify fragmented insights across technical, financial, and operational systems into one enterprise-grade interface
• Enable foresight through AI-generated alerts with estimated timelines and cross-OPU impact mapping
• Quantify monthly financial loss tied to unresolved technical inefficiencies, supporting data-driven action
• Empower users with clarity, through a redesigned Live Value Chain that highlights risk, origin, and resolution options in real time
• Support multiple user roles, from technical engineers to high-level decision-makers, with scalable dashboards and visual clarity

Design Approach

1. System Unification Strategy

• Conducted a discovery workshop to understand what's currently working and not for potential users, gathering initial mental models and expectations
• Performed contextual inquiries with users from upstream, gas, and downstream to observe tool usage and workflow in real-world settings
• Audited existing tools and visualizations across upstream, gas, and downstream with affinity mapping and brainstorming session
  
• Mapped overlaps, gaps, and conflicting user flows with card sorting
  
• Reimagined the flows as modular, reusable components within TEC using insights from co-creation sessions to ensure future extensibility and alignment with actual user logic

‍

‍

‍

2. User Research

• Interviewed 15+ users (engineers, analysts, executives, planners) and with the additions of SME of each disciplines (we started with 3 and as the project goes, we have up to 8 people now which means more requirements from each disciplines need to be handled), the workload is great but the knowledge input is even greater which help with designing TEC structure
  
• Identified friction points around delay, data trust, and lack of traceability
  
• Mapped journey needs for three core roles: strategic, operational, and technical users

‍

3. AI-Driven Design Integration

The design team worked with data teams to design insight panels showing:

• Upcoming failure forecasts
  
• Estimated timelines
  
• Chain reactions across OPUs
  
• Maintenance order cost from manufacturers

which at first we started with the idea of the normal dashboard, but it is only give information like other platforms do, so on another workshop, which the design team fully focus on this level of intelligent way of delivering information, we went with predefined chat from chatbot,

‍

‍

so on another workshop, which the design team fully focus on this level of intelligent way of delivering information, we went with predefined chat from chatbot,

‍

‍

‍

and the other additional information that is from the first draft wireframe is being put to another layer as a way to help new users to feel convinced that what the AI said is correct. This also a challenge on its own across the platform because the users still have little to no trust to rely on AI so they still want to work the way they work before and compares the result they get with the answers the AI gives. We work along side the Data Science team, Data Analyst team and AI Engineer team to make this come true,

‍

We created a template kind of design with 3 main section to dissect, the first two widget is fixed, with the top bar to show the big number for the page, the key observation and analysis widget which will always be at the bottom left, showing the summary of the page, which is also summarized by the AI and lastly, everything in between should be filled with 2-8 dynamic widgets accordingly.

‍

4. Financial Visibility Layer

• Built visualizations showing:
  
  • Monthly financial losses by OPU
    
  • Accumulated costs tied to unresolved issues
    
  • Forecasted savings from potential resolutions

‍

5. Live Value Chain Redesign

• Created a simplified, color-coded map of the enterprise chain which help the users to instantly see:
  
  • Which OPU is currently underperforming
  • How its issues affect downstream units
  • Whether the root cause comes from upstream
  • The recommended fix, pulled from AI-generated reports

‍

‍

Key Feature I Designed

• Live Value Chain Map - One-screen overview of enterprise performance, with cause-effect indicators
• Red Dot Alert System - Real-time flags tied to loss, performance, or risk
• AI Insights Panel - Forecasts upcoming failures + cross-OPU impact with timelines
• Financial Loss Tracker - Monthly breakdown of losses and where they're originating
• Role-Based Dashboard - Tailored data views for engineers, analysts, planners, and executives

What TEC Solves

• Unifies data & decision-making across 30+ OPUs into a single enterprise-grade interface
  
• Surfaces AI-driven alerts and forecasts, warning users of upcoming issues, their timeline, and cross-departmental impact
  
• Visualizes financial losses per month, linked to specific OPUs and technical inefficiencies
• Empowers strategic action through a redesigned Live Value Chain, showing dependencies, performance status, and recommended resolutions

‍

Results

• Unified visibility across 30+ OPUs in one platform
  
• Cut issue detection time from weeks to minutes
  
• Enabled proactive intervention, with AI forecasts and impact maps
  
• Improved stakeholder alignment with a shared, always-updated source of truth
  
• Recognized internally as a key strategic platform for PETRONAS’ digital transformation

Visual Design

Due to the confidentiality of this project and the sensitivity of the data involved, I am unable to share the full user interface. As per project requirements, any visuals must be blurred or redacted. I appreciate your understanding and apologize for any inconvenience.

‍

What I Learned

• Designing for clarity is designing for trust, especially in enterprise settings where action has real-world cost.
  
• True unification isn’t about putting everything in one place, it’s about aligning systems, users, and purpose.
  
• AI insights are only useful when framed in context and consequence, the design must tell a story, not just show data.

Reflection

Looking back, TEC was one of the most complex and layered projects I’ve ever led. While the core design challenges were around unifying data and improving visibility, the real difficulty was aligning across departments, each with their own language, tools, and priorities.

We didn’t just need a usable platform, we needed a shared mental model. That took time. Multiple rounds of workshops, conflicting opinions, and trial-and-error on how to balance technical accuracy with visual clarity.

If I could do one thing differently, I would have brought users into validation loops even earlier, especially those at the edges of the system, who often deal with the most friction. We did well in surfacing high-level insight, but edge-case workflows surfaced only later in QA and support feedback.

Still, I’m proud of what we achieved. TEC is more than just a dashboard, it’s a foundation for smarter, more proactive decisions at a national scale. And for me, it became a reminder that great UX isn’t just what you build, it’s who you bring together in the process.