AI Enhancement Layer
The AI Agent Concept
The AI Agent reads all operational data, identifies patterns linked to KPI gaps, and suggests prioritized R&T challenges — replacing months of manual workshops with automated, fact-based intelligence. The Committee retains full decision-making authority.
How the AI Agent Works
Read → Analyze → Detect → Suggest → Committee Validates
1
Reads
Company strategy, KPIs & targets
2
Analyzes
Incidents, shutdowns & failures
3
Detects
Underperforming KPI areas
4
Suggests
Prioritized R&T challenges
Committee
Reviews, validates & approves
Data Inputs
Incident Reports
All recorded incidents across MAA & MAB refineries
Bad Actor Lists
Equipment with repeated failures and high maintenance cost
Unplanned Shutdowns
Shutdown logs with duration, cost, and root cause data
Equipment Failure Records
Failure mode, frequency, and impact data per asset
Environmental Exceedances
KEPA limit violations and environmental incident records
Annual Reports & KPIs
Corporate KPIs, LTSP 2040 targets, and financial performance
Corporate Strategy
KPC-LTSP 2040 strategic priorities and KNPC objectives
AI Outputs
KPI-Linked Challenge List
Each challenge is directly linked to a specific KPI gap, making the business case clear and measurable.
Root Cause Analysis
AI identifies the underlying root causes from data patterns, not just symptoms.
Risk Ranking
Challenges ranked by Business Impact × Doability, replacing manual scoring workshops.
Draft Project Charters
Pre-populated charter templates for each challenge, ready for committee review.
Live Performance Dashboard
Real-time tracking of KPI progress, challenge resolution, and project execution.
Tracking Sheets
Automated project tracking sheets replacing manual spreadsheet updates.
Committee Role
The AI Agent suggests challenges based on facts and data. The R&T Committee — comprising R&T staff and department stakeholders — reviews, validates, and approves the final challenge list. Human expertise remains central to all decisions.
Manual Process vs. AI-Enhanced Process
Roadmap Step
Current (Manual)
AI-Enhanced
Time Saved
Challenge Identification
Months of workshops with 9 stakeholder departments, 254 raw challenges from manual brainstorming
AI reads operational data continuously, surfaces challenges automatically with evidence
~3 months
Validation
Multiple validation workshops to check relevance to KNPC business
Committee reviews AI-suggested list — fact-based discussion, not opinion-based
~6 weeks
Prioritization
Manual scoring of each project on Business Impact and Doability matrices
AI auto-scores using real data; committee confirms or adjusts
~4 weeks
Project Charters
R&T team manually drafts 128 project charters from scratch
AI generates draft charters pre-filled with data; engineers review and finalize
~8 weeks
Progress Tracking
Manual spreadsheet updates, periodic reporting, delayed visibility
Live dashboard with real-time KPI tracking and automated alerts
Ongoing