PT XYZ sebagai kontraktor pertambangan batu bara mengelola risiko kritis melalui program Critical Risk Management (CRM) berbasis inspeksi observasi aktivitas. Data tahun 2023–2025 menunjukkan bahwa pendekatan ini belum memadai: tingkat pencapaian observasi hanya 72,1%, sementara 9 dari 13 kejadian fatality dan Lost Time Injury berasal dari aktivitas pengoperasian dan interaksi alat-alat berat (A2B). Persoalan mendasarnya bukan hanya konsistensi pelaksanaan inspeksi, melainkan ketiadaan mekanisme yang secara sistematis memverifikasi apakah pengendalian yang paling krusial terhadap risiko fatal benar-benar andal sesuai standar kinerjanya. Penelitian ini mengaplikasikan kerangka Critical Control Management (CCM) ICMM (2015) pada dua Material Unwanted Events (MUE): kehilangan kendali unit A2B (MUE-1) dan interaksi A2B dengan unit, sarana, atau pekerja (MUE-2). Desain penelitian berupa studi kasus tunggal dengan pendekatan kualitatif. Data dikumpulkan melalui tinjauan dokumen risk register tiga lini bisnis (156 entri), wawancara mendalam terhadap 18 informan dari 7 klaster peran yang dianalisis secara tematik, dan observasi lapangan melalui Critical Control Verification (CCV). Analisis dilakukan melalui konstruksi bow-tie, seleksi kontrol dengan kriteria ICMM, penetapan critical control melalui BHP Decision Tree, perumusan performance criteria, dan verifikasi lapangan. Dari 156 entri risk register, proses penyaringan menghasilkan 13 critical control (CC-01 hingga CC-13). Hasil CCV menunjukkan 5 critical control memenuhi seluruh performance criteria (Pass) dan 8 tidak memenuhi (Fail). Critical control yang Pass terutama merupakan kontrol dengan siklus pemeliharaan teknis yang stabil, sedangkan yang Fail menunjukkan laju degradasi yang melampaui kecepatan pemulihan. Pembahasan mengungkap dua temuan analitis utama: dominasi kontrol administratif pada risk register konvensional yang menutupi kerentanan pertahanan, serta ketergantungan struktural pada manusia yang tidak terbatas pada kontrol bertipe tindakan tetapi juga melekat pada pemeliharaan kontrol berlabel objek dan sistem. Kontribusi kerangka CCM bukan pada penambahan kontrol baru, melainkan pada transformasi status kontrol yang sudah ada dari diasumsikan hadir menjadi terverifikasi memenuhi atau tidak memenuhi criteria, sehingga memberikan dasar faktual bagi alokasi sumber daya perbaikan.

PT XYZ, a coal mining contractor, manages critical risks through an activity-based Critical Risk Management (CRM) program centred on planned inspections and observations. Data from 2023–2025 revealed that this approach remained insufficient: the observation achievement rate stood at only 72.1%, while 9 of 13 fatality and Lost Time Injury events originated from heavy equipment (A2B) operations and interactions. The fundamental issue was not merely inspection consistency, but the absence of a mechanism that systematically verifies whether the most crucial controls against fatal risk are truly reliable against their performance standards. This study applied the ICMM (2015) Critical Control Management (CCM) framework to two Material Unwanted Events (MUEs): loss of control of an A2B unit (MUE-1) and A2B interaction with other units, infrastructure, or workers (MUE-2). A single embedded case study with a qualitative approach was adopted. Data were collected through document review of risk registers across three business lines (156 entries), in-depth interviews with 18 informants from 7 role clusters analysed thematically, and field observation through Critical Control Verification (CCV). The analytical sequence comprised bow-tie construction, control selection using ICMM criteria, critical control determination through the BHP Decision Tree, performance criteria formulation, and field verification. From 156 risk register entries, the screening process yielded 13 critical controls (CC-01 through CC-13). CCV results showed that 5 critical controls met all performance criteria (Pass) while 8 did not (Fail). Controls that passed were predominantly those with stable technical maintenance cycles, whereas those that failed exhibited degradation rates exceeding their recovery speed. The discussion revealed two principal analytical findings: the dominance of administrative controls in conventional risk registers that masked defence vulnerabilities, and a structural human dependency that extended beyond act-type controls to the maintenance of object- and system-labelled controls. The contribution of the CCM framework lies not in adding new controls, but in transforming the status of existing controls from assumed present to verified as meeting or not meeting criteria, thereby providing a factual basis for corrective resource allocation. Keywords: critical control management, critical control verification, control reliability, bow-tie, heavy equipment, coal mining